‘3'“?!
¢ .w'
‘ {if

‘ I
2:233:
an

1;

’
3:
«’3'?

b

was a}!
gain!
1. .31 v ‘ ‘

cg ‘

.D

 

THESIS

LIBRARIES
/ .o*--'/' MICHIGAN STATE UNIVERSITY
'- EAST LANSING, MICH 48824-1048

(“999 (320.9

This is to certify that the l
dissertation entitled

PATIENT PERCEPTION OF PROVIDER ADHERENCE
T0 TREATMENT AND SELF -CARE STANDARDS:
THE CASE OF DIABETES SELF-CARE

presented by
Ragnhild S. J. Bundesmann ‘

has been accepted towards fulfillment
of the requirements for the

PhD degree in Sociologx

 

 

/

 

F" Major PrbfesIor's Signatué

W44—

Date

MSU is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

6"01 c:/CIRC/DateDue.p65-p.15

 

PATIENT PERCEPTION OF PROVIDER ADHERENCE TO TREATMENT AND
SELF -CARE STANDARDS: THE CASE OF DIABETES SELF-CARE

By

Ragnhild S.J. Bundesmann

A DISSERTATION

Submitted to Michigan State University
In partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY
School of Sociology

2004

ABSTRACT

PATIENT PERCEPTION OF PROVIDER ADHERENCE TO
TREATIVIENT AND SELF-CARE STANDARDS: THE CASE OF

DIABETES SELF-CARE

By

Ragnhild S.J. Bundesmann

Diabetes is a serious and prevalent disease, for which there is no cure. Treatment
consists of monitoring, counseling, and self-care. Self-care can be demanding and
complex, therefore people with diabetes need training and support. Providers must teach
and encourage patients to maintain the diabetic life-style and to avail themselves of
diabetes self-care education. This teaching and encouragement is part of the practice
guidelines of the American Diabetes Association (ADA). There is evidence that
physicians do not always deliver the care according to practice guidelines, both in the
area of monitoring as well as counseling. This cross sectional study investigates the
relationship between provider information giving and the likelihood of the person with
diabetes to practice appropriate self-care behaviors. The sample for the study came from
M-CARE, a university affiliated managed care organization. The data was examined
with regression studies that controlled for patient characteristics. Provider information
giving regarding exercise, blood glucose monitoring and foot examination have a
moderate but significant impact on those behaviors. Provider modeling of foot

examination and monitoring of home blood glucose levels also have a significant effect

on those two self-care behaviors. Summary indexes for provider information giving and
provider modeling have a significant impact on aggregate measures of diabetes self-care.
It was found that older patients tend to receive less information. The study concludes that
patient self-care behaviors need to be studied while controlling for provider information

giving and provider modeling of desired behaviors.

ACKNOWLEDGEMENTS

I would like to thank all those who have supported me during the dissertation
process, particularly my committee members, Cliff Broman PhD, Tom Connor PhD,
Harry Perlstadt PhD and the Dean’s representative, Lorraine Weatherspoon RD, PhD.
Many thanks go to Stan Kaplowitz Ph.D. who patiently mentored and guided me through
the process. I also owe William Herman M.D., MPH many thanks for sharing his
TRIAD data that made the dissertation possible. Blue Care Network supported my
tuition payments for the Master of Public Health and for the Ph.D.. I am grateful for the
financial help. Many thanks go to my staff at McLaren Regional Medical Center who

consistently encouraged me for the last seven months.

More thanks go to my family, particularly our four children Evelyn, Mark,
Christopher and Michael who cheered and supported me through the long study period.
Most of all I need to thank my husband Herbert Bundesmann who faithfully encouraged

me to finish not only the dissertation, but all the studies that led to it

iv

TABLE OF CONTENTS

List of Tables ................................................................................ v
List of Figures ............................................................................... vi
Introduction ............................................................................... 1
Literature Review ........................................................................... 2
Diabetes: The disease, prevalence and treatment 2
Diabetes treatment, the sick role and physician behaviors 10
Theory of health behavior that applies to this study: The Health Belief 14
Model 21
Diabetes self-care behaviors and provider practice 34

Unresolved questions concerning provider’s practice and self-care

Hypotheses ................................................................................... 37
Provider advice and self-care behavior 37
Self-care and provider modeling 38
Self-care and provider referral and information giving 41
Severity of diabetes and physician practice 43
Demographics, provider information giving and self-care 44

Methods ....................................................................................... 47
M-CARE structure and membership 47
Variables selected for this study and their measures 5 3
Coding and variable computation 62

Results ......................................................................................... 67
Sample description 67
Hypotheses testing and data analysis 73

Self care activities 73
Diabetes education 94
Disease complications 96
Demographic variables and provider behaviors 98

Discussion of the findings ................................................................. 101
Summary of self-care behaviors 101
Summary of diabetes complications and demographics 105
Data and methods considerations 106
Background variables 108
The differences in self-care behaviors 116
Effect of provider information giving 119
Implications 122
Limitations 124
Conclusions 126

Bibliography ................................................................................ 128

LIST OF TABLES

Table 1: Adherence to practice guidelines according to various data sources ....... 31
Table 2: Ethnicity 64
Table 3: Descriptive summary of continuous variables ................................. 68
Table 4: Summary of self-care and provider behavior variables ..................... 69
Table 5: Missing Variables ................................................................ 71

Table 6: Logistic regression: The effect of patient characteristics and provider
information giving and modeling on patient home glucose monitoring .............. 75

Table 7: Logistic regression: The effect of patient characteristics and provider
information giving and modeling of foot care on foot examination .................. 77

Table 8: OLS regression: The effect of provider information giving and patient
characteristics on exercising .............................................................. 79

Table 9: Logistic regression: The effect of patient characteristics and provider
information giving on obtaining the flu vaccination .................................... 81

Table 10: Logistic regression: The effect of patient characteristics and provider
information giving on retina screening in the last 12 months ......................... 83

Table 11: OLS regression: The effect of patient characteristics and provider
information giving and modeling on office based self-care behaviors ............... 86

Table 12: OLS regression: The effect of patient characteristics and provider
information giving and modeling on home based self-care behaviors ............... 88

Table 13: OLS regression: The effect of patient characteristics and provider
information giving and modeling on aggregated self-care behaviors ................. 90

Table 14: OLS regression: The effect of patient characteristics and diabetes
education on diabetes self-care behaviors ................................................ 92

Table 15: Log regression: Receiving information and use of diabetes education. 94

Table 16 Summary of self-care and impact of various factors ........................ 95

Table 17 The impact of patient characteristics and complications on seeing a
specialist ....................................................................................... 96

Table 18: The impact of demographics on provider information giving ............. 99

vi

TABLE OF FIGURES

Figure 1: Key concepts and definition of the health belief model ..................... 20

Figure 2: Proposed relationship of patient characteristics, physician practice and 42
self-care behaviors ...........................................................................

vii

PATIENT PERCEPTION OF PROVIDER ADHERENCE TO
TREATMENT AND SELF-CARE STANDARDS: THE CASE OF

DIABETES SELF-CARE

Introduction

Diabetes is a serious and prevalent disease, for which there is no cure. Treatment
consists of monitoring, counseling, and self-care. Self-care can be demanding and
complex, therefore people with diabetes need training and support. Providers must teach
and encourage patients to maintain the diabetic life-style and to avail themselves of
diabetes self-care education. This teaching and encouragement is part of the practice
guidelines of the American Diabetes Association (ADA). There is evidence that
physicians do not deliver the care according to practice guidelines, both in the area of
monitoring as well as counseling. This study aims to explore the relationship between
provider information giving and the likelihood of the person with diabetes to practice
appropriate self-care behaviors. A central question is the following: to what degree does
failure of the provider to offer appropriate information translate into patients failing to

practice appropriate self—care.

Diabetes: 1119 disease, prevalence, a_nt_I_ treatment

Diabetes disease process and complications

This section describes diabetes, specifically type 2 diabetes, the disease process,
complications, prevalence, and treatment. The goal of diabetes treatment is glycemic
control. Glycemic control is achieved through exercise and diet, and if needed,

pharmaceuticals to help achieve blood glucose control.

Diabetes is comprised of a group of metabolic diseases that are marked by
elevated blood glucose levels. It is estimated that about 17 million Americans have
diabetes (V inik 2003). Type 1 diabetes in the vast majority of cases is due to pancreatic
islet B-cell destruction and the absence of insulin production. Therefore, people with type
1 diabetes always require insulin for survival. The onset of type] diabetes generally
occurs before age 30 and is acutel (American Diabetes Association [ADA], 2001). This
study has excluded people with type 1 diabetes and focuses on people with type 2
diabetesz. Type 2 diabetes is the most prevalent form of diabetes and results from insulin
resistance with an insulin secretory defect. A person with type 2 diabetes may produce
large amounts of insulin but there may be defects in insulin action, or diminished tissue

response to insulin (ADA 2001).

 

' Most people who develop type 1 diabetes are hospitalized for insulin administration within days
of onset.

2 There are other forms of diabetes, such as gestational diabetes and forms of diabetes caused by
malnutrition. This investigation is limited to type 2 diabetes.

Type 2 diabetes has a gradual onset and it is estimated that about 2.4 % of the US
population have type 2 diabetes but have not been diagnosed (Acton 2003). In the early
stages of diabetes, diet together with increased exercise can restore blood glucose
balance. As the disease progresses, oral medications that stimulate insulin production
and/or decrease insulin resistance may be needed in addition to the dietary and exercise
treatment. As the disease progresses insulin resistance increases, the pancreas produces
more insulin and at some point the person with diabetes may need to supplement with
insulin injections. The injected insulin is supplementary to help achieve blood glucose
control. While a non-diabetic adult produces about 20 to 30 units of insulin on any given
day, a person with type 2 diabetes may inject 100 units or more a day in order to

compensate for the body’s inability to utilize the insulin.

Cellular insulin resistance and the resulting free flowing insulin are associated
with an increase and change in lipid levels3 and large blood vessel disease
(macrovascular disease). Macrovascular disease will result in atherosclerosis, which is
the cause of heart disease, clogging of the carotid arteries and/or narrowing of the aorta
that supplies blood to the legs and other body organs. Nearly half of all people with
diabetes are also hypertensive. In people with diabetes who have these multiple factors
for cardiovascular disease the risk for coronary artery disease may be increased 19-fold

(Vinik 2003).

 

3 Lipid levels change in as far as HDL (good cholesterol) decreases and LDL (bad cholesterol)
increases. Thus it is not only the increase in total cholesterol that predisposes to artery disease by also the
composition of cholesterol.

In addition to large blood vessel disease, diabetes causes small blood vessel
disease (microvascular disease), which precedes macrovascular disease. The kidneys and
eyes are the target organs of small blood vessel changes. Damage to the retina may result
in blindness. After 15 years living with diabetes, 2% of patients become blind, and 10%
develop severe visual impairment (V inik, 2003). There is evidence that damage to the
eyes may start years before the diagnosis of diabetes. (Lecture at University of Michigan
Medical School, Department of Endocrinology, Yakko Tuomilehto 2002). Therefore,
treatment standards require that people with type 2 diabetes receive at least one annual

dilated eye examination to detect and treat damages before permanent blindness sets in.

Small blood vessel disease may damage the kidneys as well. Damage to the
kidneys results in leakage of vital proteins into the urine. Although the process can be
slowed with medication, it is estimated that about 40% of all end stage renal disease is
caused by diabetes (National Diabetes Fact Sheet 1998). People with diabetes are the
fastest growing group among dialysis and renal transplant patients. In Michigan in 2000
there were 2,015 hospitalizations among people with diabetes with the primary diagnosis

of kidney disease (Michigan Diabetes Fact Sheet, May 2002).

In addition to developing small and large blood vessel disease, people with
diabetes are prone to infections due to high glucose levels throughout their body fluids.
The glucose provides an abundance of food for pathogens while poor blood circulation
limits the body’s natural defenses against infections. Therefore, people with diabetes can
develop ulcers and pressure sores that invade the feet. While such sores are painful to the
non-diabetic person, diabetes results in nerve damage and the person with diabetes may

not notice the sores as a result of that nerve damage. There are accounts of people with

diabetes who have cut off part of their toe while trying to trim a nail or corn without
feeling any pain. Others have had foot ulcers that invaded the bone without feeling pain.
For that reason, people with diabetes are asked to inspect their feet daily and physicians
are expected to examine and monitor the feeling in the feet of people with diabetes with a
filament at least once a year. Yet, despite the standard, diabetes is the most common
cause of non-traumatic lower extremity amputations. In 2000, in Michigan 1,564
hospital discharges for diabetes documented some form of lower limb amputation

(Diabetes in Michigan Fact Sheet, May 2002).

While people with diabetes are prone to infections, infections such as the
influenza can disturb the glucose balance because the body requires greater amounts of
insulin to fight the infection. In addition the flu can affect the gastro-intestinal tract,
which may result in electrolyte imbalance. Therefore, people with diabetes are asked to

obtain an annual influenza vaccination to avoid the metabolic disturbances caused by the

flu.

As devastating as diabetes is in the more advanced stages, in the early stages of
the disease there are no symptoms that the person with diabetes can feel or detect. The
body adjusts and the brain even demands the higher glucose levels. As a consequence
only 8% of people with diabetes surveyed recognized diabetes as a serious disease
(Gilmet, 1999). Only a person who is knowledgeable about the symptoms of diabetes
can recognize when the glucose metabolism is out of balance or when early symptoms of
complications manifest themselves. However, once small or large blood vessel changes
cause overt symptoms in one or more organs, the whole body has been affected by the

disease and life threatening complications have started to develop.

Prevalence and burden of disease

It has been estimated that only one third of all people with diabetes have been

diagnosed because type 2 diabetes has such vague symptoms in the early stages.

Diabetes is one of the most prevalent and fastest growing chronic diseases in the United
States. The overall prevalence of diabetes rose from 4.9 percent in 1990 to 6.5 percent in
1998. According to the Behavioral Risk Factor Surveillance 2000, 7.1% of the Michigan
population reports that they have been told that they have diabetes (Michigan Department
of Community Health 2000 summary). Type 2 diabetes constitutes up to 95% of all cases
of diabetes (U SDHHS, CDC 2000). Type 2 diabetes prevalence increases with age; i.e.

almost 20% of people over age 65 have been diagnosed with diabetes.

In summary, diabetes is a serious disease that leads to nearly 190,000 deaths a
year in the US. and diabetes is the leading cause of new cases of blindness among
working-age adults, kidney failure, and non-accidental lower extremity amputations.
Diabetes is also very costly. In 1997 in the US. the direct cost for diabetes totaled 44.1
billion dollars, and the highest cost stems from the treatment of diabetic complications
(V inik 2003). Because of the complications, diabetes decreases the quality of life; nearly

50 percent of people with diabetes suffer limitations in their daily activities (CDC, 1999).

Diabetes treatment

As described earlier, diabetes is one of the chronic diseases for which there is no
cure, however, glucose management has proven to improve clinical outcomes (Diabetes
Control and Complications Trial Research Group, 1993; 1998; Gaster, 1998; Nasar,

1999, UK Prospective Study Group, 1998). The Diabetes Control and Complications

Trial (DCCT) was one of the first clinical trials that demonstrated the benefit of tight
glucose control. However, it targeted type 1 diabetes. Because people with type 1
diabetes develop insulin resistance over time it was hypothesized to apply to type 2
diabetes as well. It was estimated that tight glycemic control could decrease the risk for
microvascular complications by thirty percent for those with type 2 diabetes (Diabetes
Complications Control Trial, 1993). Since then research has concentrated on type 2
diabetes and demonstrated the effectiveness of tight glycemic control for preventing
complications (United Kingdom Prospective Diabetes Study Group 1998). Specifically,
for every 1% reduction in glycosylated hemoglobin there is a 22% - 35% reduction in
microvascular complications (V inik 2003; Nasar, et al., 1999) and for every .9%
reduction in hemoglobin Alc over 10 years there is a 16% reduction in heart attacks, a
25% reduction in microvascular disease and a 33% reduction in kidney damage (United
Kingdom Prospective Diabetes Group, 19998). Therefore, the goal of diabetes treatment

is tight glycemic control.

Because of its importance, glycemic control needs to be monitored. People with
diabetes monitor glycemic control at home by measuring blood glucose. Physicians
monitor control with the help of glycosylated hemoglobin Arc, which measures the
percent of glucose that is bound to protein, in this case the hemoglobin. It correlates to
the mean blood glucose level during the past two to three months. The standards of the
American Diabetes Association recommend that the hemoglobin Ale should be measured
twice a year in people with well-controlled diabetes. This is a blood test that may be

done by a laboratory or in a physician’s office.

As mentioned before, most aspects of the diabetes treatment are part of self-care
and the sole responsibility of the person with diabetes; i.e. following a prescribed eating
pattern (Medical Nutrition Therapy [MNT]), exercise, foot inspections, obtaining
influenza vaccination and dilated eye examinations. MNT and exercise are the
foundation of diabetes treatment and are a requirement even in the earliest stages of the
disease. MNT can be adapted to any ethnic needs and lifestyle factors; however, most
people with diabetes need individualized professional help to adapt their eating patterns
to meet their medical, psychological and social needs. Without tailoring, it becomes most

difficult for the person with diabetes to eat appropriately.

As research uncovers more of the causes for insulin resistance, exercise is seen as
ever more important. Exercise has been shown to delay the onset of diabetes and to
decrease or reverse insulin resistance (Tuomilehto, 2001). Therefore, people with
diabetes need to exercise daily. However for people with diabetes, exercise is like
medication and can lower blood glucose levels dangerously, therefore people with
diabetes need information on how to exercise and how to adjust their medication and
food according to exercise duration and intensity. In addition, once nerve damage
(neuropathy) has developed, exercising may be difficult or painful. Exercise generally is
difficult for obese persons and professional counseling or assistance may be needed.
Depending on the type of insurance a referral from their primary provider may be
required. Even if the insurance offers free access to diabetes education without referral,
the person with diabetes needs to be made aware that such services exist and how to
access them. Therefore the ability to practice self-care behaviors depends on the

provider’s information giving and referral to diabetes education. Thus there is

dependence on the provider and his/her staff to enable the person with diabetes to provide

his/her own self-care.

Other self-care behaviors, such as influenza vaccination and retina examination
require teamwork between the provider and the person with diabetes. The person with
diabetes has to seek out his/her provider and the provider has to use the opportunity to

provide preventive services, such as influenza vaccines and eye examinations.

To summarize, type 2 diabetes is a common and serious disease that leads to
many complications. However, in the early stages diabetes has few symptoms and once
symptoms develop, multiple body changes have already occurred. The goal of treatment
is glycemic control, which hinges on diabetes self-care, such as eating habits, exercise,
glucose monitoring, foot examinations and seeking preventive care. Few people with
diabetes can manage the complexity of self-care on their own and therefore depend on

their medical providers for education and counseling.

This study will concentrate on the people with diabetes’ self-care behaviors in
relationship to physician adherence to ADA standards for information giving and
counseling. Before we can consider the impact of physician behavior on the patient’s

self-care we need to explore physician and patient roles in the presence of diabetes.

Diabetes treatment, the §i_c_l_§ £919 and physician behaviors

This section will consider the relationship between the physician and the diabetic
patient and the difference in their current roles compared to the Parsonian model. Along
with the physician role, the paper will consider physician adherence to evidence based
treatment guidelines for diabetes specifically in the area of counseling and information

giving.

Because diabetes treatment includes complex self-care behaviors, the diabetic
patient becomes the physician’s partner. The physician guides and monitors yet the
patient is responsible for the daily treatment. As a result the division of the roles is
shifting; the diabetic patient is expected to become knowledgeable about diabetes and
cannot be dependent on the physician for daily care. This challenges the traditional

models that define physician and patient roles.

The Parsonian model

Parsons developed one of the first models that explore the physician/patient roles.
The Parsonian model was developed in the 19503, and was focused on the fimctional
roles of the physician and the patient. Parsons looked at the physician’s and patient’s role
probably according to the physicians’ own perception. At that time infectious and acute
illnesses were the norm. Thus at Parsons’ time, physician intervention was acute and
required a high degree of technical competence to facilitate the speedy recovery of the
patient (Freeman etal., 1989). The patient was afforded the “sick role” but there were

underlying assumptions and expectations concerning the sick role:

10

o The patient is not responsible for his/her illness;
0 The patient is exempted from the daily tasks and duties;
0 This sickness is undesirable;

o The patient is expected to seek out competent help.

In Parsons’ model the sick person is assumed to be helpless, and to have low
levels of knowledge and technical skills. The patient therefore is dependent on the
physician’s expertise and management of his/her illness. The emphasis is on the
physician as manager and the patient’s following physician orders because the patient
cannot recover or become well on his/her own. In return the patient is excused from

his/her social roles until recovery (Wolinsky 1980).

F riedson and others have criticized Parson’s Model because it does not
differentiate according to the severity of the disease and therefore does not allow for the
non-acute chronic type of diseases from which there is no recovery. Friedson
dichotomized the inability to perform the expected social role into minor and serious
deviation. This dichotomization permits for different reactions according to acute or
chronic conditions. Diabetes without complications would fall into the minor deviation
where the individual is expected to function in his/her role. Likewise Szasz and
Hollender’s work redefined the doctor-patient models around disease severity. They
proposed that patient passivity and physician dominance are most common in acute
disease. Less acute disease is characterized by physician guidance and patient
cooperation. (Szasz and Hollender 1956). Because Parsons did not consider disease

severity or chronic illness, the patient is granted the “sick role” until health is restored.

11

However, in most cases the diabetic person is not granted the sick role but is expected to
fimction in his/her given family/work role while caring for him/herself, changing eating
habits, monitoring his/her blood glucose levels and finding time to exercise and to seek
medical help in managing the illness. It is not unusual for a person with diabetes to
disagree with his/her environment on how much he/she should be expected to work and
how much social support is needed. While this minimized sick role is demanding on the
person with diabetes, at the same time he/she is less dependent than the patient in the
Parsonian model. While the patient in Parson’s model is passive, the modern person with
diabetes is proactive. Due to the chronicity of diabetes, the relationship shifts more
towards the Szasz and Hollender model. The physician guides and assists the patient in

the management of the illness but the patient is responsible and becomes the expert.

Thus, the informed person with diabetes limits the physician’s professional
dominance, which is the foundation of the Parsonian model. During acute stages of
complications the patient may rely on the physician’s guidance; however, the patient is
the expert regarding his/her body’s reaction in different situations. Thus, even during the
acute stages the knowledgeable person with diabetes is a partner to the treating physician.
One could state that today’s person with diabetes is expected to become his/her own
alter-physician. In order to achieve that role, the person with diabetes needs skills and
knowledge about diabetes, his/her body’s symptoms and the complications of diabetes.
The person with diabetes cannot gain those without professional help. In order for the
person with diabetes to succeed in self-management, he/she needs extensive training as
well as support from other professionals. These services need to be provided or

encouraged by the provider.

12

Thus chronic diseases and specifically diabetes, redefine both physician and
patient roles into a complex interdependent relationship. Many physicians have not been
prepared during their training for this shift in the relationship and do not follow the ADA
practice standards that require the physician to train, support, and monitor, rather than to

just “treat”.

To summarize, successful treatment of diabetes requires an active partnership
between the physician and the patient, rather than a dependent patient who assumes
passively the “sick role”. The person with diabetes needs counseling and education either
directly from the medical provider or from support personnel in order to become a
partner. Therefore, the physician’s dominance decreases and the diabetic patient’s
responsibility increases until he/she becomes the physician’s partner in the treatment of
his/her diabetes. The following section will turn from physician and patient roles to

6‘

examine self-care behaviors, which are expected in the patient’s new role”.

13

Theor_'y 3! health behavior that applies tp thjp study: IQ Health Belief

Model

One of the first models developed and used for the last five decades to explain
health and self—care behaviors is the Health Belief Model (HBM). The HBM was
developed in the early 19503; it is one of the value-expectancy models. Originally it was
developed to explain failure to take advantage of preventive health services, such as TB
screening. Although the Health Belief Model was developed for preventive health
actions, it was later expanded to include the more complex illness and sick-role behaviors
(Rosenstock, 1974; Becker 1974; Janz and Becker, 1984). The model holds that people
will take actions to improve or guard their health:

1. If they regard themselves at risk for the illness,

2. Perceived susceptibility

3. If they believe that the illness is serious or has serious consequences,

4. And they believe that they can take actions that will avoid the serious disease or

its consequences (Glanz et a1. 2002).

Perceived susceptibility

This construct refers to a person’s perceived likelihood of contracting the disease
or condition. For this study’s population, perceived susceptibility is no longer of interest.
Participants have already been diagnosed and are aware of the diagnosis (Glanz et al.,
2002). Those who do not recall that they have been told that they have the disease have

been excluded from the study.

14

Perceived severity

Perceived severity refers to the person’s belief that contracting a disease or
leaving it untreated will have serious medical consequences to their health or even life.
In the case of this study population, perceived severity pertains to the belief that lack of
self-care and high blood glucose levels will lead to serious consequences. Medical
science has proven that the probability of developing complications is a function of
glucose levels over time. According to an unpublished study of the American Diabetes
Association, only 8% of people with diabetes believe that diabetes is a serious disease
(Gilmet 1999). Hence one problem is the lack of perception of the severity of the
consequences of lack of glucose control. The difficulty is that there are no symptoms
that the patient can detect in the early stages of diabetes. Therefore, the person with
diabetes depends on medical providers to communicate the severity of diabetes and the
threat from complications; only the most educated people with diabetes will become
aware of the impact of high blood glucose levels without the help of a provider. Even
among people with diabetes who have developed heart disease, many do not realize that
diabetes is the underlying cause of their heart disease. Few people with diabetes realize
that diabetes can cause blindness unless their health care provider made them aware of
the danger. Therefore, perceived severity depends on provider communication and
information giving. One way to consider the severity of diabetes is to look at the
number of complications that have developed due to diabetes, such as heart disease,

stroke, and amputations.

15

Perceived barriers

Perceived barriers are the perceived potential negative effect or psychosocial cost
of taking the needed health actions and to avoid the serious consequences of the disease.
The perceived cost is the physical, psychological, and social cost of self-care. Perceived
cost for diabetes self-care may seem overwhelming to the person with newly diagnosed
diabetes. It includes the cost of dietary compliance, which may require a break with
family eating traditions or limiting the favorite high fat or high sugar food. Without the
help of a dietitian or diabetes educator, few people with diabetes can adjust their eating
habits to meet their own emotional, ethnic, and medical needs for food while balancing

their blood glucose levels.

Exercise poses similar challenges for people with diabetes. Without the help of
an exercise professional few people with diabetes can start an exercise program that fits
into their daily routine and accommodates other physical conditions without experiencing

hypoglycemic reactions.

Indeed, most people with type 2 diabetes are severely overweight and have
difficulty exercising secondary to their excess weight. Other people with diabetes may
have vascular insufficiency or may have developed neuropathy and experience pain while
exercising due to lack of oxygen supply or nerve damage. An exercise physiologist can
assist the person with diabetes to choose an exercise program that is appropriate for

his/her physical condition and disease stage.

Similarly, blood glucose monitoring is painful because the skin needs to be

penetrated to collect blood and the “lancets” (needles) that are used, are relatively thick.

16

People with diabetes in the advanced stages may be asked to check their glucose level up
to seven times a day, which transforms their fingertips into pin-cushion-like appendages.
One needs to consider that some people have a fear of needles; others experience distress
at the sight of their own blood. To puncture that finger again and again requires
motivation and the understanding that the discomfort at the present time lowers the risk
for greater pain or loss of independence at a later time. Likewise, receiving a flu
vaccination can cause anxiety in people who fear complications and illness as result for
the immunization. The dilated eye examination that is needed at least once a year costs
time and is inconvenient. Some people are unable to drive themselves home after the

dilation and thus have to ask for help.

The perceived cost of diabetes self-care behaviors can be lessened by professional
counseling and by tailoring the treatment to the life style of the person with diabetes.
However, few physicians are trained in counseling methods and most become frustrated
with the diabetic patient’s psychosocial barriers. Therefore the physician’s referral and
encouragement to participate in diabetes education is needed (Larme 1998). The ADA
standards call for communication, teaching and support either from the physician or by
some other health professional to whom he/she delegates the responsibility during every
office visit. The cost to the person with diabetes varies with each behavior, while some
diabetes self-care behaviors are difficult or costly behaviors, other behaviors such as foot

care pose less cost to the patient.

17

Perceived benefm

While the perception of the severity of the consequences of the disease can be a
cue to action, the person must also believe that he/she can take action to ameliorate the
consequences of the disease - perceive a benefit. Perceived benefit of self-care is the
belief that self-care can delay, limit, or prevent the complications of diabetes. Unless the
person with diabetes knows and believes that tight glucose control can prevent or delay
complications, the person with diabetes will not be motivated to exercise or to monitor
blood glucose levels. Unless the person with diabetes is aware that early detection can
limit the severity of the complication he/she will not be motivated to monitor his/her feet
or eyes. However, the patient cannot know the importance of self-care and glucose
control in the prevention of complications, unless someone communicates this. It is
the ADA standard that the physician provides counseling for the patient regarding the

importance of glucose control and preventive care during every visit.

Self-efficacy

Self-efficacy was first studied by Bandura (Bandura 1977). It can be defined, as a
person’s belief that he/she can perform the required behavior. Rosenstock, Stretcher, and
Becker held that self-efficacy should be added to the HBM as an additional construct
(Glanz et a1. 2002). The addition of self-efficacy to the HBM is particularly needed when
the model is extended from simple preventive behaviors for which it was developed to
chronic conditions and their complex self-care behaviors, such as diabetes self—care.
Self-efficacy is specific to each behavior and the magnitude of self-efficacy for exercise

can differ from the magnitude of self-efficacy for blood glucose monitoring or foot

18

inspections. Self-efficacy is increased during structured health education activities and
when the appropriate self-care behavior is modeled, such as when the provider checks the

feet or inquires about home blood glucose readings.

Perceived severity of diabetes, benefit, and barriers are all correlated. The greater
the perception of severity, the greater will be the perceived benefit. IF THE PATIENT
DOES NOT PERCEIVE SEVERITY OF THE DISEASE, HE/SHE WILL NOT
PERCEIVE MUCH BENEFIT FROM CONTROLLING IT. In order to be willing to
put up with the cost, there must be a perceived severity and benefit. While the person
with diabetes is most aware of the cost of self-care, the severity of complications and
benefit of self-care must be communicated by the providers. As mentioned earlier,
diabetes has few symptoms in the early stages and tight blood glucose control may feel
worse secondary to hypoglycemic reactions than high blood sugars4. Therefore the
understanding of the person with diabetes regarding the severity of the disease and the
need for self-care hinge on the provider who must communicate the process of the
disease a! teach the components of self-care or refer to other professionals who can
work around the barriers and empower the person with diabetes to take charge of his/her
care. It is not sufficient for the provider to state that self-care is needed, specific
instructions that the person with diabetes can remember and apply as well as tailoring to

meet psychosocial needs are required.

 

‘ Hypoglycemic have similar symptoms as anxiety attacks.

19

00.3.0000... 2... 2 5.3 200.. 8.. m. 02.00 8 0:0 4.2.0.960 0.0.9.80 0 2 2... 0.53
a. 002. .8. 05.0 as. 03.0.0 .0... 3.08 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

30080.28 0000.0... E0200 000.00...
.0... 0.235.. 02.00.. .0 02050500.. 00 :08 0.. 8.0 :00 02.02888 .00.. 2... 0.2.0200 02.00.. 0.0500055
we. 008:3 coo... 0... .0 02.00.05 ...:o..Eoo2P. 00.0.0800 0... 2.0 00.000: .0 2.0502220. .0...0> 0 >. O
.00.. 0... 0.08.05 8 2.30.0 302.200 00 :03 00 :20 .0258. 02.000. 9.300. .
03003 .52.. 6008.003 0.0 .0... 0.2% 92203 52...». 8:80.00. 2.00... mica .00» 020030.... 003 02.00 00.0. 8 2.30
0000200.. .0... 00.0.8.0 :0 2.... 2 02.05... .922. 0... ..o 03.0.32... 00.300. .. 02.00 00.5.2.2. 5 00:02:» £220.. 023.... 0.0.... 2 00:02.28 0.000 >80E9..0m
02.00 o. 000 0.0 :09. 0.. :8 002.0250 .00..
2.0 002 00.803 .82.. no wEo....oE .025... 0.00....8 0.0.2.0.... 2 02.0.0000
2.3 0.0020020... 0.2.9.0 o. .2000. 02.50. 00.2.0... 5.... 0.000.. .83. 22> 00.0.9.0 00250. 3.00.0 ..mmo.._..00.: :280
02...... .096 92...... 00.23.... 0.. 302.0 02.0.52... 00.00080 <9... 8 m20080< 5000.030 08:8... 00005.2... 2.32. 0.2.6.... 0.0.8 0.02.00 9 00.9.0.6 8 0000
0.000.. 3082.0 2.0 .020 .. :0... 9 00.2.0. 00.0.0.8 20... 9.0.. o. .000: 02.0.2000 92.000... 02.00 002.60 0... ..o 0.08
082.0... 5.3 0.000.. 0.2.5.0.. 200 b0>0 .009... 2.0 gm... b0.» 0.0 92.8.00... 62.00.00.225 .0 02.02.50 00:00:02.0. 30322.9.3 2.0 0....qu 20.50..
008...» woo... 2.0 00.0.9.0 b.0325... .200 00.000... .0 .80 002080.. 2.... .302... 0.0.0.0.. 002020.. 00.60. 2.0 bacon. 0... .00.... 02.0.. 0.0.5 002020..
0.00.0.8 00.2.0... .0 9.00:2. 0... .030 . =0. 0.. .00... 08000... 5.3 .009... .0 000000200
0.000.. 02.02.0800 0... 0522.0 .0 £05.. 2. 0.. :00 0.0... bt0>0m 32...? .o 20... 00:00.. 8 02.00
600.0. 0.0 5000.. 2.0 0.0... .0 «5.05502... .0530... 22.02358 0205.. 02009.0 0.. 8 300.0 02200.. 0... 002.60 0... ..o .3850 9:05..
8 .20.... 000. 0.0 00.2.0... .23 0.000.. .005... .00.. 20...». 0.2.0. 88:.m 0. .90... 2.0....» .20....» .32. ”00.0. 8 02.00 05.09 0... 2 00:2. 0.0.5 002000..
0.... .0090
22.02.0500 2.2.0:. 2.0 0.05. 00.82» am... 0.. 2.0 02.60.00 0 002.00. .2050
000.50.. 20002.28 0... ..o 0.030 0.. .00... 00.30... 5.3 08.0.. 0... 00.00 0.... :. 0:220:00 0... 2.0 00.. 0... ..o 002.00.00.80 20.00% 32. ..o 02.0.. 0.0.5 00200.0“.
20.... .0500 0.00220... =0
5.3 .082008 0.2.. 2.5.3083 002020.. 8.02
.2>0..0.. .0 0222.00
00.30.... 0...... x0... .0... >85. 3:022:00 :0 0.300.. 02.0....08005 0.08.00 0 ..o 0000.. 20... 02.2.8.0.— .. wagon ..o 0000.... 2.3.2.083
.830 0.... S 0.5 02.0.2.2. .330 2... o. 3...... .2. 02. 23.30000... 00:020.. 0.0>0. 0.2.. 2.0 0.2.. .0 02.0.32. 05.0.“. 0... 920.03. 02.0.. 0.000 00200.0“.

 

 

 

0.0.00.2 mozom 5.00... 0... mo 02.30.. 2.0 $300.50 >3. ”.Ifllflwm

20

Diabetes self-care behaviors m provider practice

Previous sections have explored the disease process, patient and physician roles
and behavioral issues. The following section will explore physician adherence to
evidence based treatment standards for diabetes, specifically information giving and

encouraging tight glycemic control.

Physician adherence to diabetes treatment standards

The treatment standards for diabetes were the first among the chronic conditions
to be developed. Unlike other practice standards, there is little disparity among the
standards of different disciplines. The American Diabetes Association (ADA) convenes
expert panels from various disciplines that define and update the standards for medical
management. The American Academy of Internal Medicine, the American Academy of
Family Practitioners, as well as many other professional organizations accept and adopt
the standards from the ADA and expect their professional members to practice

accordingly.

The treatment standards have been changed and updated on a yearly basis as
research provided insight into diabetes prevention, the disease process and the benefit of
tight glucose control. The first benchmarks came from the Diabetes Complications and
Control Trial (DCCT Writing Group, 1993) and the United Kingdom Diabetes Study
(N asar, 1998). Both studies demonstrated that tight glucose control decreases the
probability of developing micro and macro vascular disease. Until then medical

practitioners accepted blood glucose levels of about 200 mg/dl for people with diabetes

21

because people with diabetes “feel good” and no acute episodes were observed at those
blood glucose levels. Since the DCCT and UK studies, blood glucose levels are expected
to stay near normal, less than 140 mg/dl at all times in order to limit long-term
complications. However, when persons with diabetes keep their blood glucose levels
low, the probability of hypoglycemic reactions increases. People with diabetes who have
had high blood glucose levels and bring them into the normal ranges feel as if they are
having hypoglycemic reactions until the brain readjusts to normal levels. This may take a
few weeks. Thus the person with diabetes needs extensive cognitive skills as well as
professional support to maintain tight glucose control; i.e. the participants of the DCCT
study were supported by the services of a psychologist, a teaching nurse, a dietitian and
an exercise therapist for the almost ten year duration of the study. Most medical
practitioners cannot provide such extensive professional resources and much of the
responsibility to teach and support falls on the physician and his/her staff, who are
expected to keep up with the changing practice standards as well as motivational
techniques. This is particularly demanding on the general practitioner who has to keep
up with many conditions and prevention standards, not just diabetes. Therefore it is not
surprising that diabetes care falls short of the treatment standards, particularly in the areas

of counseling and information giving.

Part of the issue may be that physicians in general do not trust changing clinical
practice standards nor the concept of evidence based standards in general. Both of the
terms are relatively new in medical care and may be viewed as “cook book medicine”.
Mottur-Pilson (2001) found that physicians point to different reasons for failing to adhere

to standards. Some physicians cite lack of awareness or distrust towards the standards as

22

reasons; others make a professional decision that the patient does not benefit from the
service because of his/her age or comorbid conditions. In a survey sponsored by The
Quality Indicator, Physician Resource (February 2002) up to 50% of physicians reported
relying on their “professional experience” in treating patients rather than on clinical
treatment standards. However, the professional experience is not a controlled trial that
either adjusts for the different variables or compares treatment outcomes under like
conditions. Therefore, using professional experience is not a scientific or evidence based

approach to treatment.

Larme (1998) found that Primary Care Providers (PCPs) were frustrated with the
treatment of diabetes. They ranked diabetes as disease more difficult to treat than
hypertension, hyperlipidemia, angina and arthritis. They reported that the disease process
is complex and medications are difficult to regulate because of the effect of food and
exercise, the patient’s self-care behaviors. Their responses recalled “horrible struggles”
with patients over diet and life style changes. Physicians lamented that they had received
insufficient training in medical school and residencies to promote lifestyle changes. They
also cited lack of support from the health care system as part of the problem in treating

diabetes. Thus there are several reasons that can explain the lack of diabetes counseling.

In the previous sections we have explored the disease process of diabetes, its
treatment, the physician and patient roles in diabetes, behavioral theories, and physician
adherence to diabetes treatment standards. Next the study will turn the attention to
patient behaviors, particularly self-care behaviors required for glucose control in the

presence of diabetes.

23

Self-care behaviors in diabetes

This section will explore self-care behaviors and their complexity as required for
diabetes. As self-care behaviors become more complex, adherence becomes more
challenging. Diabetes self-care behaviors may be the most complex among the chronic
conditions, both from a behavioral point of view as well as from a cognitive point of

view.

Levin, Katz and Holst (1976) defined self-care “as a process by which the lay
person functions on his/her own behalf to promote health, to prevent illness, and to detect
and treat disease when it occurs.” The concept of self-care spans a variety of behaviors
that range from nutrition and exercise habits, to symptom diagnosis, self-medication, and
home remedies, as well as seeking professional assistance and the manner of interaction
with the “professional sector” (Seagal and Goldstein, 1989). All of these behaviors

together are considered self-care for diabetes.

Dean (1989) organized self-care behaviors into the following categories: a)
routine daily habits of living that affect health, such as smoking, drinking, sedentary
behavior, etc; b) consciously sought health maintenance behaviors such as exercise and
blood glucose monitoring6 c) behavioral responses to symptoms of illness such as seeking
professional assistance for insulin adjustment, etc. All of these self-care categories

impact glucose control, which determines the extent of the complications.

 

6The difference between #1 and #2: #1 consists of behaviors of the daily routine; #2 are
consciously sought behaviors that effect health.

24

Thus, diabetes requires complex self-care behaviors; it demands an understanding

of the body’s glucose and lipid metabolism, the disease process, and diabetes

complications as well as the recognition of their symptoms. In addition to these cognitive

skills, self-care requires difficult behavioral changes that may conflict with the self-image

and lifestyle of the person with diabetes.

(Anderson et a1. 1997, Lupton 1994, Larrne 1998). People with diabetes need:

Regular exercise to reduce insulin resistance, correct the lipid metabolism, and to
enhance the circulatory system;

A diet that keeps glucose levels balanced and lipid levels low;

Home blood glucose monitoring to evaluate the outcome of self-management;
Daily foot inspections to discover and treat infections;

An annual influenza vaccine to prevent the complications of the flu;

An annual dilated eye examination to detect and/or treat diabetic eye damage.

Obtaining an annual influenza vaccination and retina examination are office based

self-care behaviors. One could argue that they are under the control of the physician,

however, while the physician can provide the services, the patient can obtain dilated eye

examination from an optometrist and the influenza vaccination from community centers

or even in the supermarket. No referral or prescription is needed. Thus, while the

physician can provide the services for the patient, the patient also can take initiative and

act on his/her behalf. Therefore the two behaviors will be included with other self—care

behaviors.

The timing of exercise, the monitoring of blood glucose and the time required to

care for feet and eyes impact on every phase of the life of the person with diabetes’ as

well as on that of his/her family. Yet, self-care is an integral part of treatment and the

25

key to avoiding complications. Because of the complexity of self-care, most people with
diabetes require counseling and education, such as exercise counseling and diabetes
education for blood glucose monitoring. Counseling and education in turn increase self-

efficacy (Sackett 1979; Becker 1975; Lorig 2001).

Considering the complexity of diabetes self-care and considering that this self-
care is needed for the duration of the life-span, it is not surprising that many people with
diabetes do not adhere to self-care standards as expected. According to Lutfey (1999)
and Kurtz (1990) only one third to three fourth of people with diabetes adhere to self-care

guidelines.

The next section will explore the relationship between diabetes, self-care or the

lack of it, and the physician’s information giving according to treatment standards.
Non-adherence and provider practice in diabetes

Non-adherence in patients has been studied since the beginning of the 20th
century. Both Haynes and Sackett et al. have listed almost 2,000 references on non-
compliance research (Phillips 1988). The terminology for non-adherence has changed
over the years. In the early part of the century, the patient was considered to be
“defaulting” (Wright 1982). From “defaulting” evolved the concept of “non-compliance”
in the 19708 (Becker 1972) and in the 1990s the same behavior was defined as “non-
adherence”. Studies of non-adherence and non-compliance have focused on the patients’
behaviors of taking prescribed medication or appointment keeping (Phillips 1988). Those
two behaviors are relatively simple, not as complex as exercise and home blood glucose

monitoring. Adherence to diabetes self-care behaviors may be linked to patient/physician

26

communication as well as other physician behaviors, such as information giving and
encouragement to practice diabetes self-care behaviors. Therefore, the physician’s
communication and information giving must be considered when we study self-care

behaviors in people with diabetes.

Non-verbal as well as verbal physician communication has been shown to affect
compliance. Terry et al. (2000) found that physician advice contributed to overall
satisfaction with care, which has been found to correlate to adherence (Pumilia 2002;
Clemenhagen 1994). Roter (1994), Phillips (1988), Beck et al. (2002), and Moci (2001)
found that physician support for the patient’s questions, and providing health specific
information increased patient adherence. A meta analysis showed that advice by primary
care physicians to quit smoking increased the odds ratio for quitting to 1.69 (Silagy,
Stead LF 2002). The American Lung Association study (1999) found that patients who
have been advised by their physician to stop smoking are 15 to 30% more likely to stop
smoking. The Brown University Digest (2002) reports that problem drinkers benefit
from brief physician advice to limit drinking. Fleming et al. (2002) found that problem
drinkers who received information and counseling from their physicians were likely to
reduce their drinking both at six month and four year measurement points. The Writing
Group for the Activity Counseling Trial (2001) reported success for all patients who
received physician advice to exercise. Beresford et al. (1997) found that low intensity
interventions, such as counseling by primary care physicians resulted in changes in eating
patterns such as fat and dietary fiber consumption. Taveras et al. (2003) found that
physician advice and support for breastfeeding increased the number of mothers who

breastfed. The duration of success after counseling varies according to behaviors and

27

setting from four months to four years (The Writing Group for Activity Counseling 2001 ,
Silgagy 2002, Fleming 2002, Taveras 2003). The success of physician counseling or
advice may also depend on the physician’s own modeling (Hugins 2003) of the health

behavior.

If the physician does not counsel during the regular medical visits, then the person
with diabetes needs the assistance of other health professionals who can provide the
behavioral interventions. However, that requires that the physician informs the person
with diabetes about diabetes education opportunities and encourages the person with
diabetes to seek out such assistance. Working with a diabetes educator or a dietitian can
lower the perceived cost for self-care by tailoring the self-care behaviors to the schedules,
ethnic and family norms of the person with diabetes. Furthermore, such health
professionals can reinforce the physician’s communication regarding the threat of

complications and the benefit of tight glucose control.

McGlynn (2003) et al. found in a national study that physicians provide 56.1% of
recommended care to people with chronic conditions in general, 45.4% of recommended
diabetes treatment and 18.3% of recommended counseling or education. Conrad (1994)
found that only 48.9% of insulin treated type 2 people with diabetes and 23.7% of non-
insulin treated type 2 people with diabetes had any counseling or diabetes education at
any time during their disease process. He found that the average length since the last
diabetes education was reported to be 4.5 years. People with type 2 diabetes who do not

use insulin were least likely to report education or counseling.

28

Considering that physicians provide only 18.3% of the counseling and about one
quarter of all people with diabetes receive formal diabetes education, it is not surprising
that patients with diabetes do not adhere to self-care recommendations; they may not
even be aware of the importance of self-care (Lutfey 1999; Kurtz, 1990). Once patients
have received diabetes education or participated in self-help groups they are more likely

to adhere to self-care guidelines (Deyo 2001, Lorig 1999).

In summary, there is evidence that physicians do not adhere to diabetes treatment
standards in general and to counseling standards in particular. As a result people with
diabetes may not have the cognitive skills and behavioral capacity to practice the desired

self-care.

Because the difficulties with diabetes treatment and the lack of adherence to the
diabetes clinical practice guidelines is known, accrediting and watchdog agencies require
monitoring and evaluation of diabetes treatment, such as the National Committee for
Quality Assurance (NCQA) and Health Employer Data Information System (HEDIS)
measures. However, their measures have limitations. At the present time neither NCQA
nor HEDIS require monitoring medical providers’ information giving for diabetes or
referrals for dietary counseling or self-care classes (Von Korff 1997). The selection and
choice of evaluation criteria encourages or even forces medical care providers to pay
attention to that which is perceived to be important. Physician adherence to guidelines
appears lower for behavioral or patient focused practices compared to physiological
practice components (Glasgow 1999). As the next section will show, all diabetes care

indicators are lower than one would expect. The data that are explored in the following

29

sections are taken from Healthy People 2010, Diabetes in Michigan, HEDIS measures,

and M-CARE’s Diabetes Health Status Survey Questionnaire.

Outcome measures for diabetes treatment

One of the data sources that look at diabetes treatment is Healthy People 2010.
Healthy People 2010 reports national data from the 1998 Behavioral Risk Factor
Surveillance System by the Centers for Disease Control. The information is self-reported
and was collected from a stratified random sample of Americans, some of whom have no
health insurance; others have a variety of different insurance coverage. Hence the
diabetes data are different from other sources that gather data only from people with

diabetes who are covered by medical care insurance.

Local treatment data comes from the Michigan Department of Community
Health. Their Diabetes Fact Sheet combines information from the Michigan Behavioral
Risk Factor Surveillance Survey (self-reported data) as well as claims data from
Michigan’s diabetic population. The data is specific to Michigan and includes people
with a variety of insurance plans, no health insurance, as well as those who are covered

by managed care.

The Health Employer Data Information Set (HEDIS) reports data gathered by
Managed Care Organizations (MCOs). HEDIS reports contain data fi'om claims as well
as information from chart audits. HEDIS requires several measures for diabetes care.
The national HEDIS data for 1998 come from over 70 million people who receive
medical care through managed care organizations. Sampling and data collection methods

are strictly defined and adherence is verified by audits.

30

The most specific data source for this paper is the M-CARE Diabetes Health
Status Questionnaire. It is a survey that is mailed to a random sample of M-CARE’s
diabetic population. One of its purposes is to collect M-CARE’s data for the HEDIS
report. Hence one could consider the data a subset of the national HEDIS data. The data

is also self-reported.

The following table lists the different data sources, the diabetic treatment
measures and the percent of people with diabetes who received the treatment and the data
source. Each column lists the percentage of surveyed people who received the services.
Although the American Diabetes Association Standards call for a minimum of two
hemoglobin A1c measures per year, HEDIS and NCQA for whom most of the measures

were collected, require only one An measure per year.

Table l: Adherence to practice guidelines according to various data sources

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[Trgatment' ‘ J. I Healthy People Diabetes-in H HEDIS " TIM—CARE Health. ”
" i I .2910 . j , “Michigan 4; . Stags Survey"

Patient based self-care

Diabetes Self-Care 40% NM‘ 24% 24%

Education

Home Blood Glucose 42% NM NM NM
Provider based self-care

Foot Examination 55% 64% 55% 39%

One hemoglobin A], 24% 21% 80% 86%

measure per year

Self-care that can be initiated by either provider or patient
Influenza Vaccination N/M* 46% NM‘ 73%
Eye Examination 56% 71% 50% 68%

 

 

 

 

 

 

 

N/M= not measured or not reported

31

According to the data collected during 1998 for Healthy People 2010, as well as
the HEDIS data and the data from the Michigan Behavioral Risk Surveillance Survey
data, people with diabetes are not receiving self-care education, hemoglobin Ale
monitoring, foot examinations, influenza vaccinations or dilated eye examination
according to ADA standards; all measures could show 90 % compliance. The ADA
standards call for ongoing dietary counseling, a minimum of two hemoglobin Alc tests
per year, (not one as reported in Table l) and an annual eye examination, foot
examination, influenza vaccination as well as home blood glucose monitoring for the

majority of people with diabetes.

Some may argue that physicians may have ordered diabetes education and the
patient may not have complied. However, people with diabetes see their physicians at
least twice a year, therefore the physician can insist on compliance before the next visit or
before he/she re-orders diabetes medications. Hence, even though the person with
diabetes has to cooperate, the physician can exert control. The hemoglobin Alc test is
totally under the physician’s control; it does not require fasting and can be performed in
the physician’s office or at least blood can be drawn in the office and sent to the lab. The
annual dilated eye examination requires the cooperation of the person with diabetes to
keep the appointment for the eye examination. However, the physician can exercise
control, mail reminders or call to remind the patient with similar methods that are used by
dentists and veterinarians to remind their clients. The foot examination is performed by
the physician or the nurse and is the physician’s sole responsibility. The influenza
vaccination is under the physician’s control. It can be given between October and

February. Whenever the patient sees his/her doctor at least two times a year, the

32

physician has the opportunity to vaccinate. Home blood glucose monitoring requires the
cooperation of the person with diabetes, as well as training and reinforcement from the

physician.

Therefore when one examines the lack of expected diabetes self-care one has to
consider the lack of physician adherence to clinical practice guidelines and
communication with the diabetic person. This relationship between lack of physician
information giving and self—care behaviors is not clear and the magnitude of that

relationship is not known.

33

Unresolved Questions concerning provider’s practice and self-care

While there is growing awareness in the literature that many physicians do not
practice according to evidence based standards (Simon et al. 2003), little is known about

the impact of physician non-adherence on diabetes self-care behaviors.

Because of the importance of self-care and the common lack of it, much research
has focused on diabetes self-care. However, most research has studied the patients, their
attitudes and beliefs without controlling the provider’s information giving or adherence to
ADA practice standards and the impact on the people with diabetes’ self-care (Von Korff
1997, Heissler 2002). Some studies have focused on physician adherence, others on
diabetes self-care behaviors. The link between provider adherence to standards and the
diabetic person’s adherence to self-care behaviors is poorly understood and needs to be
explored. Therefore, this study aims to explore the concept of self-care in the presence of
diabetes and its relationship to physician communication and information giving

according to treatment standards.

Furthermore, existing research regarding the self-management of diabetes has
focused on patients that are treated in specialized diabetes clinics, which offer a range of
support services and may have diabetes educators among their staff. It is reasonable to
assume that patients whose diabetes is treated by specialists are in the more advanced
stages of diabetes and at that time they have become aware of the severity of diabetes
complications. As has been shown in the context of the HBM, disease severity of
diabetes is important because in the early stages there are no symptoms and the presence

of complications affects perceived severity of diabetes. The cost of blood glucose

34

monitoring and exercise may appear to be less once the person with diabetes is faced with
the increasing threat of loosing a foot, leg, or eyesight. While general practitioners are
expected to keep abreast of most conditions, new developments, and treatment standards,
specialists can concentrate on the treatment of diabetes and its complications. Therefore,
specialist and staff become expert in treatment expectations and disease process and
information giving regarding diabetes. In the US, primary care physicians provide more
than 90% of diabetes care (Heissler 2002, Glasgow 1999). While primary care
physicians are quite capable of caring for people with diabetes in the earlier stages of
diabetes, there is little knowledge regarding primary care provider adherence to
standards, information giving and the impact on diabetes self-care (Koltun 1986). There
is reason the suspect that there is a difference in communication from specialists and their
specialized support staff compared to the primary care physician and his/her staff (
Zgibor, et al. 2002). What is true for the specialist may not be true for the primary care
physician. This study aims to explore the diabetic person’s recall of communications
from their primary provider and their supporting staff and the impact on self-care

behaviors.
Patient recall as a measure of physician behavior

This study concentrates on the self-report of self-care behaviors and provider
information giving as recalled by the patient. One could argue that recall is not an
accurate measure of provider information giving. Previous studies have shown that
patient recall of information given by medical providers varies greatly. For example
recall of medication instruction was as high as 89.6% when instructions were reinforced

in writing or as low as 26.6% from verbal provider instruction (Crichton 1978). More

35

extensive teaching, such as pre-operative teaching may be of similar interest to the patient
as diabetes self-care education. Cheng (2002) found recall as high as 96% among
surgical patients. In our particular situation (recalling having received diabetes self-care
information) it is suspected that the level of recall should be relatively high. Patients are
being asked if they received some information - not if they remember anything specific.
In addition, if physicians did provide the information and patients do not recall receiving
it, then providers need to devise methods of conveying such information that are better
retained by the patient. The person with diabetes should have a minimum of two
physician visits per year; most people with diabetes average three visits per year (The
writing group for activity counseling in Primary Care, 2001). The standards require
counseling during each of the visits. Thus the patient should receive self-care
information several times per year and be able to remember. Furthermore, if the person
with diabetes cannot remember receiving the information then it is impossible for him/her
to act upon the information and practice appropriate self-care; this study explores the

effect of information on self-care.

36

HYPOTHESES

The following section contains the hypothesis statements of the dissertation.
While there are several different hypotheses and some of them may be related, each

hypothesis is assuming that other factors are equal.

Provider advice and self-care behavior

In a large California study of people with diabetes enrolled in a managed care
organization, members who had seen a nutritionist or were part of a diabetes support
group were more likely to also have received retina examinations, foot examination and
glycosylated hemoglobin testing (Lorig, 1999). In other words, it appears that those who
receive diabetes self-care teaching also receive medical care according to treatment

standards from their physicians in other areas.

Likewise, Terry et al. (2000) Pumilia (2002), Clemenhagen (1994), and Silagy
(2002) found that physician advice to quit tobacco use improved the odds ratio for smoke
cessation. Fleming (2002) reported that drinkers reduced drinking after physician advice.
The Writing Group for the Activity Counseling Trial (2001) reported that physician
advice increased physical activity. Smoking and alcohol consumption are addictive
behaviors and it is not clear how they relate to diabetes self-care behaviors; exercising
however, is one of the diabetes self-care behaviors and the study demonstrated that it was
increased by physician advice. According to the Health Belief Model, the provider’s
advice should increase the perceived benefit of the self-care behavior and give a cue to

action.

37

Hypothesis 1:

Hypothesis 1.1: People with diabetes who recall their provider’s advice to
engage in specific diabetes self-care activities (blood glucose testing,
exercise, foot examination) are more likely to report practicing that self-care

behavior.

Hypothesis 1.2: Those who receive more self-care advice from their
provider are more likely to engage in office based self-care behaviors

(influenza vaccination and retina examination).

The second hypothesis is stated because there are such situations where we
do not know whether the provider gave specific advice on that behavior and a
general information giving index measure will be used as a predictor. As
mentioned previously, the two office based self-care behaviors are under the
provider’s control, but the person with diabetes can secure these services

independently without provider assistance.

Self-care a_nr_i_ provider modeling

According to the HBM, the person with diabetes must believe that he/she can

perform the required self-care activity. This belief is also known as self-efficacy.

Home blood glucose monitoring is the foundation of diabetes self-monitoring and
the only information that enables the person with diabetes to intervene as soon as blood
glucose levels rise or fall. According to American Diabetes Association standards,

(2000, 2001) home blood glucose monitoring is recommended for most people with

38

diabetes along with a recording system of blood glucose levels. By examining the home
blood glucose record, questioning the results, and suggesting adjustments for high or low
blood glucose readings, the provider models and demonstrates how to make decisions
based on blood glucose levels. When this occurs regularly during visits, the person with
diabetes learns to follow the provider’s example and how to interpret blood glucose
readings and how to adjust medication, diet and exercise. As the person with diabetes
sees the success of his/her decisions or gets help from the provider, the person with
diabetes learns that he/she is capable of controlling his/her blood glucose levels. Thus,
self-efficacy for controlling blood glucose levels increases. Self-efficacy has been found
to be one of the better predictors for self-care behaviors (Lorig 2001, Plotinikoff 2000,
Johnson 1997, Skelly 1995). Provider review of the glucose log is one of the American

Diabetes Association practice standards.

Similarly, professional foot examinations should take place at least yearly for
people with diabetes during the annual physical, more often for those who have
difficulties with glucose control or their feet. Nationally 55% of people with diabetes
recall having an annual foot exam. 39% of M-CARE’s diabetic members recalled having
had a foot examination from their physician within the last twelve months (See Table 1).
In addition to provider examination, people with diabetes need to monitor the condition
of their feet themselves daily. As mentioned before, due to the prevalence of nerve
damage, the person with diabetes may not feel pain in the presence of injury or infection.
It is theorized that when the physician examines the feet and tests for nerve damage, the
person with diabetes learns that his/her feeling may not be reliable; that in turn increases

the perceived susceptibility (HBM). He/she learns to follow the provider’s modeling and

39

to inspect his/her own feet for signs of injury or infection; that in turn increases self-
efficacy (HBM). Thus, provider modeling of self-care behaviors, or provider
demonstration of self-care such as reviewing the home blood glucose log and examining
the feet, will increase the self-efficacy of the person with diabetes for those two

behaviors.

In addition to increasing self-efficacy for those two behaviors, the providers’
modeling of the behavior communicates to the person with diabetes that the behaviors are
important in preventing the complications of diabetes. The monitoring of the home blood
glucose log communicates that glucose logs and glycemic control are important for
preventing the complications of diabetes. The inspection of the feet sends the message
that early detection of infections or injury can prevent further complications. Simon
(1999) found in California that those who received foot exams were also more likely to
receive other diabetes care, such as retina exams. Hugins (2003) found that physician

modeling of self-care behaviors appears to effect self-care behaviors.

Hypothesis 2: When the provider models self-care behaviors the patient is more likely to

practice self-care behaviors.

Hypothesis 2.1 :When the person with diabetes recalls that the provider examined
the home blood glucose log; the patient is more likely to monitor his/her home
blood glucose levels.

Hypothesis 2.2: When the person with diabetes recalls that the provider
examined his/her feet, the person with diabetes is more likely to report

examining his/her own feet at home.

40

Self-care 1151 provider referral 2M information giving

Discussing self-care and encouragement to use diabetes education are part of the
treatment standards for diabetes (American Diabetes Association 2000, 2001). Larme
(1998) found that physicians report that they received little training in medical school and
during the clinical years regarding counseling for self-care behaviors and assistance with
life-style changes. Therefore, many physicians feel frustrated and unable to change the
diabetic person’s behavior. Glasgow (1999, 2000) and McGlynn (2003) found that
performance measures for support for self-management and counseling are lower than
those that monitor biological markers. Because of the difficulties in the primary care
setting and the lack of skill and time required for counseling, people with diabetes need

counseling from specialized support providers, such as dietitians or diabetes educators.

Hypothesis 3:

Hypothesis 3.1: People with diabetes who recall receiving information

 

regarding diabetes education are more likely to take advantage of diabetes
education.
Hypothesis 3.2: Those who receive diabetes education are more likely to

practice appropriate self-care behaviors.

41

Figure 2: Proposed relationship between patient characteristics, physician practice and

self-care behaviors

 

Patient Characteristics:
Age
Education

 

Ethnicity

Gender

Income

Complications

Years since diagnosis with diabetes

1

Provider Behavior Variables
> Adherence to clinical practice guidelines by providing
patient self-care education
> Modeling self-care behaviors

VVVVVVV

 

 

 

 

 

 

 

 

Patient Self Reported Self-care Activities
Exercise

Home blood glucose

Foot examinations

Influenza vaccination 4—
Eye examination

VVVVV

 

 

 

42

 

Severity 91' diabetes in; physician practice

The severity of the disease, and the progression of diabetes complications may
influence physician referral to specialty and support services. Once complications set in,
physicians refer the person with diabetes to medical specialists, self-care education etc.
One measure for the severity of diabetes is time since diabetes diagnosis, and presence of
complications, such as cardiovascular conditions and neuropathy. Complications develop
according to glycemic control and time lived with diabetes (Comrad 1994, Anderson

1994).

Also, seen from the Health Belief Model, once complications become
symptomatic the person with diabetes is more likely to recognize the severity of diabetes,
i.e. once a person with diabetes has had the first heart attack or bypass surgery, he/she is
more likely to consider the seriousness and try to maintain the appropriate self-care
behaviors. However, once a person has had a stroke or an amputation, exercise may be
limited. Therefore disease severity has to be considered when we study the relationship

between treatment, and diabetes self-care behaviors.

The issue is this: Self-care in the early stages can delay or prevent complications,
but referrals and encouragement for self-care appear to be offered in the later stages of
diabetes when the disease process limits the person’s independence secondary to the

complications of diabetes.

43

Hypothesis 4:

Those who are in the more advanced stages of diabetes are more likely to

report using the services of a specialist.

Demographics, provider information giving 113! self-care

Dean (1989) found socioeconomic status, including education was found to be the
greatest predictor for self-care. David Williams (1990) found low socioeconomic status
as predictor for high body weight and Bucher and Ragland (1995) found a relationship
between high serum cholesterol levels and low education. Similarly, Sorlie et al. (1995)

found that a higher education level is associated with lower mortality.

Provider information giving and encouragement for self-management of diabetes
may vary according to patients’ economic status, ethnic background, gender, or education
level (Roter & Hall 1992, 2001, Wiggers 1997, Scott 1996, Kaplan 1989, Waitzkin
1991). There is growing evidence that health outcomes and technical medical care vary
by education, age and ethnic group (Anderson 1995, Sorlie 1995, Scott 1996, Callahan
2000, Roter 2002). Thus Lipton (1998) found that physicians frnd it more difficult to
communicate with patients with low reading levels or low language proficiency. When
physicians have difficulty communicating, then patients may not understand and be able
to be motivated to follow self-care instructions (Roter & Hall 1992, Boyer 1996,

Anderson 1997, Wagner 2001).

44

Education influences diabetes control in other ways; education is correlated to
income. The better-educated person generally earns more than the less educated person
does. Those who have greater financial resources have the ability to buy foods that are
helpful in the diabetic diet, such as fish and fresh vegetables (Sorlie et al. 1995).
Furthermore, the better-educated person has an easier time accessing medical care as well

as obtaining needed information in the library or on the Internet.

There is no difference in diabetes prevalence between males and females once age
is controlled for. However, the diabetic life style may affect men and women differently.
Women tend to be more willing to diet because the American ideal is a thin woman (Dan,
1994) and it is the cultural expectation. Therefore, women in Western societies are more
likely to severely restrict their food intake, which should make it easier to adhere to the
diabetic medical nutrition plan. Men on the other hand tend to engage in more vigorous
physical exercise, which is one of the diabetes related health behaviors. There are also
gender differences in self-treatment and men are less likely than women to utilize
medical preventive care (Waldron, 1988). Thus health behaviors are culturally dependent

and gender based, and we will consider the gender of the subjects.

Age is important because the prevalence of type 2 diabetes increases with age.
Also as mentioned earlier, small and large blood vessel diseases take time to manifest
themselves. Therefore, the longer the person lives with diabetes, the more complications
tend to develop and age and time lived with diabetes are correlated. According to the
Health Belief Model, once the person with diabetes becomes more aware of symptoms,

the perceived severity of diabetes increases. In addition, it is common to treat the 80-

45

diabetes and there is evidence that providers may communicate differently with those

who are very old (Callahan 2000, Hall 1994).

Hypothesis 5:

Hypothesis 5.]: Those who are better educated are more likely to receive
information on diabetes self-care from their providers.

Hypothesis 5.2 : Those who are better educated are more likely to practice
appropriate diabetes self-care behaviors when we control for information
giving.

Hypothesis 5.3: Those who are older are less likely to be counseled on, or

referred to diabetes self-care education.

46

METHODS

This study focuses on care provided within the setting of M-CARE, a managed
care organization. All services required for medical treatment are a covered benefit for
diabetic members, although members have varying amounts of co-pays for the services.
Therefore, the study may not apply to the general population who has little or no
coverage for services such as physician office visits, counseling or diabetic supplies, i.e.
glucose monitoring strips. The study includes M-CARE’s Medicare population, which
has limited prescription coverage, but this study does not examine medication
compliance, rather it focuses on provider information giving and modeling and patient
self-care behaviors, such as exercising, foot examination, glucose monitoring, and getting

preventive care.

M-CARE structure a_ng it_s membership

M-CARE is part of the University of Michigan Health Systems (UMHS). M-
CARE was founded to provide affordable medical care to the university’s staff and
faculty. In subsequent years M-CARE expanded and enrolled private sector employers,
such as GM, EDS, Chrysler and Pfizer. During the time of data collection, M-CARE was
responsible for 13,000 Medicaid members, 177,000 commercial HMO members, and

6,000 Medicare members].

 

7 The membership numbers are given for the year 2001. In the year 2002 Medicare membership
was available in only two counties.

47

Approximately 50% of M-CARE’s medical care is delivered through the
University Hospital System and its satellite centers. These centers, like the University’s
Medical School are part of the University of Michigan Health System. Evidence based
treatment and adherence to medical practice standards are emphasized. Furthermore, the
University of Michigan Medical School receives funding for research and for
demonstration projects. Therefore, funding is available for members with chronic
conditions, including diabetes, to cover services under research grants that are not a
covered benefit by other insurers. Thus, financial barriers are limited or removed for
diabetic M-CARE members. M-CARE does not offer financial incentives to physicians
for low utilization, and physicians are not financially responsible for the cost for diabetes
education or diabetes self-care supplies. Therefore, there are no financial incentives for
physicians to discourage person with diabetes to participate in diabetes self-care

education.
M—CARE ’s diabetes population

M-CARE provides coverage to about 5,500 members with diabetes who live in
the Southeastern Michigan area. HMO covers 57%, Medicare 38%, and Medicaid 5% of
diabetic memberss. M-CARE sends a survey to its diabetic membership that explores
recall of information and services received as well as attitudes for practicing self-care
behaviors. Some of the questions were worded the same as the questions in the TRIAD

survey from which the data for this study comes. According to the Diabetes Health

 

8 The overall diabetes rate at M-CARE is low because the mean age of members is only 29 years.
Diabetes prevalence increases with age.

48

Status Questionnaire, the diabetic population is approximately 80% white, 10% Afi'ican-

American and 10% other ethnicities.

According to M-CARE’s Diabetes Health Status Questionnaire in 1998, 74% of
M-CARE’s people with diabetes have some college education and 15% have benefited
from graduate studies. In comparison, data from 1997 Blue Care Network diabetic
surveys show that only 5.1% of the diabetic membership had a college degree of any
kind. The BCN survey did not ask for graduate studies, therefore comparison at that level
cannot be made, but it appears that M-CARE’s diabetic members are well educated. This
may be important because a person who has a sound general education has less difficulty

assimilating diabetes information compared to someone who lacks formal education.

Data for this paper came from Translating Research into Action for Diabetes
(TRIAD), a study that has been funded by the CDC. TRIAD is a multiple site national
study however, the study population for this paper is limited to M-CARE’s type 2
diabetic membership. People with diabetes who are included in the study were diagnosed

at least twelve months prior to the survey.

49

People with diabetes were identified through:

0 Pharmacy records: At least 1 prescription for insulin or oral anti-diabetic
agent.

0 Hospital discharge diagnosis: At least one hospitalization with primary or
secondary diagnosis of diabetes.

0 Laboratory test records: At least two hemoglobin Alc tests or fasting
blood glucose measures in the past year.

0 Self-identification to the disease management nurse as person with
diabetes and one outpatient visit or hospitalization during the past 18

months.

Study subjects came from a random sample of diabetic M-Care members. At the
UMHS where access to lab data is available, members who met above criteria and had a
Hemoglobin A1,, > 9.0 were over sampled to ascertain adequate representation of

participants with different levels of glucose control.

50

The following exclusion criteria applied to potential subjects:
1. Patients who report having type 1 diabetes;
2. Patients with onset of diabetes before age 309 ;
3. Patients who report m); having “diabetes or a blood sugar problem”'°;
4. Patients who are not able to consent for themselves;
5. Patients who do not speak English;
6. Patients who do not receive their diabetes care from a primary care

physician (i.e., patients who receive their diabetes care from an

endocrinologist).

From each health system about 400 patients who met selection criteria were
selected, with over-sampling of patients with HbAlc > 9.0 % (those are members who are

poorly controlled, hemoglobin Ale should be below 7 %). In order to increase people
with diabetes from minority groups and those of lower SES, practices that serve

disproportionately low income and racial minority patient populations were included.

Self-administered questionnaires were mailed to eligible subjects. The surveys
explored the member’s perception of his/her health, life style activities, diabetes care,
provider information regarding self-care, as well as participation in diabetes education.
The questions were formulated before I joined the study team and therefore are not

worded exactly to correlate to the measures of this study.

 

9 Young patients are excepted because most people with diabetes who develop the disease before
age 30 are type 1 and the study is limited to type 2 diabetics.

'° Those who report that they do not have diabetes are either in the very early stages of the disease
or may have landed by error in the diabetes registry.

51

The initial contact letters were sent from M-CARE on M-CARE letterhead
stationery and signed by the M-CARE Medical Director. With the initial letter, contact
information was provided and a pre-addressed stamped envelope was enclosed to allow
potential study subjects to opt out of the study before the Informed Consent form and
survey were mailed. At least two weeks were allowed for responses from mailings. The
researchers did not contact potential subjects who opted out in order not to pressure any
one to participate; this was a mandate of the University of Michigan Internal Review
Board approval. However, according to protocol, members who did not opt out were
mailed reminders or contacted by phone if the survey was not returned. This resulted in a

total of 1649 surveys from members.

52

Variables selected M flip study m their measures

Provider behavior
Provider information giving and modeling:

There are five questions that explore recall of provider advice for a specific area
of self-care. According to clinical guidelines, the diabetic diet, and blood glucose should
be reviewed and monitored during each visit (American Diabetes Association 2001;
Michigan Quality Improvement Consortium 2000). One could argue that it would be
possible that the person with diabetes might have been taught or advised by the team, but
the patient may not remember. However, according to American Diabetes Association’s
guidelines, people with diabetes need to be seen by their provider a minimum of two
times a year, more often if there are complications or if control is lost. It is difficult to
imagine that someone reviewed diet, exercise and blood glucose levels twice in the last
twelve months and the patient cannot remember being advised a single time.
Additionally, if a patient cannot remember being asked about diabetic diet, exercise or
blood glucose monitoring, then the communication was ineffective. Kravitz (1996)
found that recall of self-practice recommendations varied from 90 to 24% but the study
measured will from one visit. not from repeated visits and repeated recommendations.
Sometimes when someone does not remember it indicates that if it was indeed
communicated, the communication was not made relevant and/or it was not clear so that

the patient could not assimilate the information.

53

Question 16: For the next set of items, please indicate if your current doctor or other
health care provider (such as diabetes educator or nurse) in your doctor’s office

explained to you, showed you, or gave you information about the following:

 

Yes No Unsure

 

A How to care for your feet?

 

C How to exercise appropriately?

 

 

 

 

 

 

D What is a good number for your blood sugar?

 

Question 35: During the past 12 months, have you received any of the following types

of diabetes related information from your doctor’s office or health care plan?

 

Yes No Unsure

 

Diabetes materials (e. g. pamphlets or newsletters, audiotapes or
videotapes)

 

 

Information about diabetes education (such as support groups or
one-on-one counseling, advice services, or Internet sites)?

 

 

 

 

 

Provider modeling

During the regular office visit the provider performs some of the same activities
that the patient is expected to practice at home. Thus the provider reviews the blood
glucose log from home with the patient and demonstrates or models how to make
decisions based on the recorded blood glucose levels. Likewise the provider inspects the

feet in similar ways as the patient is expected to examine his/her feet at home. Thus the

54

 

 

provider demonstrates and models the self-care behaviors that are expected from the

person with diabetes during regular office visits.

Question 20: Does your doctor or some other health care provider in your doctor’s office
review your home blood/urine sugar test results at each visit? (Yes, No, Not
Sure)

Question 22: During the past year, how ofien did your doctor or some other health care
professional examine your feet with your socks off?

Every visit; most visits; at least one of the visits; none of the visits; not sure

Question 23 .' When was the last time a doctor or other health professional tested the
feeling in your feet or legs by touching them with a monofilament (which looks

like a short piece of fishing line)?

During the past 12 months; more than a year but less than 2 years; more than 2 years;

never; not sure

Participation in Self-Care Education

Diabetes self help classes for this population are certified by the Michigan
Department of Community Health. The curriculum is standardized and includes
information about the disease process, complications, diet, exercise, diabetes medication,
influenza vaccination, retina examinations and blood glucose testing as well as medical
standards of care. Likewise patients are referred for individual teaching to Certified
Diabetes Educators who follow the standards of the American Diabetes Association and

the Association of Diabetes Educators and cover the same areas of self-care.

55

Question 36: During the past 12 months, have you used any of the following diabetes-

related services or attended any of the following diabetes-related programs

 

Yes No

 

A diabetes support group

 

 

One-on-one or group diabetes education

 

 

 

 

 

Measures for diabetes self-care

There are several questions concerning self-care in the survey. Self-care that can
be divided into home based self-care, such as exercise, blood glucose monitoring and foot
examinations. Then there are two office based self-care behaviors, such as influenza
vaccination and retina examination. However, as mentioned earlier, the patient can

obtain those services independently without provider support.

Exercise

Question 95 : Think about the extra time you yourself spend taking care of your diabetes-
related health problems. (Related problems might include high blood pressure,

high cholesterol, or heart and circulation problems).

On a typical day, how many minutes do you spend exercising?

Blood Glucose Monitoring

Question 1 7: Do you test your blood sugar levels at home? (Yes, No)

56

Foot Examination

Question 42: Do you or someone in your home check your feet for sores every day? Yes,

No

Flu Shots

Question 27: Did you get a flu shot during the past 12 months? Yes, No

Retina Screening

Few people with diabetes will understand the difference between a “regular visit

to the ophthalmologist” and a dilated eye exam, hence this question is rather explicit.

Question 21: When was the last time you had an eye exam in which your pupils were

dilated (drops in your eyes that make you temporarily sensitive to bright light)?

During the past 12 months; more than a year but less than 2 years ago; more than 2 years;

never; not sure

57

Background variables

Complications (severity of disease)

The person with the more severe diabetes is more likely to be referred for diabetes
counseling or to be counseled by the practitioner (Koltun 1986). Therefore it is important
to control for the severity of the disease. As stated earlier, type 2 people with diabetes
who have developed symptomatic complications are more severely ill and in the
advanced stages of diabetes. Complications increase with time lived with high blood

glucose levels.

The survey asks directly for the presence of heart disease, which is the outcome of
large blood vessel disease, and it is preceded by small blood vessel disease. Therefore,
people with diabetes who have been told that they have or had heart disease, coronary
bypass surgery, stroke or angioplasty are experiencing the complications associated with
advanced diabetes. According to the Health Belief Model, this makes the threat from
diabetes complications more salient to the patient. Also some people may be willing to
change their life-style once their life is threatened. Therefore the ratio of cost of self-care
to the threat from diabetes is changing in favor of self-care when complications develop.

Question 38: Have you ever been told by a doctor or someone in your doctor’s office that

you have had a heart attack, a “coronary” or a “myocardial infarction”?
Yes No

Question 39: Have you ever been told by a dbctor or someone in your doctor’s office that
you have had a stroke, cerebrovascular accident, blood clot or bleeding in the

brain or a transient ischemic attack or “mini-stroke”? Yes No

58

Question 40: Have you ever had any of the following procedures:

A Surgery to bypass or unclog arteries to your heart. Yes, No, Unsure
B Angioplasty or a balloon to unclog arteries to your heart or leg.

Yes, No, Unsure

C A toe, foot, or leg amputation Yes, No, Unsure
Demographic measures

Gender

Question 10: What is your gender? Male Female
Age:

Question 9: What is your birth date? (month/day/year)

Education and Income

Question 100: What is the highest grade of school that you completed?

8th grade or less

Some high school, but did not graduate
High school graduate or GED

Some college or 2-year college degree
4-year college graduate

More than 4-year college degree

59

Question 99: Which income category below best describes your total annual household
income before taxes?

Less than $5,000

$5,000 to under $7,500
$7,500 to under $10,000
$10,000 to under $12,500
$12,500 to under $15,000
$15,000 to under $20,000
$20,000 to under $25,000
$25,000 to under $30,000
$30,000 to under $35,000
$3 5,000 to under $40,000
$40,000 to under $75,000
$75,000 to under $100,000
$100,000 or above

Ethnicity

Question 102: What is your race? Please select all that apply.

American Indian or Alaska Native

Asian, if Asian then Chinese, Filipino, Japanese, Korean
Asian (or East) Indian

Native Hawaiian

Pacific Islander

Black or African American

White

Other, specify

60

The data was collected via surveys by M-CARE as part of the national
Translating Research into Action for Diabetes (TRIAD) research project. The data

needed to be formatted and coded for analysis.

61

Coding 9&1 variable computation

The following section describes the recoding process and quality assurance after
the recoding to assure accuracy of the data and adherence to the study protocol. All
subjects who stated that they were diagnosed with diabetes before age 30 were removed
from the data to follow the original study protocol and to ascertain that the data contain
only subjects with type 2 diabetes. That left 1439 cases for analysis. Next the variables
were recoded. Respondents who answered that they were “not sure” whether they
received information from their providers were coded as “no”. As previously mentioned,
if a person with diabetes cannot remember being taught or receiving information then for
this study’s purposes the person with diabetes could not have acted upon the information.
Female was coded as “0”, male was left as “1”. Thus female becomes the reference.
Next the subjects’ ages were calculated based on the birth date. The data on retina
examination was dichotomized into those that had a retina examination in the last 12

months and those who did not.

Income was originally coded in uneven intervals; therefore it was recoded as
income in thousands with the midpoint in each category. Afier the recoding was
completed, summary frequencies of original and recoded variables were compared. They
matched. A random check on every 25th case comparing values within cases showed no

errors and the data were considered clean and ready for analysis.

62

Aggregated measures and indexes

The next paragraph describes the method for developing aggregate measures that
were used for various analyses, either to investigate the effect of aggregated behaviors or
to create proxy measures when specific measures were not available. First aggregate
measures for self-care were computed. Self-care at home includes minutes of exercise
per day, blood glucose monitoring and foot care. Foot care and blood glucose testing
were dichotomous categorical variables. Minutes of exercise ranged from “0” to “300”
per day. Therefore, exercise was standardized. An index score for home-based self-care
was calculated by taking the mean of standardized exercise, home blood glucose
monitoring and foot inspections. In the same manner an index score for office-based
self-care was created by averaging the responses to “influenza vaccination” and “retina
screening”. Then the two index scores for office and home based self-care were added to

create an index for “overall self-care”.

Next an index score for “complications” was calculated by adding the number of
complications such history of stroke, heart attack, amputations, coronary artery bypass
graft and angioplasty. In the same manner an index score for provider information giving
was created to be used as proxy for information on influenza vaccination and retina
examination because there were no specific questions for information giving regarding
flu vaccinations and retina examinations. Likewise an index score for diabetes education
was created by calculating the mean of one on one education and using a diabetes support

group.

63

Because of the interest of diabetologists in Asians, there were six choices for
Asian ethnicity, but all six categories totaled only 38 subjects1 I; there were only 34
subjects who classified themselves as “Hispanic”; 187 subjects who classified themselves
as “Black”; 84 who classified themselves as “other” and 1095 subjects who classified
themselves as “White”. Therefore the variable was collapsed into three categories:

White, black and other.

Table2: Ethnicity

 

 

 

 

 

 

 

 

Outliers

The data was examined for outliers. One case was identified as outlier for
exercise. The person stated that he exercised 8 hours every day because of his
diabetes. The subject was a Japanese male; age 47 years with a graduate degree
and a BMI of 35. He had been diagnosed with diabetes at age 33 and was using
three insulin injections per day. This raised several concerns: Any physically
active person diagnosed at age 33 using 3 insulin injections per day may be a

person with type 1 diabetes. Furthermore, it is difficult to believe that a working

 

" Tuomiletho 2001 stated that Japanese type 2 diabetics are different from western type 2
diabetics, because their bodies remain lean throughout the disease while westerners become obese.

64

adult will exercise 8 hours a day because of the diabetes and still have a BMI of 35,
which is considered “morbid obesity”. There were two more people with diabetes
who stated that they exercised five or four hours every day. However, the question
did not define the intensity of exercise, thus someone might play golf for four hours
a day. Those two subjects did not have a high BMI and therefore those two cases

were not excluded from the analysis.

The minimum weight was only 85 lbs, which is surprisingly small in a
diabetic sample. However, upon investigation, this turns out to be an 80-year old
black female who was only 5’3”tall who had a stroke and did not feel well.

Therefore, this is a believable weight and was not excluded from the analysis.

The maximum weight was 500 lbs and initially seemed excessive.
However, upon investigation this is a 5’8” tall white female who does not exercise
and does not state that she follows a diabetic eating pattern. Her sole method of
diabetes treatment is oral medication, therefore she was not excluded either because
people with diabetes develop morbid obesity unless the metabolic disorder is

controlled. Under those conditions her weight becomes believable.

Both OLS and logistic regression studies were used for analysis. In the first
block the background variables were entered in order to control for them. Those
included demographics (age, education, gender, ethnicity, income), the
complications index score (patient had history of heart attack, coronary bypass
surgery, angioplasty, amputations) and years lived with diabetes. In the second

block the variables of interest were entered, such as provider information giving

65

and provider modeling of diabetes self-care behaviors. Whenever information
regarding the dependent variable or modeling is part of standard diabetes education
curriculum, seeing a diabetes educator was also entered as dependent variable

because the diabetes educator can be looked upon as “part of the provider staff”.

For regression studies a “listwise” deletion method was used which explains
varying sampling sizes in the analysis. The only exception is the “provider blood glucose
log review”. That question had too many missing values that needed to be replaced with

the mean.

For each OLS regression the standardized residuals were plotted against the
standardized expected values. The residuals were randomly distributed along the
expected regression line; for minutes exercised a square root transformation was

needed to achieve a more normal distribution of residuals.

66

RESULTS

The preceding sections have explored diabetes treatment, physician and patient
roles, theories of health behavior, developed the hypotheses and the methodology. The
following section describes the results of the data analysis. As mentioned earlier the data
were collected from a sample of M-CARE diabetic members and are part of the national
TRIAD study, which is ongoing and funded by the Centers for Disease Control. The
survey was developed and the data collected before I became involved with the research.
Therefore, the wording of the survey questions is not always as specific to the measure as
it could be and at times indexes were created by aggregating similar measures and have

been used for analysis.

Sample Description

During the next few paragraphs we will look at the sample compared to M-
CARE’s diabetic membership, distribution of continuous variables, self-care behaviors,

provider practice behaviors and finally missing data.

Once the data were cleaned to fit the study protocol, there were 1438 subjects in
the sample. 728 subjects (50.6%) were female, and 709 (49.3%) male, one subject did
not identify the gender. The mean age of the sample was 65.2 years, the median age was
68.2 and the mean education was just short of a four-year college degree. Mean income
was somewhere around $41,300 per year; the median income was $27,500. The

following table describes the distribution of the continuous data collected, demographics,

67

years since diagnosed with diabetes and the number of minutes exercised per day to

control diabetes.

Table 3: Descriptive summary of continuous variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tvariables' ‘ '1 N ”Minimum ' Maximum M Mean . Median fitandard
. . W. ,, , . , ,, dev.
Age. 1364 30 91 65.2 68.2 11.9
Weight 1405 85 500 201.4 195.0 49.8
BMI” 1397 15 78 31.9 30.5 7.1
Years since diabetes 1360 1 51 12.0 9.5 8.9
diagnosis
Complications index 1329 0 l . 14 0.0 .22
Income in thousands 1142 <5 > 100 41.3 27.5 37.9
Education 1377 <8‘h grade > 4yrs college Some college H.S grad 10 yrs
Minutes of exercise/day 121 l 0 300 21.3 15.0 28.1
Square root transformed 121 l 0 17.3 3.5 3.9 3.0
exercrse
Home based self-careI 1438 0 2 .49 .5 .29
Office based self-care2 1432 o 3 1.66 2.0 .46
Overall self-care 3 1432 o 4 2.15 2.25 .56
Provider information4 1319 0 1 0.7 0.8 0.3
giving index
Provider modeling index 1428 0 3 1.2 1.3 0.9

 

" Anyone less then 30 years of age was deleted from the data to ascertain that the database contained only type 2
diabetics. That was in line with the TRIAD study protocol.

' BMI: up to 25 is normal weight; 25-30 is considered over weight, 2 30 is considered obese

This sample is comparable to other samples taken from M-CARE’s diabetic

population. Subjects had been diagnosed with diabetes for at least one year. As

discussed previously, the prevalence of diabetes increases with age, thus the diabetic

 

' Home based self-care is the mean of monitoring feet, blood glucose and standardized exercise
2 Office based self-care is the mean of obtaining an influenza vaccination and retina examination
3 Overall self-care is the sum of office based and home based self-care

‘ The provider information giving index is the mean of all information giving by the provider and staff

68

 

population is older than the general population and most are obese or overweight, as is

expected in a population with type 2 diabetes in the United States.

The next table gives a description of the dichotomous variables and the overall

distribution of self-care, provider behavior and diabetes complications. The table is

intended to provide a general sense of patient and provider behavior in this sample.

40.8% of participants had developed at least one complication. The table indicates the

number of answers to each question and the subjects who answered affirmative.

Table 4:_Summary of self-care and provider behavior variables

 

 

 

 

4 . .. -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

gvaaasrzg‘”" " ‘ "’ " ' ' “Tatar Answered ‘ " "%‘fofsmpre’

;j , ._ , .. L . j 3 _ ., . . Answers Yes . .. answered “yes”
Patient Self-care Variables

Patient monitors blood glucose 1432 1164 80.9
Patient examines feet 881 470 32.7
Patient had influenza vaccination in last 12 month 1419 935 65.0
Patient had retina examination in the last 12 months 1414 1086 75.5
Patient used one on one diabetes education 1302 163 11.3
Patient used diabetes support group 1307 61 4.2
Patient Reported Provider Behaviors

Provider gave blood glucose information 1281 1060 73.7
Provider gave exercise information 1293 816 56.7
Provider gave foot care information 1312 894 62.2
Provider checked glucose log 101 l 479 33.3
Provider checked feet 1411 1081 75.2

 

Only 33.3% of the subjects stated that their providers checked their blood glucose

log, while foot monitoring with the recommended filament was provided to 75.2% of the

diabetic members. 32.7% checked their own feet at home and 80.9% of the subjects

monitored their home blood glucose and 75.5% stated that they had a retina examination

69

 

within the last twelve months. Only 11.3% of the subjects stated that they received

diabetes education in the last year.

Participants were permitted to skip questions and therefore there are not 1438
responses to all questions, thus the varying subject numbers in the tables and the
regression studies. The following table shows the number of responses to each question
and the percent of missing responses. Income, patient checks his/her feet and provider
checks blood glucose log are the questions with the most missing answers. The
following table shows the number of replies to each question and the percent of non-

missing and percent missing answers.

70

Table 5: Missing Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. -, . . _. , w VéfiEEIES —. .. , 53743911 I”? M19196"; "
Patient Demographic Variables

Age 1429 0.7
Education 1 3 77 4.2
Ethnicity 1438 0.0
Gender 1437 0.0
Income 1142 20.6
Patient exercises 121 1 15.8
Patient checks feet 881 38.7
Patient monitors blood glucose at home 1432 0.4
Patient received retina examination 1414 1.7
Patient received influenza vaccine 1419 1.3
Complications

Patient had a stroke 1315 8.5
Patient had a heart attack 1314 8.6
Patient had amputation 1291 10.2
Patient had coronary bypass surgery 1314 8.6
Patient had angioplasty 1305 9.2
Years since diabetes was diagnosed 1360 5.4
Provider Behaviors f

Provider gives information on foot care 1312 8.8
Provider gives information on blood glucose 1281 10.9
Provider gives information on exercise 1293 10.1
Provider checks patient’s feet with filament 1237 14.0
Provider checks patient’s glucose log 1011 29.7
Provider examines patient’s feet 1411 1.9

 

 

 

Now that we have a general understanding of the overall distribution of patient
self-care, provider information giving and health status the paper will proceed to display
the results of hypothesis testing. Each self-care behavior was tested independently to

measure the impact of provider information giving or provider modeling while

 

controlling for patient characteristics, such as demographics and diabetes complications.
The patient characteristics were entered in the first block; the main independent

variable(s) were entered in the second block.

72

Hypotheses testing gig data analysis

The following section will display the results of hypothesis testing for specific
self-care behaviors as dependent variables and then for aggregate self-care behaviors.
Thereafier we will look at the results of hypothesis testing for diabetes education and

demographics as dependent variables.
Self-care activities

Self-care behaviors were examined comparing each specific self-care behavior to
specific advice/information given or specific provider modeling whenever possible.
Whenever there was no specific question to match the self-care behavior (retina
examination and influenza vaccination), the provider information giving index5 was used
as proxy. Thereafter the summary indexes for home based self-care, office based self-
care and total self-care were examined using the same method. The results are reported

according to self-care behaviors and all hypotheses that deal with the behavior.

Hypothesis 1.1: People with diabetes who recall their provider’s advice to
engage in specific diabetes self-care activities (blood glucose testing,
exercise, foot examination) are more likely to report practicing that self-care

behavior.

 

5 The information giving index is the mean of all information giving by the provider and staff

73

Hymthesis 1.2: Those who receive more self-care advice from their
provider are more likely to engage in office based self-care behaviors

(influenza vaccination and retina examination

Hypothesis 2: When the provider models self-care behaviors the patient is more

likely to practice self-care behaviors.

Hypothesis 2.1:When people with diabetes recall that the provider examined the
home blood glucose log; the patient is more likely to monitor his/her home blood

glucose levels.

Hypothesis 2.2: When people with diabetes recall that the provider examined
his/her feet, the person with diabetes is more likely to report examining his/her

own feet at home.

Hypothesis 3.2: Those who receive diabetes education are more likely to practice

appropriate self-care behaviors.

Hypothesis 5.2: Those who are better educated are more likely to practice

appropriate diabetes self-care behaviors.

Specific self-care behaviors

Home blood glucose testing: Those who test their blood glucose levels at home

and those who do not were compared and whether they received information on blood

glucose levels and whether their provider examined the blood glucose log during office

visits. Table 5 displays the results of Logistic Regression listing the Beta, Exp(B)[odds

74

ratio], and Wald statistic. As always, demographics, complications and years since

diabetes diagnosis have been controlled for.

Table 6: Logistic regression: The effect of patient characteristics and provider

information giving and modeling on home blood glucose monitoring

 

Dependent variable: Home blood glucose monitoring

 

 

 

 

 

 

 

 

 

 

 

 

     

 

 

 

 

 

 

Variables Hypothesis 1: Provider Hypothesis 2: Provider

information giving n=102~5 . .. (modialingn?= 1057

B 5 ’ Exp (Bf Ward 13' ' Exp-(B) ' Ward
- 19411:" toads .
Ratio] ' Ratio} *

Constant 1.505 4.598‘ 2.615 15.325
Age -.025 .980 5.555 -.025 .976 8.480“
Education —.1 16 .891 2.316 -.099 .906 1.850
Black ethnicity -.360 .698 1.747' -.367 .693 1.976
Other ethnicity -.676 .509 5.180 -.637 .529 5.490“
Gender -.297 .743 2.556 -.252 .777 2.027
Income in thousands .002 1.002 .541. .002 1.002 .356
Complications index .263 1.300 6.204' .314 1.369 9.493"
Years since diabetes diagnosis .081 1.085 32.964‘” .067 1.070 27.4621‘”
Patient saw diabetes educator 7.811” 10.741".
Provider gave blood glucose 31.379”.
information

 

Provider monitors log

 

 

 

 

 

 

 

 

 

 

Cox-Snell R2 .094
Nagelkerke R2 .156 .106
*p 5 0.05; " p 5 .005 m p 5 .00 (all probabilities are two tailed)

 

All patient characteristics were entered into the regression; the Nagelkerke R2 was

.089; when provider information giving was added it increased to .156. Likewise the Cox

& Snell R2 increased from .054 to .094 when provider information giving was added to

the patient characteristics. The numbers come from the first block analysis and are not

given in the table. The block is significant at p 5. 001. When we control for

75

demographics and diabetes complications, seeing a diabetes educator is significant with
an odds ratio of 2.788 and p=. 005. However, the strongest predictor, controlling for all
other factors, is provider information giving regarding blood glucose testing. The odds
ratio is 3.204 and it is significant at p g .001. The model is significant and information
giving about blood glucose levels is the most significant predictor for monitoring blood
glucose. The data support the Hypothesis 1.1: People with diabetes who recall their
provider’s advice to engage in specific diabetes self-care activities (blood glucose testing)

are more likely to report practicing that self-care behavior.

Provider modeling by evaluating the home blood glucose log and evaluating the
control is not significant; however receiving individual diabetes education6 is significant
at p 5 .001. All patient characteristics without information giving resulted in a
Nagelkerke R2 of .085; when modeling was added it increased to .106. Likewise the Cox
& Snell R2 increased from .051 to .064 when provider information giving was added.
The model is significant at p _<_ .001; the block is significant at p=. 007. The increase is
due to seeing a diabetes educator who will monitor the home blood glucose log in order
to monitor the outcome of self-care. The data support hypothesis 2.1:When people with
diabetes recall that the provider examined the home blood glucose log the patient is more

likely to monitor his/her home blood glucose levels.

More schooling or education does not increase home blood glucose monitoring.

Education is not significant and the Exp(B) is only .891. The data disconfirrn hypothesis

 

6 During individual diabetes education, the educator checks the blood glucose log and thus models how to
evaluate glucose levels and adjust medication and life style accordingly.

76

5.2: Those who are better educated are more likely to practice appropriate diabetes self-

care behaviors.

Patient foot monitoring: Next the data was analde for foot monitoring at home
in the same manner (see Table 6). The independent variables are information giving
regarding foot care and provider modeling foot examination. Demographic factors,

complications and years lived with diabetes have been controlled for.

Table 7: Logistic regression: The effect of patient characteristics and provider

information giving or modeling foot care on foot examination.

 

Dependent variable: Foot monitoring

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables Hypothesis ’1: Provider Hypothesis 2: Provider
information giving F698 modeling n=703
B Exp (B) Wald B Exp (3) Wald
- IOrIds [Odds ‘
Ratio} Katie]
Constant .021 .659 —.326 .270
Age .002 1.002 .068 .000 1.000 .002
Education -.O3O .970 .178 -.027 973 .150
Black ethnicity .070 1.072 .283 .119 1.127 .872
Other ethnicity -.053 .948 .278 -.098 .907 .990
Gender -.151 .860 .767 -. 169 .845 1.008
Income in thousands -.075 .928 6.027' -.007 .993 8.079“
Complications index 1.028 2.796 4.445' .207 1.238 6.691 W
Years since diabetes -.014 .986 .955 -.004 .999 .005
diagnosis
Patient saw diabetes .519 1.680 4.351‘
educator
Provider gave foot care 1.271 3.564 52.291.
information W
Provider monitors feet .554 1.740 50.205."
Cox-Snell R2 .1 19" .106‘
Nagelkerke R2 .159” .142“
‘p 5 0.05; "

p 5 .005 “" ; p 5 .001 (all probabilities are two tailed)

77

 

 

When only patient characteristics were entered into the regression, the Nagelkerke
R2 was .05; when provider information giving was added it increased to .159. The Cox &
Snell R2 increased from .037 to .119 when seeing a diabetes educator and provider
information giving were added. Those numbers came from the first step and the numbers
are not given in the table. Provider information giving is the strongest and most
significant predictor for patient foot care. The provider giving the information is more
effective than seeing a diabetes educator. The data support Hypothesis 1.1 for foot
examinations: People with diabetes who recall their provider’s advice to engage in
specific diabetes self-care activities (foot examination) are more likely to report

practicing that self-care behavior.

When provider modeling foot examinations was added to the patient
characteristics the Nagelkerke R2 increased fiom .053 to .142 and the Cox & Snell R2
increased from .040 to .106 when provider modeling was added to patient characteristics.
The block and the model are significant at p z .001. Thus hypothesis 2.2 that provider
modeling of foot examination increases patient foot care is supported. The model was

significant at p 3.001 and the odds ration was 1.740.

The impact of the patient’s education level on foot monitoring is not significant

and the Exp (B) is only .97. The data disconfirrn hypothesis 5.2 for foot monitoring.

Exercise behaviors: Exercise was stated in “extra minutes spent on the average
day, due to diabetes”. As previously, demographics, complications and years lived with

diabetes have been controlled for. As mentioned earlier a square root transformation was

78

performed to achieve a random distribution of residuals. The following table displays the

results of OLS regression of exercise and exercise information giving.

Table 8: OLS regression: The effect of provider information giving and patient

characteristics on exercising (square root transformation of exercise)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Variable: Minutes exercised per day n=912
Variables Unstandardized Standardized t
. . , , Coefficient Coefficient

Constant 1.984 2.644"
Age 6.098E-03 .024 .633
Education .185 .080 2249‘
Black ethnicity .218 .047 1.463
Other ethnicity 3.282E-02 .009 .277
Gender 8.280E-02 .014 .422
Income in thousands -5.894E-03 -.076 -2002‘
Complications index -3.817E-02 -.013 -.393
Years with diabetes -1.475E-02 -.042 -l .240
Patient saw diabetes educator .581 .067 2.119.
Provider information on exercise 1.006 .161 5.098".
R2 .041
F value 4.638."
*pg 0.05;

” pg .005

m pg .001 (all probabilities are two tailed)7

All patient characteristics were entered into the regression, the Nagelkerke R2 was

.002; when provider information giving was added it increased to .041; the increase is

 

7 ' Home based self-care is the mean of blood glucose monitoring, checking of feet and standardized
exercise

b Office based self-care is the mean of influenza vaccination and retina examination

° Overall self-care is the sum of home based and office based self-care

‘ Provider information giving index is the mean of all information given

‘ Provider modeling is the mean of provider monitoring the blood glucose log and examining feet.

79

significant at pg. 001, a beta of .161 and an unstandardized coefficient of 1.006; diabetes
education is also significant at p=. 034 but the beta is only .067 and unstandardized
coefficient is .581. Overall the slopes are very small because of the square root

6619’

transformation of minutes exercised. However a square root increase of could for
example increase the square root from 6 to 7 thus the amount of exercise increases from
36 minutes to 49 minutes per day. The data support the hypothesis 1.1: People with

diabetes who recall their provider’s advice to engage in specific diabetes self-care

activities (exercise) are more likely to report practicing that self-care behavior.

The patient’s education level has a significant effect on exercising. The beta is
.08 but it is significant at p = .025. The data support hypothesis 5.2 for exercise: Those
who are better educated are more likely to practice appropriate diabetes self-care

behaviors, in this case exercise

This concludes the analysis of home based self-care behaviors and the next

section examines self-care behaviors outside the home.

Influenza vaccination: We will look at influenza vaccination. According to
clinical treatment standards the influenza vaccination should be given once a year
between October and February. There is no specific question in the survey whether or
not the provider gave information regarding the influenza vaccine; therefore the
information giving index was used as proxy. The information giving index has a range
from “0” to “1” as mentioned earlier. Demographics, complications and years lived with
diabetes have been controlled for and entered in the first block; the provider information

giving index is entered in the second block. The following table displays the results of

80

logistic regression concerning influenza vaccination and the provider information giving

index.

 

Table 9: Logistic regression: The effect of patient characteristics and provider

information giving on obtaining influenza vaccination

Constant

Education
Black ethnicity

Other ethnicity
Gender
Income in thousands

index
Years since diabetes
Provider information
Cox-Snell
Nagelkerke

 

‘ p_<_ .005
‘" pg .001 (all probabilities are two tailed)

When all patient characteristics were entered into the regression, the Nagelkerke
R2 was .054; when provider information giving was added it increased to .094. Likewise
the Cox & Snell R2 increased from .062 to .067 when provider information giving was
added to the patient characteristics. The model is significant at p 5.001. The provider
information giving index has an odds ratio of 1.716, a Wald statistic of 6.01 8 and a

significance of p= .014. The data support hypothesis 1.2: Those who receive more self-

81

care advice from their provider are more likely to engage in self-care behaviors

(influenza vaccination).

Education also has a significant effect on obtaining the influenza vaccination.
The p=. 007 and the Exp(B) = 1.78. The data on influenza vaccination support
hypothesis 5.2: Those who are better educated are more likely to practice appropriate

diabetes self-care behaviors. Next we will examine retina screening.

Retina screening: As mentioned in the methods section, the measure was
dichotonrized into those who obtained a retina examination in the last 12 months and
those who did not. Again there was no specific question whether the subject received

retina examination information and the provider information giving index was used as

proxy.

82

Table 10: Logistic regression: The effect of patient characteristics and provider
information giving on retina screening in the last 12 months

Constant

Education
Black
Other ethnicity

Gender

Income in thousands
index

Years since diabetes

Provider information

Cox-Snell

 

<

" p 5 .05
m p 5 .001 (all probabilities are two tailed)

When all patient characteristics were entered into the regression, the Nagelkerke
R2 was .053; when provider information giving was added it increased to .06. Likewise
the Cox & Snell R2 increased from .035 to .039 when provider information giving was
added. Again these R2 values come from the stepwise entry method and the data are not
displayed in the table. The model is significant at p = .027 and the increase is significant
at p 5.001. Thus the data support hypothesis 1.2: Those who receive more self-care
advice from their provider are more likely to engage in self-care behaviors (retina

examination).

83

Education also has an effect on obtaining retina examination with an Exp(B)
=1 .159 and p = .029. The provider information giving index has an odds ratio of 1.710
and the data support hypothesis 5.2: Those who are better educated are more likely to

practice appropriate diabetes self-care behaviors.

In the previous section we have measured the effect of provider information
giving and behavior modeling, when appropriate on specific self-care behaviors such as
blood glucose testing, foot examination, exercising, obtaining influenza vaccination and
retina screening while controlling for demographics, complications and years lived with
diabetes. The following section will test index measures for office based, home based as
well as the overall self-care behavior indexes. Demographics, complications and years

lived with diabetes are again controlled for.

Aggregated Self-Care Analysis

Aggregated self-care behavior indexes are

1. The mean of reported home based self-care behaviors, in this case
standardized minutes exercise, foot monitoring and home blood glucose
monitoring.

2. The mean of reported office based self-care behaviors, in this case

influenza vaccination and retina examination.

3. Overall self-care, which is the sum of office, based self-care behavior and

home based self-care behavior.

The aggregated self-care behaviors are the dependent variables and the provider

information giving index and the modeling index are the independent variables.

84

Oflice based self-care

Office based self-care is the mean of obtaining influenza vaccination and retina

661”

examination. Both are dichotomous variables with values of “0” and . The main
independent variable is the provider information giving index and provider modeling
index. The provider information giving index is the mean of all information giving. The
modeling index is the mean of modeling the evaluation of the glucose log and

demonstrating foot examinations. The following table displays the results of OLS

regression.

85

Table 11: OLS Regression: The effect of patient characteristics and provider information

giving and modeling on office based self-care behaviors

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

Dependent variable: Office based self-care
Variables Hypothesis 1: Provider Information. . . Hypothesis 2; Provider modeling
' . ' giving n=1047 ‘ ‘ 114052
Unstandardizet Standardiud . t Unstandardiur Standardized t
Coefficient Coefficient Coefficient ' Coefficient
Constant .812 7.346'“ .896 8546'“
Age 8.624E-.03 .224 6.216’“ 8.026E-03 .210 5885‘"
Education 3.655E-.02 .108 3.157“ 3.082E-02 .092 2.683"
Black -3.277E-.02 -.048 -1.565 -3.854E-02 -.057 -1.845
ethnicity
Other ethnicity 1.276E-02 .023 .759 9.668E-03 .018 .586
Gender -3.526E-03 -.004 -. 127 6.062E-04 .053 1.451
Income in 6.323E-.04 .054 1.497 -9.769E03 -.01 l -.353
thousands
Complications -1.473E-02 -.034 -1.068 -7.822E-03 -.018 -.575
Index
Years since 50171303 .098 2.998” 4.8020E-03 .095 2.924‘
diabetes
diagnosis
Patient saw 5.078E-02 .039 1.281 4.925E-02 .038 1.253
diabetes
educator
Provider
information
giving index

 

 

 

 

 

5.744E-02

   
 

Provider
modeling
index

 

 

R2 .078 .077

 

 

 

 

 

F value 8743‘" 8.695‘“

 

‘p 5 0.05;
" p 5 .005
‘" p 5 .001 (all probabilities are two tailed)
When all patient characteristics were entered into the regression, the Nagelkerke

R2 was .067; when provider information giving was added it increased to .078. The

86

.095. Seeing a diabetes educator does not increase office based self-care behaviors; it is
not significant at p=.200. The data support hypothesis 1.2: Those who receive more self-

care advice from their provider are more likely to engage in self-care behaviors.

Office based self-care is also increased by provider modeling. The R2 increased
from .063 to .077 and it is significant at p5.001, the beta was .110, and it was significant
at p5.001. Seeing a diabetes educator was not significant in this model with p =. 210 and
a beta of .03 8. Thus the data support hypothesis 2: When the provider models self-care

behaviors the patient is more likely to practice self-care behaviors.

Education also is significant for obtaining office based self-care. The beta = .21
and .224 and the p value is p= .002. Thus the data support hypothesis 5.2: Those who

are better educated are more likely to practice appropriate diabetes self-care behaviors.

Home based self-care

Next, the effect of information giving on home-based self-care was examined.
The home-based self-care index is the mean of blood glucose monitoring, checking of
feet and time spent exercising (standardized). The independent variables are provider

information giving index and provider modeling index.

87

Table 12: OLS Regression: The effect of patient characteristics and information giving

and modeling on home based self-care

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
 
 

 

 

 

     

     

 

 

 

 

 

 

 

Dependent variable: W Home based self-care
villainous: , . . mouthed-Imam pat; “ ' 2t.:-.::-- ‘ j " _n=
* momentum“ «1:943 . _ 1974.4
- genus-riders: :‘SMdlrdiifll at. amusement-durum ' t
dCoefieient- Coefficient- , , . Coefficient . enemas-rt .
Constant .265 4.458‘“ .412 6967'“
Age l.535E-.04 .007 .206 -7.249E-O4 -.034 -.946
Education 1.293E-.O3 .007 .208 -8.849E-04 -.005 -.137
Black -l.471E-.03 -.004 —.130 -9.139E-05 .000 -.008
Other -1.215E-.02 -039 -1.338 -1.144E-02 -.037 -1.228
Gender -7.812E-.03 -.016 -.521 -4.462E-04 -.O68 -1.893
Income in -4.537E-.04 -.070 4990‘ -1.118E-02 -022 -.717
thousands
Complications 7.93 7E-.03 .032 1.068 1.622E-02 .066 2.1 17'
index
Years since 2.248E-.03 .078 2.497‘ 3.155E—03 .109 3.397'"
diabetes diagnosis
Patient saw .105 .143 4.931‘” .120 .162 5.409'“
diabetes educator
Provider .245 .305
information giving
index
Provider modeling 4.901E-02
index
R2 .145 .087
F value 17.688‘” 9.884‘”
‘p50.05;
" p5.005

‘“ p 5 .001 (all probabilities are two tailed)

When all patient characteristics were entered into the regression, the Nagelkerke
R2 was .019; when provider information giving was added it increased to .145 and it is
significant at p 5 .001. The beta for provider information giving is 0.305 and for diabetes

education it is .143 when we control for patient characteristics. The data support

88

hypothesis 1.2 for home based self-care behaviors: Those who receive more self-care

advice from their provider are more likely to engage in self-care behaviors.

When all patient characteristics were entered into the regression and the R2.
Adding the provider modeling index to the patient characteristics increased the R2 from
.029 to .087 and is significant at p 5. 001. The beta for provider modeling is .165 it is
significant at p 5 .001. Seeing a diabetes educator is also significant at p 5 .001 with an
unstandardized coefficient of .120 and a beta of .162. The data for provider modeling
also support hypothesis 2: When the provider models self-care behaviors the patient is

more likely to practice self-care behaviors.

Education did not have a significant effect on home based self-care behaviors and
the beta was negative (-.005) for using the provider modeling index. The data do not
support hypothesis 5.2 for home based self-care behaviors: Those who are better

educated are more likely to practice appropriate diabetes self-care behaviors.

Next we will study the relationship between provider information giving and

modeling and the overall score of self-care combined office and home based self-care.

Total self-care (Sum of home based and office based self-care indexes)

The provider information giving index, one on one education and provider
modeling index are the major independent variables, total self-care is the dependent

variable.

89

Table 13: OLS regression: The impact of patient characteristics and provider information

giving and modeling on aggregated total self-care8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Provider
modeling index

 

 

 

 

Dependent variable: Aggregate of all self-care behaviors
Variables Hypothesis 1. Provider Hypothesis 2:Prov1dermodeling
. informnongiwrngn=1038 Ill'lrm
Unstandardiz standardize r Unstandardinet Standardized r
d Coefficient costumer, , _ Coeflielent Coefficient
Constant 1.073 8.352‘“ 1.306 10.586’“
Age 8.859E-.03 .193 5.496'” 7.334E-03 .161 4569'”
Education 3.772E-.02 .094 2.804“ 2.976E-02 .074 2201‘
Black ethnicity -3.437E-.02 -042 -1413 -3.864E-02 -.048 -1.571
Other ethnicity 5.358E-.04 .001 .027 -1.800E-03 .003 -.093
Gender -1.118E-.02 -011 -.346 1.620E-04 .012 .329
Income in 1.805E-.04 .013 .368 -2.062E-02 -020 -.634
thousands
Complications -7.158E-.03 -014 -.447 8.214E-03 .016 .513
index
Years since 7.288E-.03 .119 3.748‘“ 7.953E—03 .131 4.099‘“
diabetes
diagnosis
Patient saw .156 .223 3.389" .169 .109 3.657'"
diabetes
educator
Provider .380 .100 7.455“
information
giving index

 

 

  
 
 
 
 

 

 

 

 

 

 

 

 

R2 .124 .103
F value 14.726" 11.942‘”
*p 5 0.05;
“ p 5.005

*” p 5 .001 (all probabilities are two tailed)

When all patient characteristics were added to the regression the R2 was

.062. When we add provider information giving to the patient characteristics, the

 

' Aggregated self-care is the sum of home based self care (mean of exercise, foot monitoring and blood
glucose testing) and office based self-care behaviors (mean of obtaining infl

90

information giving was significant at p 50.001. When we control for demographics
and diabetes complications and provider information giving resulted in the greatest
increase in self-care behaviors; The data support the Hypothesis 1.1: People with

diabetes who recall their provider’s advice to engage in diabetes self-care activities

are likely to report practicing that self-care behavior.

When we add the provider modeling index (mean of provider examines feet and
provider examines home blood glucose log) to the patient characteristics the R2 increased
from .060 to .103. Provider modeling is significant at p 5. 001with a beta of .170 and an
unstandardized coefficient of .106. Seeing an individual diabetes educator is also
significant at p 5.001 with an unstandardized coefficient of .169 and a beta of .109. The
data support hypothesis 2: When the provider models self-care behaviors the patient is

more likely to practice self-care behaviors.

In this model education has a significant effect on overall self-care behaviors with
a p=. 005 and a beta =. 094. The data support the hypothesis 5.2: Those who are better

educated are more likely to practice appropriate diabetes self-care behaviors.

This concludes the results for self-care behaviors and provider behaviors. In
summary the data support the hypothesis 1.1 for home blood glucose monitoring,
exercise, and foot examination, hypothesis 1.2 for influenza vaccination and retina
examinations and overall self-care behaviors, both home based and office based. The
data also support hypothesis 2 that provider modeling will increase diabetes self—care
behaviors. The data also support hypothesis 5.2 that those who are better educated,

practice more appropriate self-care.

91

behaviors. The data also support hypothesis 5.2 that those who are better educated,

practice more appropriate self-care.

The next analysis explores using diabetes education and the impact on self-care,
in this case the mean of blood glucose testing, exercising, checking feet, obtaining
influenza vaccine and retina examination. Diabetes education is the mean of attending a

diabetes support group and seeing a diabetes educator one on one.

Table 14: OLS Regression: The effect of patient characteristics and diabetes education on

 

diabetes self-care behaviors

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Variable: Aggregate of diabetes self-care behaviors n7 = 977
Variables Unstandardized Standardized t

_ Coefficient ”Coefficient . , _
Constant 1.396 11.373'"
Age 6.993E-03 .152 4.307‘"
Education 3.704E-02 .092 2704'
Black ethnicity -3.157E-02 -.039 -l .281
Other ethnicity -2.938E-03 -.004 -.151
Gender -1 .410E-02 -.013 -.429
Income in thousands 3.626E-04 .026 .725
Complications 1.698E-03 .003 . 105
Years with diabetes 9.607E-03 .157 4.934".
Patient used diabetes educator or .323 .140 4.721“.
support group
R2 .082
F value 10.553".
*p 5 0.05;
" p 5 .005

‘“ p 5 .001 (all probabilities are two tailed)
When we entered the patient characteristics, the R2 was .063; when use of
diabetes education was added the R2 was .082. Using diabetes education or a diabetes

support group is a significant factor in increasing appropriate self-care behaviors with a p

92

Next we will explore the effect of provider information giving on participation in

diabetes education activities.

93

Diabetes Education:

Hymthesis 3.1: People with diabetes who recall receiving information
regarding diabetes education are more likely to take advantage of diabetes

education.

The question asking, “Did you receive information about diabetes classes?” was

used as predictor. The dependent variable is diabetes education.

Table 15: Log Regression: Receiving information and use of diabetes education

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables B Exp(B) [Odds ratio] Wald ,
Constant -l .974 6.964"
Age -.014 .987 1.984
Education .130 1.139 2.374
Black ethnicity -.058 .943 .167
Other ethnicity -.253 .776 2.963
Gender -.451 .639 5207'
Income in thousands -.002 .998 .357
Complications index -.076 .927 .530
Years with diabetes .008 1.008 .416
Information about diabetes 1.029 2.797 18.250'"
education
Cox & Snell R2 .034‘“
Nagelkerke R2 .064’”
‘p 5 0.05;

" p 5.005

‘” p 5 .00] (all probabilities are two tailed)

When all patient characteristics were entered into the regression, the Nagelkerke
R2 was .027; when provider information giving was added it increased to .064. Likewise
the Cox & Snell R2 increased from .014 to .034 when provider information giving was

added. Receiving class information has a significant impact at p 5 .001 and odds ratio of

94

2.797 on participation in diabetes education. The data support the hypothesis 3.1:
Persons who have diabetes and receive information regarding available diabetes

education are more likely to use diabetes education.

This concludes any testing for self-care in relationship to provider information
giving, provider modeling and diabetes education. Table 15 summarizes the impact on

self-care when we add the different factors into the model.

Table 16: Summary of self-care and the impact of various factors

 

 

 

 

 

 

 

 

 

 

 

 

' Dependent Variable: Self-Care
Variables; entered k I h i P value of model, R square of model.
Demographics 6.872'” .037
Complications index 6.701 .042
Years with diabetes 7.553'” .054
Provider information 14564” .110
including diabetes education
Provider modeling 11.942' .103
*p 5 0.05;
" p 5.005

"W p 5 .001 (all probabilities are two tailed)

As the above table summarized, demographics and years since diabetes diagnosis
have a small but significant effect on self-care. Provider information giving and provider

modeling have a significant relatively strong effect on self-care behaviors.

Next we will investigate complications and their possible impact on seeing a

specialist and on practicing diabetes self-care.

95

Disease complications

Hymthesis 4.1: Those who are in the more advanced stages of diabetes

are more likely to report using the services of a specialist.

As mentioned earlier, for diabetes severity several complication measures have
been aggregated, such as history of heart attack, amputation, stroke, and coronary bypass
surgery. Years lived with diabetes was consistently associated with better self-care
behaviors in contrast to the complications index, which has a negative effect on self-care.

Therefore it was not aggregated in the complications index but kept separately.

Table 17: Logistic regression: The impact of patient characteristics and complications on

seeing a specialist

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I Dependent variable: Seeing a specialist in the last 12 month n =1073
Variables B Exp(B)/OR Wald

Constant -499 .846
Age .008 1.008 1.133
Education .204 1.226 10712“
Black ethnicity -.064 .938 .367
Other ethnicity -.071 .932 .698
Gender .020 1.020 .019
Income in thousands -7.l E-7 1.000 .099
Years with diabetes .013 1.014 2.236
Complications index .281 1.324 12.158".
Cox & Snell R2 .029‘”
Nagelkerke R2 .042”.
*p 5 0.05;

" p 5 .05

*" p 5 .001 (all probabilities are two tailed)

When all patient characteristics were entered into the regression, the Nagelkerke

R2 was .02; when provider information giving was added it increased to .04 and the

96

model is significant at p 5 .001. The presence of complications has a significant effect
although the odds ratio is only 1.3. Therefore those who have more complications are
more likely to see specialists than those who do not yet have developed complications
from diabetes. The data support the hypothesis 4.1: Those who are in the more advanced

stages of diabetes are more likely to report using the services of a specialist.

This concludes the testing of the relationships of self-care behaviors and
complications from diabetes. The next section (hypothesis 5) will investigate how
demographic factors influence provider information giving when we control for

complications and patient self-care behaviors.

97

Demographic variables app provider behaviors

In the previous sections we have focused on the diabetic person’s self-care
behaviors, and the relationship to provider behavior, diabetes education and
complications from diabetes. In this final analysis we will try to tease out how patient
factors such as age and education affect physician information giving and self-care
behaviors when we control for complications. Specifically we will look at age and its

impact on provider information giving.

Hypothesis 5.1: Those who are better educated are more likely to receive

information about self-care from their providers.

flvpothesis 5.3: Those who are older are less likely to receive information

regarding self-care from their providers.

The impact of age and general education level on provider information giving is
the subject of interest. In the first block the demographic variables and diabetic
complications were entered without education or age. In the second block education

level and age were entered.

98

Table 18: The impact of demographics on provider information giving

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent variable: Provider Information in .= 01070 8
Variables Unstandardized Standardized t
, , y , - p . Coefficient . Coefficient
Constant .861 11727”
Black Ethnicity 2.166E-02 .045 1.470
Other Ethnicity 2.116E-02 .054 1.790
Gender -6.025E-03 -.010 -.307
Income in thousands 2.54OE-O7 .031 .845
Complications index 2.997E-02 .098 3.093”
Years with diabetes 4.802E—O3 .135 4161‘"
Education -2.786E-O3 -.012 -.340
Age -4.404E-O3 -.164 -4.561
R2 .041
F value 5.726."
*p 5 0.05;
“ p 5.005

*** p 5 .001 (all probabilities are two tailed)

Education level was not significant at p = .734, the R2 did not increase by adding
the education level and the F value decreased from 6.533 to 5.726. The data do not
support this part of the hypothesis; education level does not appear to effect provider
information giving in this sample when we control for other demographic factors,

complications and years lived with diabetes. Hypothesis 5.1 is disconfirmed.

When we control for education, ethnicity, gender, years lived with diabetes and
complications, those who are older receive significantly less information with a beta of -
.164 and p 5 .001. Adding age to the patient characteristics increased the R2 from .023 to
.041, the model and the block are both significant at p 5.001. The data support

hypothesis 5.3 those who are older receive less information.

99

This concludes the presentation of the results of analysis and in the next section

we will discuss the results.

100

DISCUSSION OF THE FINDINGS

In this final section the paper will summarize the findings, explore considerations

and possible explanations for the findings.

Summagy pf self-care findings

Hypotheses addressing self-care:

Hyppthesis 1.1: People with diabetes who recall their provider’s advice to
engage in specific diabetes self-care activities (blood glucose testing,
exercise, foot examination) are more likely to report practicing that self-care

behavior.

Hypothesis 1.2: Those who receive more self-care advice from their

provider are more likely to engage in self-care behaviors.

Hypothesis 2.1:When the person with diabetes recalls that the provider
examined the home blood glucose log, the patient is more likely to monitor

his/her home blood glucose levels.

Hypothesis 2.2: When the person with diabetes recalls that the provider
examined his/her feet, the diabetic is more likely to report examining

his/her own feet at home.

Hyppthesis 2.3: When the person with diabetes recalls that the provider

modeled behaviors he/she is more likely to engage in self-care behaviors.

101

Hymthesis 3.1: People with diabetes who recall receiving information
regarding diabetes education are more likely to take advantage of diabetes

education.

Hypothesis 3.2: Those who receive diabetes education are more likely to

practice appropriate self-care behaviors.

Hypothesis 5.2: Those who are better educated are more likely to practice

appropriate self-care behaviors.

In general hypotheses 1.1 and 1.2 for self-care and provider information giving
are supported, diabetic patients who receive more advice regarding self-care are more
likely to practice appropriate self-care behaviors. Hypotheses 2.1, 2.2, and 2.3 for
provider modeling are also supported. When the provider models foot examinations
and/or blood glucose evaluation, the person with diabetes is more likely to practice those
self-care behaviors. Hypothesis 3.2 that those who receive diabetes education are more
likely to practice self-care behaviors is also supported. Hypothesis 5.2 that those who are
better educated practice better self-care behaviors is supported, the effect is strong and
significant for office based self-care and exercise, but not for home blood glucose
monitoring and foot care. This concludes the smnmary of the general results, next we

will summarize the findings for self-care behaviors.

Specific self-care behaviors

The following paragraph will summarize the findings itemized by specific self-
care behaviors and their hypothesized predictors. Home blood glucose monitoring as

hypothesized is strongly and significantly influenced by provider information giving as

102

well as by individual patient education. Provider evaluation of the blood glucose log and
modeling decision making based on the home blood glucose log is significantly and
moderately effective; seeing a diabetes educator who also reviews the log and models
decision making is significantly and strongly predictive of home blood glucose
monitoring. Education level has no significant influence on home blood glucose

monitoring.

Foot examination is significantly increased by provider information giving as well
as by individual diabetes education. The effect is stronger for provider information
giving than for diabetes education. Provider foot examination modeling also has a strong
and significant effect on the diabetic person’s foot monitoring. General education does

not have a significant effect on foot monitoring.

Provider information giving and individual diabetes education are significantly
and strongly related to exercise. Education level has a significant but weak effect on

increasing exercise time.

The relationships between obtaining flu vaccination and retina examination and
the provider information giving index are weak but significant. Education also has a

significant impact on obtaining a retina examination and influenza vaccination.

The provider information giving index is a significant and strong predictor for the
office based self-care index, seeing a diabetes educator is not a significant predictor for
office based self-care. Education level has a strong and significant impact on office
based self-care behaviors. The home based self-care index measure is significantly and

strongly predicted by diabetes education, provider information giving index but not by

103

education level. Diabetes education, the provider information giving index and education
level all have a significant and strong effect on the overall self-care index measure. In
summary hypothesis 1.1, 1.2, hypothesis 2.1, 2.2, hypothesis 3.2 and hypothesis 5.2 are

supported by the data.

Hypothesis 3.1 is also supported; people with diabetes who recall receiving
information about diabetes self-help classes are more likely to attend. The effect is

strong and significant.

This concludes the summary for self-care behaviors and provider information
giving; next the paper will summarize the findings addressing diabetes complications,

demographics and provider information giving.

104

Summary o_f diabetes complications Ed demographic findings:

Hymthesis 4: Those who are in the more advanced stages of diabetes are more

likely to report using the services of a specialist.

Hypothesis 4 is supported. Those who have more complications (aggregated in
the complications index) are more likely to see a specialist for their diabetes. That effect

is weak but significant.

Hyppthesis 5.1: Those who are better educated are more likely to receive

information on self-care from their providers.

Hypothesis 5.3: Those who are older are less likely to receive advice from

their provider.

Hypothesis 5.1 is disconfirmed. Those who are better educated may receive less
information from their provider rather than more, but the relationship is weak and not
significant. Hypothesis 5.3 is supported, those who are older are less likely to receive

information from their provider. That relationship is moderate and significant.

In conclusion, all of the hypothesis statements are supported by the data, with the
exception of hypothesis 5.1, those who are better educated pp po_t receive more
information from their providers when we control for other demographic factors,
complications, and years lived with diabetes. This concludes the summary of the

findings and next the paper will discuss data and methods considerations.

105

Data 3% methods considerations

The following section will consider data and measurement issues and their
possible effect on the results. Some data issues are rooted in the fact that the survey was
developed and the data collected before I became involved with this research project.
Therefore, some questions were not specifically worded to match the hypothesis

statements.

While the survey asked about some recommendations that the physician may have
made, it lacked specific questions asking if the provider had recommended office based
self-care behaviors, such as retina screening and obtaining influenza vaccination.
Therefore it is not surprising that the more general questions on physician behavior were
not as strong a predictor of these specific behaviors. The effect of physician behavior on
home based self-care behaviors is greater because all of the questions regarding provider
information giving were specific and addressed home based self-care (blood glucose

monitoring, foot inspections and exercise).

While physician behavior was a good predictor for these behaviors, we now
discuss why age was a strong predictor as well for these behaviors. Influenza vaccination
recommendations and retina examination standards are not limited to people who have
diabetes: There is a general recommendation to obtain the influenza vaccination for all
people who are age 55 and older and for people of any age who have heart disease or
lung disease or a compromised immune system. Therefore many community groups,
including churches and public service announcements encourage people to obtain the

influenza vaccinations as soon as they become available. The situation is similar for

106

retina examinations. As people age they become more likely to experience deterioration
of the retina and should have their retina examined. Likewise, those who have
hypertension (high blood pressure) are likely to develop microvascular diseases,
including retina detachment. Like diabetes, hypertension prevalence increases with age.
Thus the office based self-care behaviors are not specific to diabetes. That may effect the
importance of provider advice or the lack of provider advice secondary to the messages
from churches and community groups. That would be one possible explanation why age
had such a significant and strong effect on retina screening and eye examinations. The
age effect on influenza immunization and retina examination is large enough to impact

the aggregate measurement indexes for self-care.

As one looks at the unstandardized coefficients in the OLS regression tables, they
appear to be extremely small, yet they are significant. The reason for their small values
can be found in the summary table of the continuous variables. All of the index measures
are very small because they represent the mean of complications or self-care behaviors.
Complications and self-care behaviors, with the exception of exercise and retina
examination, were dichotomous measures. Those had a range from “0” to “1” and thus
the mean of home based and office based self-care behavior is close to one. The mean,
rather than the sum was used to create the index measures to avoid decreasing the sample
size due to the many missing measures. Therefore it is more helpful to look at the

standardized regression coefficients.

Now that we have considered the measurement issues we will examine and discuss some

of the factors for which we have controlled for in the analysis.

107

Background Variables

The background variables include use of a specialist, diabetes complications
index, and demographic variables. Utilization of a specialist was particularly interesting
because 71.1 % of the respondents stated that they had seen a specialist during the last
twelve months. This is a very high percentage compared to only 16.4% of a diabetic
sample in California that reported having seen a diabetes specialist (Simon et a1, 1999).
This high utilization of specialists could be important because several studies have
demonstrated that those who see a diabetes specialist receive better diabetes care (Simon
et a1 1999; Zgibor et a1, 2002). One of the possible explanations for the high rate of
specialist use may be found in M-CARE’s administrative structure and its affiliation with
the University of Michigan Medical School and its endocrinology department. M-
CARE’s Associate Medical Director is also the head of the endocrinology department
and the principal investigator for the TRIAD study from which the data for this analysis
came from. It is possible that people with diabetes that had seen the specialist felt
obligated or pressured to participate in the study, thus creating a selection bias. That is

one possible explanation for the high utilization of specialists.

There is another possible explanation that is rooted in the wording of the question.
The question did not specify the type of specialist seen and whether that specialist was a
diabetologist, it just asked: “. . .. did you see a specialist for your diabetes”. Depending
on how the subjects interpreted the question, the specialist could have been a podiatrist
for a foot problem or an ophthalmologist for the yearly retina examination rather than a

diabetologists. Yet podiatrists are not likely to address diabetes issues, other than the

108

feet, and ophthalmologists are not likely to teach about diabetes, but rather focus on their

expertise, the eyes.

Another possible explanation can be found in the link between M-CARE and the
University Health System. The Health System employs the specialists as well as half of
all primary care physicians who care for M-CARE’s members. In other managed care
organizations, the specialists and primary care physicians are seldom employed by the
managed care organization. That makes a referral to the specialist more costly for both
the primary care physician and the managed care organization; some of the managed care
organizations have disincentives for primary care physicians to refer to specialists. Thus
it is possible that M-CARE primary care physicians may refer more freely to specialists
because the specialists are part of the same system and there is no financial penalty for
the refenal. That explanation is supported by the fact that seeing a specialist was
dependent on seeing the primary physician for routine maintenance care. Those who saw
their primary care provider for routine care were more likely to be referred to a specialist.

Referral to a specialist was also dependent on diabetes complications.

Those who had more complications were more likely to see a specialist. However,
subjects who had emergency department use in the last twelve months (67.6%) were less
likely to see a specialist and the emergency care utilization in this sample was unusually
high. One would assume that a person who needs emergency medical treatment would
consult a specialist, either in the emergency room or as follow-up. In this sample this

does not appear to be the case.

109

One possible explanation for the high emergency room utilization may be found
in M-CARE’s fee structure that existed at the time the information was collected.
Members had a $15.00 copay for a scheduled office visit, and office appointments can be
difficult to get. Usually appointments are scheduled according to the physician’s
availability, rather than the patient’s needs. Emergency visits on the other hand are not
scheduled and can be planned on by the member and the copay was only $25.00. Thus,
the member could use emergency care at a convenient time for only an increase of $10.00
compared to a visit during regular office hours and a time that may have been
inconvenient. Therefore emergency department use may not be reflective of the subject’s
health status; rather it may be the result of the fee structure and availability of office

appointments and the patient’s motivation to see his/her own physician.

Thus the high utilization patterns of specialists and emergency centers could be
the outcome of the unique medical care that is provided in a medical school managed
care setting; the use of specialists and emergency centers may be more acceptable than in
other managed care organizations, because all are providers for the same organization.
Related to utilization patterns are diabetes complications. The next section will explore

the diabetes complications in this sample and the relationship to self-care.
Diabetes complications

As mentioned before, the complications from diabetes increase over time. The
microvascular changes are the first to occur but there are ofien no symptoms until the
advanced stages, when heart attack, stroke, retina damage, blindness or kidney disease

occur. Macrovascular changes occur over decades before “the major event” occurs.

110

Thus length of time since diabetes was diagnosed should be a strong predictor advanced
disease. However, the number of years lived with diabetes has a greater impact on self-
care behaviors compared to the impact of complications. The influence of years lived
with diabetes on self-care is consistently stronger than the actual complications. The
reason for this may be as follows, the longer a person lives with diabetes, the more time

there is to learn to live with diabetes and to adjust the life-style accordingly.

Vice versa, it could be that those who practice good self-care stay healthy and
therefore live longer with diabetes. Consistent with that theory, when the complications
were analyzed, we found that amputation, angioplasty, stroke or heart attack all had
negative effects on self-care. This is surprising because according to the Health Belief
Model, complications should increase the perceived threat from complications (another
heart attack or another amputation). The lack of increase in self-care behaviors after
complications may be explained by limited behavioral capacity', which is not considered
in the Health Belief Model. Many people who have had strokes are afraid of falling or
experience weakness or other physical limitations and therefore have a difficult time
exercising. Others who have heart disease or neuropathy may experience pain when they

exercise.

Similarly people who have developed retinopathy may have difficulty seeing and
reading their blood glucose levels, examining their feet or driving to the provider’s office.

Thus, it may be more difficult for the person with diabetes who has developed

 

‘ Behavioral capacity is the ability to perform a behavior. Behavior can be
limited by physical or psychosocial factors.

111

 

complications to practice recommended self-care. This causes a sad predicament for the
person with diabetes. The diabetic person has to practice self-care behaviors to prevent
complications. Once complications become symptomatic and the person with diabetes
becomes aware of the threat, he/she is limited in the ability to practice the required self-

care behaviors.

Another possible explanation for lack of increase of self-care behaviors once
complications have developed, may be the presence of depression. Once complications
set in, the quality of life decreases and depression may become more prevalent. Those
who are depressed are less likely to exercise and those who exercise are less likely to
become depressed because exercise produces brain chemicals that enhance the mood.
There are no depression indicators in this data set and there was no way to control for it.
It is however, known that people with depression have poorer health outcomes for almost
any chronic condition, such as heart failure, asthma, diabetes or even back pain. It would
be worthwhile in the future, to study the effect of depression on self-care because
depression can be treated. If depression is the underlying cause of lack of self-care, the
depression could be treated and self-care would improve. Another possible explanation
is that those who cannot motivate themselves to practice good self-care, develop
complications. The lack of self-care will result in higher blood glucose levels and those,
in turn, lead to a faster occurrence of complications, such as stroke, heart attack and
amputations. In that case, the lack of self-care would precede the complications and the
complications may just further de-motivate the person. Because this is a cross sectional

study, the data do not permit us to differentiate.

112

Demographics

At this point we have looked at the summary of findings, utilization (emergency
care use and seeing a specialist) and factors associated with diabetes complications.
Demographic factors are the last remaining group of background variables and will be
considered and discussed in he next section. As mentioned in the “results” section, it was
disappointing to see the sample to be mostly “white”. “White” populations have lower
diabetes prevalence and lower complication rates than other groups once diabetes has
developed. Yet 76.1% of the sample classified themselves as “white” even though
University Medical Centers that serve minority populations were oversampled. M-CARE
does not collect ethnic information on its membership and therefore it is difficult to say
whether the general membership has an appropriate ethnic composition. Because of the
lack of minority etlmicities in this sample the data do not show the ethnic differences that
have been found in other studies. Ethnicity was not significant in any of the analyses,

presumably because of lack of subjects in the non-white ethnicities.

Education was one of the more interesting demographic variables. It was a
significant factor in most of the analyses, with the exception of home blood glucose and
foot monitoring. One possible explanation is that subjects with more education may be
younger and exercise more, although we controlled for age, it could be that they are able
to control their blood glucose levels better and therefore worry less about their glucose
levels and neuropathy; those were the two behaviors that were not increased by education
level. When education was studied as dependent variable, the complications index was
significant and the slope negative. Thus, those who are better educated appear to have

fewer complications. It was surprising to see that those who have more formal education

113

received less information from their providers. Since we live in an age of the Internet,
providers may assume that well educated patients do not need information from their
providers, because they have the tools to find any information they need. That
explanation is supported by the data. When education was set as dependent variable,
getting information through the Internet had a strong and significant association to
education level. Thus the providers may not need to give the information, because the
person with diabetes may ask the questions that indicate that he/she has the needed
information. That would also explains why those who enjoyed more formal education
make better use of the medical care system, they not only know what is needed but have

the communication skills to access available services.

Age was another powerful demographic variable that had a great effect on the
dependent variables. Those who were older had significantly less education, had less
income and had more complications. Yet, those who are older are given less information.
It is difficult to discern whether this may be caused by the aged person's lack of short-
terrn memory or by the providers’ lack of information giving. Unfortunately, the data do
not tell whether those who are older receive less information or whether their ability to
recall the information is decreased. While there may be a recall problem, the provider
should be aware if bis/her patient has decreased short-term memory and should
reinforce important information with written materials in order to treat effectively.
There are simple instruments to assess mental acuity that can assist the provider in
identifying memory problems. It is the provider’s responsibility to provide

appropriate treatment including age appropriate information giving. If the patient

114

cannot remember what he/she was told to do, the patient cannot act upon the information

and the self-care behavior cannot be practiced.

Another possible explanation is that providers found it more difficult to
communicate with the older population, because they were older and had less education.
There is evidence that few physicians like working with the aged populations. There is a
shortage of geriatric specialists in the US. despite the fact that the population is aging.
One of the possible reasons may be that the treatment of the older population tends to be
less technical. People over age 65 are less likely to receive transplants or other
experimental treatments and may not wish to receive drastic life prolonging interventions.
Thus the provider will not be able to practice his/her high technical skills, which are
considered prestigious and therefore the old frail patient may not be very desirable from a
physician’s point of view. The other demographic variables did not have a strong or
significant impact on the dependent variables, thus this concludes the discussion of the
demographic factors and we will turn our attention to the differences in the self-care

behaviors.

115

The differences in self-care behaviors

As the data indicate, the self-care behaviors that are considered in this study are
rather different from each other in cost to the patient and in complexity of the tasks; they
should not be aggregated without considering the difference between the behaviors. The
office-based behaviors such as obtaining an eye examination and the influenza
vaccination are needed only once a year and do not require special skills and are not
painful for the diabetic person. The main requirement is the realization that the treatment
is needed and beneficial and the ability to access care. As a consequence, education is a
strong and significant factor for office based behaviors in this sample. Office based self-
care behaviors are not specific to the diabetic population, as mentioned earlier, age and
other conditions require the same treatment, therefore age was a significant predictor next

to provider information giving for office based self-care.

The home based self-care behaviors measured in this study include exercise,
blood glucose monitoring, and foot examination; they should be practiced every day.
They also differ in complexity and “cost” to the patient. It is important to consider the
cost and the complexity of the behavior, generally the simpler and the more convenient
the behavior the greater is the likelihood of adherence to the behavior. Exercise is one of
the more difficult behaviors, particularly for women (Waldron 1988). Exercise is not
specific to diabetes, it is also indicated for better mental functioning, preventing artery
disease, some of the cancers and maintaining a healthy weight. It was surprising to see
that the mode for exercise is “zero minutes per day”, 35.3% of this sample do not

exercise at all and only 37.2% exercise 20 minutes or more per day. Considering that

116

exercise is the foundation of diabetes treatment and the only treatment that can decrease
insulin resistance, more interventions are needed to support exercise behaviors in people
with diabetes. 56.7% of the population stated that they received “advice to exercise”.
However, the questions only talks about “advice” it does not give information on how
practical and how detailed that advice was. Usually complex behaviors need tailored
support to help the individual to achieve and maintain the complex behavior. Actually it
is encouraging that even the simple advice increases the desirable complex behavior. The
other 43.3% cannot even remember that they were told that exercise is part of diabetes
treatment. That is not medical practice according to evidence based clinical practice

standards.

Home blood glucose monitoring is also one of the “painful” and “complex”
diabetes self-care behaviors. It takes some dexterity to handle the strips and the monitor
and to obtain sufficient blood on the strip for accurate measurement. It requires cognitive
skills to evaluate the results and treat accordingly. It demands much motivation, because
it is not comfortable to penetrate the skin repeatedly. Although it is one of the more
costly behaviors, 80.9% of the sample stated that they monitor their glucose at home.
That is a good ratio for this behavior, because “most people with diabetes” but not

everyone needs to monitor their blood glucose at home.

Foot monitoring is a simple, low cost behavior. It does not hurt and it does not
require much time or skill to look at the sole of the foot to ascertain that there is no
injury. The question asked whether someone in the house checks the feet of a person
with diabetes, therefore a morbidly obese person who has problems seeing the feet may

ask help from a family member. Yet only 32.7% of the sample stated that they check

117

 

their feet. This is most surprising. The only explanation that is plausible is that the
people with diabetes are not aware of the risk for amputation secondary to infection and
neuropathy. According to standards, everyone should inspect their feet because the cost
is low and the benefit is the potential saving of the toe or even foot. This lack of
monitoring is reflected in Michigan’s 1,564 amputations for diabetes in the year 2001 at
the cost of $29.5 million. All of these are potentially preventable with proper foot care
(Michigan Department of Community Health, 2003). It would be worth researching why

foot monitoring is not more commonly practiced.

Considering the differences in costs and complexity of the self-care behaviors, the
next section of the paper will discuss the relationship of provider information giving and

provider modeling on those self-care behaviors.

118

 

Effect o_f provider information giving

The following section will discuss the relationship between provider information
giving and self-care behaviors. The data leave little doubt; provider information giving is
an important predictor for diabetes self-care. This is true for home blood glucose
monitoring, examining the feet and exercising - those behaviors had specific measures in
the survey. For home based self-care behaviors provider information giving was not only
significant but the effect was strong. The office based behaviors (flu vaccination and
retina examination) did not have a specific information giving question in the survey
nevertheless, provider information giving index was a significant predictor for office
based self care behaviors. The provider information giving index is the single most
important factor that increases influenza vaccination and retina examination. The effect
would probably have been greater if there would have been a question addressing office
based self-care information giving. The effect on the home based behaviors was large
enough to create a strong effect in the overall self-care index score. However, while
provider information giving influences self-care behaviors, better self-care behaviors
appear to lead to more provider information giving. There seems to be reinforcement
from provider to patient and when the patient practices appropriate self-care the provider
offers more information. This may point to the providers’ discomfort or unease and lack
of training in working with patients to change their life-styles. Thus, if the patient is
willing to practice appropriate self-care behaviors, then the provider may feel
comfortable giving more information regarding life-style and effect on the disease. It is
also possible that the underlying causal factor is the diabetic person’s motivation, i.e. the

patient who is motivated practices the proper self-care and may also request more

119

information. The question is whether the more motivated person with diabetes requests
diabetes information and therefore recalls information given by the provider and then is
able to practice more appropriate self-care, or whether the patient practices good self-care
and then the provider offers more information. The question can only be settled by a
longitudinal study that measures attitudes towards health and motivation at the beginning
of the study and then either randomly assigns subjects to diabetes education/provider
information giving group or controls statistically for motivation. Unfortunately this was
a cross sectional survey and there was no question in the survey that could shed light on

the subject’s motivation and whether or not the diabetic person requested information.

The data on diabetes educators was surprising. It appears that diabetes educators
are more effective with complex behaviors, such as blood glucose monitoring as
compared to inspecting the feet, a simple behavior. Particularly in the “Modeling table”
(Table 5) the diabetes educator appears to have the greater impact. Intuitively one would
expect that the individual educator focuses on the diabetic patient’s needs and spends on
the average one half to one hour with the diabetic person, trying to tailor instructions and
assist with behavior changes. That should have greater impact than mere information
giving. Providers and their office based staff do not have much time available; in most
cases the actual encounter with the physician lasts less than ten minutes and the
communication part may take only two to five minutes during any routine visit. One
possible explanation may be found in the relationship between the provider and the
patient. The provider and his/her staff usually have established rapport with the patient,
while the diabetes educator has to establish that rapport and trust in order to be effective.

Much of the time the diabetes educator spends with the patient is used to gather the

120

medical history and identifying the learning style of the diabetic person. Therefore the
actual teaching time may be less than anticipated. The office staff on the other hand has
all the information available from the medical record. Therefore tailoring and
personalizing to meet the medical need is less difficult ( i.e. “For you Mr. Jones, it is
particularly important that you exercise. Exercise can increase your HDL and your HDL
is too low”). This type of tailoring is thought to be most effective because it is relevant to
the patient. Another possible explanation is that the patient associates the provider
information giving and accommodates other physical conditions from the office staff as
“physician prescription”. The physician prescription is personal and carries more weight
than general information that may be provided by the diabetes educator. In addition,
considering that the provider directs the staff, there is little chance that there is conflict in
the message sent to the person with diabetes, while such conflict can happen with the
diabetes educator, particularly if the provider does not practice according to evidence
based treatment standards for diabetes. Again, this is an issue that would benefit from
further study, which behaviors can be modified by diabetes educators and which

behaviors are better addressed by the provider.

At this point we have looked at some of the findings and tried to explore them in
light of measurement issues and what is known about diabetes self-care behaviors. Next

we will try to examine the implications and limitations of the findings.

121

Implications

The implications for this study consider the importance of provider information
giving and opportunities to strengthen provider information giving. While information
giving by the provider and his/her staff is one of the most important predictors for self-
care behaviors, most physicians are not trained in behavior modification and nurses
receive limited training in behavior modification and education unless they are
specialized. Furthermore, physicians seldom are reimbursed for time-spent teaching or
counseling. Thus it would help to change the training of physicians as well as the
reimbursement regulations. The American Medical Association is currently considering
providing counseling education and experience during the academic and clinical training.
That would give new physicians the skill and confidence to provide effective counseling
and help with life-style changes. There are also changes in the expectations of the
regulating bodies. For example, the American Hospital Association is monitoring the
quality of patient education in hospitals. Coordinated education tailored to the patients’
specific abilities is required for accreditation. However, patients with diabetes who are
hospitalized may be too ill and in too much pain to be ready to learn; modeling is not
easily accomplished in the hospital. In managed care there are other problems. Health
insurers evaluate or rank each provider by information that can be obtained from claims
data. Thus they measure hemoglobin Arc testing but not necessarily whether the blood
glucose has been controlled. Counseling and teaching are not evaluated because the
information cannot be obtained from claims data, - providers do not bill for it because
they do not get reimbursed for it. Obtaining teaching and counseling information

requires costly medical record review. Thus current measures lack sophistication,

122

specificity and often accuracy. The evaluation process becomes useless for monitoring of
counseling and does not motivate providers who need feedback on all aspects of their

patient care.

Physician time is costly and employed physicians are expected to see as many as
forty patients a day in a hospital or managed care setting. It might be a better solution to
improve the training of the provider’s staff, such as medical assistants or nurses. The
barrier is that at this time hospital based nurses are pressed for time as well and due to
lack of time do not provide intensive patient education. Another solution would be to
create a communication vehicle that makes it easy for the provider to communicate with
the diabetes educator and direct the educator to address the patient’s most important
issues. Thus the message sent by the educator could be a direct reflection of the
provider’s concerns about health behaviors and could be targeted and tailored. However,
that requires that the physician refers the patient to a diabetes educator and that the
patient actually sees the educator to be helped. In this sample there were no monetary
problems for the diabetic person, but outside a managed care organization, most
insurances do not reimburse for diabetes education and Medicare reimburses only part of
it. Thus while diabetes educators can be effective, the reimbursement system does not
support them well. In Michigan the diabetes education programs lose money. Thus there

appears to be growing awareness of the problem, but the solution is not easy to find.

After considering the implications of the findings and possible solution, we need

to consider the limitations of the findings.

123

Limitations

The following section will consider the limitations of the findings due to sample
selection, and medical care provision that is unique to a managed care organization in a

university setting.

The sample came from a managed care organization in a university setting. Thus,
as discussed in the data section of the discussion, medical care and utilization may be

unique to the setting as well as the population in this study.

> The study was limited to people who are proficient in the English

language. Language barriers make it more difficult to provide information
and those who do not speak English receive less information (Simon et al.
1999). Thus any findings are limited to English speaking populations that
can read and comprehend. Lack of English skills is more common in
ethnic minority populations who also are at the greatest risk for diabetes
complications. The data do not permit us to draw any conclusions on how
provider communication affects ethnic minorities and their diabetes self-

care .

> The questionnaire did not contain a reliable question to evaluate adherence
to the medical nutrition plan. Since nutrition is a key-stone in diabetes

self-care, the lack of measure limits the data.

> This study has shown that the provider’s advice makes a substantial

difference in patients’ behaviors. However the population is this study has

124

generous insurance benefits with small copays on diabetes supplies or
office visits. Therefore the provider’s advice to monitor glucose or to
obtain flu vaccination or retina examination imposes minimal consumer
costs. When there is no insurance and great expense, the person with
diabetes may not be willing or able to pay the costs associated with self-
care. Advice and desire to practice appropriate behaviors does not remove

access barriers that exist in many other medical care settings.

> The subjects in this sample receive generous benefits and have no copays
for any diabetes education. They have access to diabetes case managers.
This is not true for all populations or for people who do not have health
care insurance. A recent unpublished study demonstrated that people with
diabetes are less likely to practice preventive behaviors when the cost is
greater. Herman et al (2003) found in an unpublished study that copays
deter from participation in diabetes education or blood glucose monitoring
programs. Therefore, the study may not apply to the general population

who has little or no coverage.

Thus, while the findings are important, there are difficulties in the health care
system that need to be resolved before provider/patient communications about needed

self-care can improve on an industry wide scale.

125

Conclusions

The conclusions for this paper will demonstrate that the findings of this study are
important; when researchers focus on patient adherence or lack of compliance, it is the
patient and the patient’s personal environment that are usually studied. That implies that
the patient or his/her personal environment bear the responsibility and need to be changed
to obtain adherence. Yet, the data suggest that much of the adherence depends on
provider counseling and information giving, possibly financial barriers and access to
medical care. Provider counseling is most powerful, the ten minutes spent in the
provider’s office impact the next six months of life-style even when we control for age,
education, income, ethnicity, and diabetes complications. For the amount of resources
spent that is an important impact, particularly since the questions did not require the
patient to remember specifics, only whether they were given information. The data leave
little doubt that part of the responsibility for the diabetic person’s self-care lays with the
providers and their information giving- their behaviors and adherence to evidence based

clinical practice guidelines.

The other issue is the lack of reimbursement by Medicare and traditional medical
insurers for diabetes education. Diabetes is a very complex disease and most of the cost
from complications could be avoided with better self-care behaviors. We know that
diabetes education increases desirable self-care behaviors, which in turn lower the
glucose levels. We know that good blood glucose control decreases and delays
complications. Yet we still blame the patient for lack of adherence rather than looking at

the medical structure and what could be changed to assist the person who struggles with

126

diabetes. The diabetic patient and his/her personal environment do not bear sole

responsibility for his/her health behaviors.

127

Bibliography

Acton K, Burrows N, Geiss L, Thomson T, 2003, “Diabetes prevalence among American
Indians and Alaska Natives and the overall population —United States 1994-2002,
Mortality Weekly 52(30)

American Diabetes Association, 2002, “The diabetes prevention program, description of
lifestyle intervention”, Diabetes Care, 25(12): 2165-21 72

American Diabetes Association, 2000, “Clinical practice recommendations”, The Journal
of Clinical and Applied Research and Education, 23, (suppl. 1)

American Diabetes Association, Expert Committee on the Diagnosis and Classification
of Diabetes Mellitus, 2001, “The report of the expert committee on the diagnosis
and classification of diabetes mellitus”, Diabetes Care, Vol. 24, 81, January 2001
pp 5:133.

American Association of Clinical Endocrinologists, 2001, “ACE consensus conference
on guidelines for glycemic control Endocrine Practice, Nov/Dec 2001

Anderson R, Fitzgerald J, Gorenflo D, Oh M, 1993, “A comparison of the diabetes-
related attitudes of health care professionals and patients”, Patient Education and
Counseling, 21, 41-50

Anderson R, Funnell M, Butler P, Arnold M, Fitzgerald J, Feste C, 1995, “Patient
empowerment: Results of a randomized controlled trial” Diabetes, 18:943-9.

Anderson R, Funnell M.; 2000, “Compliance and adherence are dysfunctional concepts in
diabetes care”, Diabetes Educator, 26:597-604.

Anderson L, Janes G, Ziemer D, Phillips L, 1997, “Diabetes in urban African Americans,
body image, satisfaction with size, and weight change attempts” The Diabetes
Educator, Vol. 23(3) pp 301-308

Anderson R; Hiss R; Stepien C; Fitzgerald J; Funnel M; 1994, “The diabetes education
experience of randomly selected patients under the care of community
physicians”, The Diabetes Educator, Vol. 20(5) pp 399-405

Anderson R, Donelly M, Dedrick R, 1990, ”Measuring the attitudes of patients towards
diabetes and its treatment”, Patient Education and Counseling, 162231-245

Angela M, Gneiss L, Minnie D, Arians C, Wagner E, Friedman N, Hurley J, Trinkets K,
1998, “Use of services by diabetes patients in managed care organizations.
Development of a diabetes surveillance system” Diabetes Care: 21(12) 2062-
2068

Annon, 1999, ”Standards of medical care for patients with diabetes mellitus”, Diabetes
Care; Vol. 22 (suppl. 1): $32: 41

128

Aubert R, Herman W, Waters J, Moore W, Sutton D, Peterson B, Bailey C, Koplan J,
1998, “Nurse case management to improve glycemic control in diabetic patients

in a health maintenance organization. A randomized, controlled trial”, Ann Intern
Med, 129(8): 605-612

Beck R, Daughtridge R, Sloane P, 2002, “Physician-patient communication in the
primary care office: A systematic review”, Journal American Board of Family
Practice, 15(1): 25-38

Becker M, “The Health belief model and personal health behavior”, Health Education
Monographs, 1974 (2) entire issue

Becker M, 1978. “Understanding patient compliance: The contributions of attitudes and
other psychosocial factors.” Cohen. New Directions in Patient Compliance

Becker M, Green L,l975, “A family approach to compliance with medical treatment: A
selective review of the literature”, International Journal of Health Education,
Vol. 18(3)

Beckles G, Engelgau M, Narayan K, Herman W, Aubert R, Williamson D, 1998,
“Population based assessment of the level of care among adults with diabetes in
the US”, Diabetes Care, 21(9) 1432-1438

Beresford S, Curr}l S, Kristal A, Lazovich D, Feng Z, Wagner E., 1997, “A Dietary
intervention in primary care practice: The eating patterns study”, American
Journal of Public Health, Vol. 87(4) 610:616

Blomgarden Z, 2001, “Clinical highlights in diabetes”, 61 " Scientific Sessions of the
American Diabetes Association, June 23, 2001

Boyer B, Lerman C, Shipley T, McBreartuh J, Quint A, Goren, 1996, “Discordance
between physicians and patient perceptions in the treatment of diabetes mellitus:
A pilot study of the relationship to adherence and glycemic control”, The
Diabetes Educator, Vol. 22(5) pp: 493-499

Brechner R, Cowie C, Howie L, Herman W, Will J, Harris M, 1993, “Ophthalmic
examination among adults with diagnosed diabetes mellitus”, Journal of the
American Medical Association; 270(14): 1714-1718

Brown SA. 1990, “Studies of educational interventions and outcomes in diabetic adults:
A meta-analysis revisited” Patient Education Counseling; 16(3): 189-21

Brownlee-Duffeck M, Peterson L, 1987 “The Role of health beliefs in the regimen
adherence and metabolic control of adolescents and adults with diabetes
mellitus”, Journal of Consulting and Clinical Psychology, Vol. 55(2) 139-144

Bucher M, Ragland D, 1995, “Socioeconomic indicators and mortality from coronary
heart disease and cancer: A 22-year follow-up of middle-aged men”, American
Journal of Public Health, September 1995, Vol. 85(9) pp. 1231 :45

Cabana, R, Powe N, et al., 1999, “Why don't physicians follow clinical practice
guidelines? A framework for improvement” [Review] JAMA. 282: 1458-65.

129

Cabana, Davis M, 2002, “Improving physician adherence to cholesterol management
guidelines”, Managed Care, Supplement,

Calahan E, Bertakis K, Azari R, Robbins J, Helms J, Chang D, 2000, “The influence of
patient age on primary care resident physician-patient interaction” Journal of the

American Geriatrics Society, 48:30-35

Caplan R, Van Harrison R, Wellons R, French J, 1980, Social support and patient
adherence, experimental and survey findings, Survey Research Center Institute
for Social Research, The University of Michigan

Carter J, Pugh J, Monterrosa A, 1996, “Non-insulin-dependent diabetes mellitus in
minorities in the United States Annals. 125:221-32.Notes: COMMENTS: 1996
Comment in: Ann Intern Med Aug 1; 125(3): 237-9

Centers for Disease Control, 1995, “Prevention and managed care: opportunities for
managed care organizations, purchasers of health care, and public health
agencies”, Morbidity and Mortality Weekly Review; 44(RR-l4)

Centers for Disease Control, 2002, “Prevalence of heath-care providers asking older
adults about their physical activity levels, United States, 1998 ” MMWR; 51(19),
412:4

Centers for Disease Control, 1997, “Diabetes surveillance”, Department of Health and
Human Services

Centers for Disease Control, 1997, “Diabetes-specific preventive-care practices among
adults in a managed care population-Colorado, Behavioral Risk Factor
Surveillance System”, Morbidity and Mortality Weekly Review, 46:1018-1023

Centers for Disease Control, 1999, “Diabetes in managed care work groups. Exploring
and expanding the research agenda for diabetes in managed care: a report of a
CDC-managed care workshop” Diabetes Care; 22(48) 1734-1738

Charlson M, Sax F, MacKenzie C, Fields S, Braham R, Douglas R, 1986, “Assessing
illness severity: Does clinical judgment work”, Journal Chronic Disease,
Vol.39:439-452

Cheng B, Chen P, Cheng D, Chu C, So H, 2002, Recall of preoperative anaesthesia
information in Hong Kong Chinese patients, Hong Kong Medical Journal, 8(3)
181:4

Ciechanowski P, Katon W, Russo J, Walker E, 2001. “The Patient-Provider Relationship:
Attachment Theory and Adherence to Treatment in Diabetes”, American Journal
of Psychiatry Vol. 158(1): 29-35

Clark N, Dodge J. 1999, “ Exploring self-efficacy as a predictor of disease management”,
Health Education Journal 26:72-89

Clemenhagen C, 1994, “Patient satisfaction: The power of an untapped resource”,
Association Medical Canadienne, Vol. 150(11), pp 1771 :72

Clark N, Gong M, 2000, “Management of chronic disease by practitioners and patients:
are we teaching the wrong things?” [Review] British Medical Journal; 320:5 72-5

130

Clark N, Janz N, Dodge J, et al., 2000, “Changes in functional health status of older
women with heart disease: evaluation of a program based on self-regulation”
Journals; 55:81 17-26

Colberg S, 1999, “Use of clinical practice recommendations for exercise by individuals
with type 1 diabetes”, Journal of the American Association of Diabetes
Educators, Vol. 26, 265-2 71

Conrad B, Betschart J, Harris M, 1994, “Frequency and determinants of diabetes patient
education among adults in the US. population”, Diabetes Care; 17(8)852-858

Cowie C et a1, 2003, “CDC Surveillance Summary”, Morbidity and Mortality Weekly
Report, Summer52z833-837

Crim C, 2000, “Translating guidelines into practice: Implementation and physician
behavior change, clinical practice guidelines vs. actual clinical practice: The
asthma paradigm Chest, Vol. 118(2)

Cummings M, Becker M, Maile M, 1980, “Bringing the models together: An empirical
approach to combining variables used to explain health actions ”, Journal of
Behavioral Medicine, Vol. 3(2) pp. 123:45

Dan A, 1994, Refiaming Women ’s Health, Sage Publications, Thousand Oaks, London,
New Delhi

Daniel E, 1993. “Taking Sides, clashing views of controversial issues in health and
society ” Guilford, Ct. The Dushkin Publishing Group

Dean K, 1989,“Self care components of lifestyles: The importance of gender, attitudes
and the social situation”, Social Science Medicine, Vol. 29, pp. 137:152, 1989

Detmar S, Muller M, Wever L, Schomagel J, AaronsonN, 2001, “The patient-physician
relationship. Patient-physician communication during outpatient palliative

treatment visits: an observational study”, Journal of the American Medical
Association, Vol. 285(10) 135 1-7

Deyo R, 2001, “New methods improve treatment”, Health Care Quality Special Reports,
Winter 2001, pp 11:16

Deyo R, Cherkin D, Ciol M, 1992, “Adapting a clinical comorbidity index for use with
ICD-9-CM administrative databases” Journal Clinical Epidemiology; 45(6): 613-
619

Diabetes Control and Complications Trial Research Group, 1993, “The effect of intensive
treatment of diabetes on the development and progression of long-term diabetes
mellitus”, New England Journal of Medicine, 329:977

Donohue M, 1998, “Comparing generalist and specialty care: Discrepancies, deficiencies
and excesses.” Archives Internal Medicine. Vol. 158(15): 1596-608

Dube C, O’Donnell J, Novack D, 2000, “Communication skills for preventive
interventions”, Academic Medicine 75: $45-54, 2000

131

Duncan K, Prozehl B, 2002, “Staying on course: The effects of an adherence facilitation
intervention to home exercise participation”, Progress in Cardiovascular Nursing,
1 7(2):59-65, 2002

Eckert T, Junker C, 2001, “Motivation for smoking cessation, what role do doctors
play?” Swiss Medical Weekly, 131:521-526

Evangelista L, Dracu K, 2000, “A closer look at compliance research in heart failure
patients in the last decade, Progress in Cardiovascular Nursing, (15 (3): 97-103

Fain J, Nettles A, Funnel M, Chadron D, 1999, “Diabetes patient education research: An
integrative literature review”, The Diabetes Educator, Vol. 25, (6): 7-16

F eldman D et al., 1998. “Effects of managed care on physician-patient relationships,
quality of care and the ethical practice of medicine: A physician survey.” Archives
of Internal Medicine Vol. 158:1626-1632.

Fitzgerald U, Davis W, Connell C, Hess G, Funnell M, Hiss R, 1996, “Development and
validation of the diabetes care profile”, Evaluation & the Health Profession, Vol.
I 9( 2) pp208-230

Fitzgerald J, Funnell M, Hess G, et al., 1998, “The reliability and validity of a brief
diabetes knowledge test”, Diabetes; 21 :706-10.

Fitzgerald J, Anderson R, Funnel M, Jacober S, Grunberger G, 1997,“Differences in the
impact of dietary restrictions on African Americans and Caucasians with
NIDDM, The Diabetes Educator,_Jan/Feb. 1997, Vol. 23(1) pp. 41-4

Fleming M et al. Al., 2002, “Problem drinkers benefit from brief physician advice”,
Brown University Digest of Addiction Theory & Application 21(4): 1:6, 2002,
Minisses Communication Group Inc.

Freeman H, Levine 8,1989, Handbook of Medical Sociology, Englewood Cliffs, New
Jersey Prentice Hall

Geffner D, 1992, “Diabetes care in health maintenance organizations” Diabetes Care 15
Supplement 1:44-50

Genel M, 1997, “Socioeconomic status, health, and health systems”, Pediatrics 99(6)
888:9

Gilrnet P, 1999, “Diabetes disease management in managed health care plans”, Drug
Benefit Trends, 11(1 lsb): 35-42, Cliggot Publishing

Glanz K, Lewis M, Rimer, B, 1999, “Health Behavior and Health Education, Theory
Research and Practice” Jossey-Bass, Publisher, San Francisco

Glanz K, Rimer B, Lewis M, 2002, “Health Behavior and Health Education, Theory
Research and Practice; Third Edition ” J ossey-Bass, Publisher, San Francisco

Glasgow R, Boles S, Calder D, Dryfoos J, Bagdade J, 1999,”Diabetes care practices in
primary care: Results from samples and three guidelines”, Diabetes Care, 25: 755-
763

132

Glasgow R, Fisher B, Anderson B, LaGreca A, MarreroD, Johnson S, Richard R, Cox D,
1999, “Behavioral science in diabetes: contributions and opportunities”, Diabetes
Care, 22(5): 832-843

Glasgow R, Strycker L, 2000 “Level of preventive practices for diabetes management:
patient, physician, and office correlates in two primary care samples” American
Journal of Preventive Medicine, 19:9-14

Glasgow R, Hampson S, Strycker L, Ruggiero L,1997 “Personal-model beliefs and
social-environmental barriers related to diabetes self-management”, Diabetes;
20:556-61

Glasgow R, Shawn B, Calder D, Dreyer L, Babdade J, 1999, “Diabetes care practices in
primary care: Results from two samples and three measurement sets”, The
Diabetes Educator, Vol. 25(5) 755- 763

Glasgow R, Toobert D, Hampson S, 1996, “Effects of a brief office-based intervention to
facilitate diabetes dietary self-management”, Diabetes Care: 19(8)835-842

Glasgow R, Wagner E, Kaplan R, Vinicor F, Smith L, Norman J, 1999, “If diabetes is a
public health problem, why not treat it as one? A population-based approach to
chronic illness” [Review]. Annals 21 :159-70

Glasgow R, Hiss R, Anderson R, Friedman N, Hayward R, Marrero D, Taylor B, Vinicor
F, 2001,”Report of the health care delivery work group: Behavioral research
related to the establishment of a chronic disease model for diabetes care”,
Diabetes Care 24(1): 214-130

Gochman D,1988, Health Behavior, Emerging Research Perspectives, New York,
Plenum Press

Graber A, Elasy T, Quin D, Wolff K, Brown A, 2002, “Improving glycemic control in
adults with diabetes mellitus: Shared responsibility in primary care practices”,
Southern Journal of Medicine, 95(7) 684-690

Green L, Ottoson J, 1994, Community Health, Seventh Edition. Chicago, Mosby

Green L, 1980, The Professional and Scientific Literature on Patient Education, Gale
Research Company, Detroit, Michigan

Greene G, Rossi S, Reed R, Willey C, Prochaska J, 1994, “Stages of change for reducing
dietary fat to 30% of energy or less ”, Journal of the American Dietetic
Association, Vol. 94(10) pp. 1105-10

Greenfield S, Nelson E, Zubkoff M, Manning W, Rogers W, Kravitz R, Keller A, Tarlov
A, Ware J Jr., 1992, “Variations in resource utilization among medical specialties
and systems of care. Results from the medical outcomes study”, Journal of
American Medical Association; 267(12) 1624-1630

Greenfield S, Roger W, Mangotich M, Carney M, Tarlov A, 1995, “Outcomes of patients
with hypertension and non-insulin dependent diabetes mellitus treated by different

systems and specialties. Result from the medical outcomes study , , Journal of
American Medical Association; 247(18): 1436-1444

133

Grimshaw J, Russell 1, 1993, “Effect of clinical guidelines on medical practice: a
systematic review of rigorous evaluations”. Lancet: 342(8883) 1317-1322

Hall J, Palmer H, Orav J, Hargraves J, Lee M, Writh E, Louis T, 1990, “Performance
quality, gender, and professional role”, Medical Care, Vol. 28(6) pp.489 501

Hall J, Irish J, Roter D, Ehrlich C, Miller L, 1994, “Satisfaction, gender, and
communication in medical visits”, Medical Care Vol. 32(12): 1216-1231

Hall J, Stein T, Roter D, Rieser N, 1999, “Inaccuracies in physicians’ perceptions of their
patients”, Medical Care, 3 7(1 1): 1 1 64-1 1 68

Harris M, 1995, Summary: National diabetes data group, Diabetes in America. 2nd ed.
Washington DC: Government Printing Office: 1-14 (NIH publication no. 95-
1468)

Harris M, Cowie C, Howie L, 1993, “Self-monitoring of blood glucose by adults with
diabetes in the United States population”, Diabetes Care; 16(8): 1116-1123

Harris M, Eastman R, Cowie C, Flegal K, Eberhardt M, 1999, “Racial and ethnic
difference in glycerrric control of adults with type 2 diabetes”, Diabetes Care;
22(3): 403-408

Harris M, 1993, “Undiagnosed NIDDM: clinical and public health issues, Diabetes Care;
16(4) 642-652

Hays R, Shaul J, William V, Lubalin J, Hasrris Jojetin L, Sweeny S, Cleary P, 1997,
Psychometric properties of the CAI-IPS 1.0 Survey measures. Consumer
Assessment of Health Plans Study. Med Care; 37 (Suppl. 32M822-31)

Hayes-Bautista, 1976, Modifying the Treatment: Patient Compliance, Patient Control and
Medical Care, Social Science Medicine

Heissler M, Bouknight R, Rodney H, Smith D, Kerr E, 2002,”The Relative importance of
physician communication, participatory decision making, and patient
understanding in diabetes self-management”, Journal of General Internal
Medicine, 17(4): 243-252

Hiss R, 1996,” Barriers to care in non-insulin-dependent diabetes mellitus. The Michigan
experience, Annals Internal Medicine 124 (1 ): 146-148

Ho M, Marger M, Beart J, Yip I, Shekel] P, 1997, “Is the quality of diabetes care better in
a diabetes clinic or in a general medicine clinic?” Diabetes Care; 20(4): 472-475

Holman H, Lorig K, 2000, “Patients as partners in managing chronic disease. Partnership
is a prerequisite for effective and efficient health care” [editorial; comment]
British Medical Journal: 320:526-7.

Notes: COMMENTS: Comment on: British Medical Journal, 2000 Feb 26;
320(7234): 550-4, Comment on: British Medical Journal 2000 Feb 26; 320(7234):
569-72

Hooper J, Veneziano L, 1995, “Distinguishing starters from nonstarters in an employee
physical activity incentive program” Health Education Quarterly, Vol. 22(1),
49:60

134

Howard-Century C, 2001 “Counseling, persistence, and multiple telephone calls along
with low-literacy level educational material impact screening for diabetic
retinopathy in inner-city African American communities” Department of Health
and Behavior Studies, Center for Health Promotion, Columbia University
Teachers College, unpublished presentation at 129th Annual meeting of the
American Public Health Association

Howard-Century, 2001,“ Prevention of vision loss from diabetes”, Presentation 129th
APHA Annual Meeting October 21-25, Atlanta Georgia

Hunt L, Pugh J, Valenzuela M, 1998, “How patients adapt diabetes self-care
recommendations in everyday life”, [see comments]. Journal of Family Practice
Medicine; 46:207-15.Notes: COMMENTS: 1998 Cement in: Journal of Family
Practice Medicine; 46(3): 205-6

Hunt L, Valenzuela M, Pugh J, 1998, “Porque me toco a mi? Mexican American diabetes
patients' causal stories and their relationship to treatment behaviors” Social
Science Medicine. Vol. 46: 959-69.

Huggins C, 2003, “Patients less likely to heed advice of obese physician”, Preventive
Medicine, 36:41-44

Hurley S, Huggis R, Jolley D, Reading D, 1994, “Recruitment activities and
sociodemographic factors that predict attendance at a marnmographic screening
program”, American Journal of Public Health, October 1994, Vol. 84(10) pp.
1655:61

Jacques C, Jones R, Houts P, Bauer L, Dwyer K, Lynch J, Casle T, 1991, “Reported
practice behaviors for medical café of patients with diabetes mellitus by primary-
care physicians in Pennsylvania”, Diabetes Care; 14(8): 712-717

Janz and Becker, ”The Health Belief Model: A decade later”, Health Education
Quarterly, 1984 (11) 1-47

Johnson J, 1997, “Self-efficacy theory as a framework for community pharmacy-based
diabetes education programs”, The Diabetes Educator, 22(3): 237-241

Kandrack M, Grant K, Segall A, 1991, “Gender differences in health related behavior:
Some unanswered questions”, Social Science Medicine, Vol. 32(5) 579-590

Kaplan S, Greenfield S, Ware J, 1989 “Assessing the effects of physician-patient
interactions on the outcomes of chronic disease [published erratum appears in
Med Care Vol.27 (7): 679]. Medical Care: 27:8110-27.

Keeler E, Sloss E, Brook R, Perskalski B, Goldberg G, Newhouse J, 1987, “The effects
of sharing on physiological health, health practices and worry” HSR; 22(279-306

Kenny S, Smith P, Goldschmid M, Newman J, Herman W, 1993, “Survey of physician
practice behaviors related to diabetes mellitus in the US. Physician adherence to
consensus recommendations” Diabetes Care; 16(11): 1507-1510

Kolturn A, Stone G, 1986, “Past and current trends in patient noncompliance research:
Focus on diseases, regimens — programs, and provider-disciplines”, The Journal
of Compliance in Health Care, Vol.1 (1) pp 21-31

135

Kraft S, Marrero D, Lazaridis E, Fineberg N, Qiu C, Clark C, 1997, “Primary care
physicians’ practice patterns and diabetic retinopathy. Current levels of care”,
Arch Fam Med; 6(1): 29-37

Kralewski J, Rich E, Bernhardt T, Dowd B, Feldman R, Johnson C, 1998, “The

organizational structure of medical group practices in a managed care
environment”, Health Care Management Rv; 23(2): 76-96

Kralewski J, Wingert T, Barbouche M, 1996. “Assessing the culture of medical group
practices.” Medical Care, Vol. 34(5): 377-388

Kravitz R, Hays R, Sherbourne C, DiMatteo R, Rogers W, Ordway L, Greenfield S,
1996, “Recall of recommendations and adherence to advice among patients with
chronic medical conditions.” Archives of Internal Medicine, Vol. 153(16):

1 8691 878

Kravitz R, Linn L, Shapiro M, 1990, “Physician satisfaction under the Ontario Health
Insurance Plan”, Medical Care, Vol. 28(6) 502- 512

Kreuter M, Strecher V,1995, ”Changing inaccurate perceptions of health risk: Results
from a randomized trial ” Health Psychology, Vol. 14(1):56:63

Kurtz M, 1990, “Adherence to diabetes regimens: empirical status and clinical
applications”, Diabetes Educator 16:50-56

Laine C, Davidoff F, 1996, “Patient-centered medicine. A professional evolution” [see
comments]. Journal of American Medical Association 275: 152-6.Notes:
COMMENTS: JAMA 1996 Jan 10; 275(2): 147-8, Cement in: JAMA 1996 Apr
17; 275(15): 1156; discussion 1157, Comment in: JAMA 1996 Apr 17; 275(15):
1156-7

Landrine H, Klonoff E, 1992, “Culture and health-related schemas: A review and
proposal for interdisciplinary integration”, Health Psychology, Vol.11 (4) 267-
276

Larme A, Pugh J, 1998, “Attitudes of primary care providers toward diabetes: barriers to
guideline implementation”, Diabetes Care 21(9): 1391-6

Larsson J, Aperlquist J, 1995, “Towards less amputations in diabetic patients. Incidence,
causes, cost, treatment, and prevention — a review. Act Orthop Scand; 66(2): 181-
192

Leaf P, Livingston M, Tischler G, Freeman Jr. D, Weissman M, Myers J, 1988, “Factors
affecting the utilization of specialty and general medical mental health services”,
Medical Care, Vol. 26(1) 9-26

Legoretta A, Peters A, Ossorio R, Lopez R, Jartulis D, Davidson M, 1996, “Effects of a
comprehensive nurse managed diabetes program: an HMO prospective study”,
American Journal of Managed Care; 2: 1024-1030

Ley P, 1986,“Cognitive variables and noncompliance”, The Journal of Compliance in
Health Care, Vol. 1(2) pp. 171:188

Levin L; Katz A; Holst E; 1976, “Self Care: Lay Initiatives in Health, New York: Prodist

136

Lipkin M, 2001, “New findings and approaches in patient-physician communication”,
25’” Annual Meeting of the Society of General Internal Medicine

Lipton R, Losey L, Giachello A, Mendez J, Girotti M, 1998, “Attitudes and issues in
treating Latino patients with type 2 diabetes: views of healthcare providers”,
Diabetes Educator, 24(1): 6 7- 71

Lorig K, 1996, “Patient Education: A Practical Approach Sage Publications, Thousand
Oaks, London and New Delhi

Lorig K, Sobel D, Stewart A, et al., 1999, “Evidence suggesting that a chronic disease
self-management program can improve health status while reducing
hospitalization: a randomized trial”, Medical Care; 37:5-14.

Lorig K, Sobel D, Ritter P, Laurent D, Hobbs M, 2001, “Effect of a self-Management
program on patients with chronic disease”, Efl'ective Clinical Practice, Nov/Dec
2001, Vol. 4(6) 256-262

Lupton D, 1994, “Memory and meaning the symbolic and social nature of food events”,
Sociological Review, Vol.42 #4, pp 664:722

Lutfey K, Wishner W, 1999, “Beyond "compliance" is "adherence". Improving the
prospect of diabetes care. [Review] [44 refs]. Diabetes Care.22: 635-
9COMMENTS: Diabetes Care 1999 Dec; 22(12): 2090-2

M-CARE, 2001, Making A Healthy Diflerence: An Overview of HEDIS 2000 Results, M-
CARE Web Page

Maillet N, D’Eramo M, Spollett G, 1996, “Using focus groups to characterize the health
beliefs and practices of black women with non-insulin-dependent diabetes”,
Diabetes Educator; Vol.22 (1): 39-46

Marshall C, Bluestein M, Chapin C, Davis T, Gersten J, Harris C, Hodgin A, Larsen W et
a1. 1996, “Outpatient management of diabetes mellitus in five Arizona Medicare
managed care plans”, Am J Med Qual; 11(2): 87-93

Marshall C, Bluestein M, Brier E, Chapin C, Darling B, Davis K, Davis T, Gersten J,
Harris C, Hodgin A, Larsen W, Mabb D, Rigberg H, Watson D, Krishnast V,
2000, “Improving outpatient diabetes management through a collaboration of six
competing, capitated Medicare Managed Health Plans”, American Journal of

Medical Quality, Vol 15(2)pp 65:72

Martin T, Selby J, Zhang D, 1995, “Physician and patient prevention practices in
NIDDM in a large urban managed-care organization”, Diabetes Care; 18(8):
1124-1132

Maxwell A, Hunt 1, Bush M, 1992, “The effects of social support groups as an adjunct to
diabetes training on metabolic control and psychosocial outcomes”, The Diabetes
Educator; 18:303-3-9

Mayfield J, Rith-Najarian S, Acton K, Schraer C, Stahn R, Johnson M, Gohdes D, 1994,
“Assessment of diabetes care by medical record review. The Indian Health
Service Model”, Diabetes Care; 17(8): 918-923

137

McCulloch D, Glasgow R, Hampson S, Wagner E, 1994, “A systematic approach to
diabetes management in the post-DCCT era”, Diabetes Care; 17(7): 765-769

McCulloch D, Price M, Hindmarsh M, Wagner E, 1998, “A population based approach to
diabetes management in a primary care setting: early results and lessons learned”,
EflClin Pract; 1(1): 12-22

McGlynn E; Asch S; Adams J; Keesey J; Hicks J; DeCristofaro A; Kerr E; 2003’ “The
Quality of Health Care Delivered to Adults in the United States”, The New
England Journal of Medicine, Vol 348 (26) 2635:2645

McNabb W, 1997,”Adherence in diabetes: can we define it and can we measure it?”
[Review] [30 refs]. Diabetes Care; 20:215-8

Michigan Department of Community Health, 2003, Michigan Diabetes Strategic Plan,
October 2003, Kimberlydawn Wisdom, MD

Michigan Department of Community Health, 2002, Diabetes in Michigan, May 2002
Fact Sheet

Michigan Department of Community Health 2000, Health Risk Behaviors 2000

Moci K, 2001, “Effective clinician-patient communication”, Physician News Digest, 2(1)

Mottur-Pilson C, Snow V, Bartlett K, 2001, “Physician explanations for failing to comply
with ‘Best Practices”, Eflective Clinical Practice, Vol. 4(5) 207-213

Nasr C, Hoogwerf B, Fairman C, et al. 1999, “United Kingdom Prospective Diabetes
Study, Effects of glucose and blood pressure control on complications of type 2
diabetes mellitus”, Cleve Clinical Journal of Medicine, 66:247-253

National Committee for Quality Assurance, 2002, State of Managed Care: Quality 2001
Comprehensive Diabetes Care

National Forum of Quality Improvement in Health Care, 2001, Improving Care for
People with Chronic Conditions, 13‘11 Annual Meeting, Orlando Florida

National Health Information LLC, 2001, “Diabetes, disease management strategies &
programs, Volume II”, 2001 National Health Information, LLC, Atlanta, GA
30333-0429

Nichols-English G, Poirier S., 2000, “Optimizing adherence to pharmaceutical care
plans”, Journal American Pharmaceutical Association, 40(4): 475-485

O’Connor P, Rush W, Peterson J, Morben P, Chemey L, Keogh C, Lasch S, 1996,
“Continuous quality improvement can improve glycemic control for HMO
patients with diabetes”, Arch F am Med; 5(9): 502-506

Oung-Hyman D, 1999, “Provider impact in diabetes education”, The Diabetes Educator,
Vol. 25, suppl. November/December pp34-43

Pattern K, White B, 1996,“Do beliefs, knowledge, and perceived norms about diet and
cancer predict dietary change?” American Journal of Public Health, Vol. 86(10)
1394:1400

138

Phillips D, 1996, “Medical professional dominance and client dissatisfaction”, Social
Science Medicine, Vol. 42(10) 1419-1425

Phillips L, 1988, Patient Compliance, New light on Health Delivery Systems in Medicine
and Psychotherapy, Hans Huber Publishers, Lewiston, New York

Plotnikoff R, Brez S, Hotz S, 2000 “Exercise behavior in a community sample with
diabetes: understanding the determinants of exercise behavioral changes”,
Diabetes Educator, 26(3):450-459

Prochaska J, Velicer W, Rossi J, Goldstein M, Bess H, Rakowski W, Fiore C, 1994,
“Stages of change and decisional balance for 12 problem behaviors”, Health
Psychology Vol. 13(1) 39:46

Pumilia C, 2002, “Psychological impact of the physician-patient relationship on
compliance: A case study and clinical strategies”, Program Transplant, 12(1),
1 0-1 6, 2002

Rask K, Kohler S, Wells K, Williams J, Diamond C, 2001, “Performance improvement
interventions to improve delivery of screening services in diabetes care”, JCOM,
Vol.8 (11) pp23: 29

R1220 J, 1993, “Physician uncertainty and the art of persuasion Social Science
Medicine, Vol. 37(12) pp 1451-1459

Rosenstock I, 1974, “Historical origins of the Health Belief Model”, Health Education
Monographs, 1974(2), 328-335

Roter D, 1982, “Participation in the Patient-Provider Interaction: The Eflects of Patient
Question Asking on the Quality of Interaction, Satisfaction, and Compliance” The
SOPHE Heritage Collection, Volume3, Research and Evaluation in Health
Education: Patient

Roter D, Hall J, Doctors Talking with Patients/Patients Talking with Doctors: Improving
communication in medical visits, Auburn House, Westport, Connecticut, 1992

Roter D, Hall J, 1994, “Strategies for enhancing patient adherence to medical
recommendations”, Journal of the American Medical Association, 271(1) 80-81

Roter D, Hall J, Rolande M, Nordstrom B, Cretin D, Svarstad B, 1998, “Effectiveness of
interventions to improve patient compliance: A Meta-Analysis”, Medical Care,
36(8) 1138-1161

Roter D, Hall J, 2001, “How physician gender shapes the communication and evaluation
of medical care”, Editorial, Mayo Clinic Proceedings, 76(7). 673-676

Roter D, Hall J, Aoki Y, 2002,”Physician gender effects in medical communication: A
meta review”, Journal of the American Medical Association 288(6): 75 6- 764

Russel N, Roter D, 1993, “Health promotion counseling of chronic —disease patients
during primary care visits”, American Journal of Public Health, Vol. 83 (7),
979:988

Sackett D, Haynes B, 2000, “Compliance with Therapeutic Regimens John Hopkins
University Press, Baltimore and London

139

Scott A, Shiell A, King M, 1996, “Is general practitioner decision making associated with
patient socio-economic status?” Social Science Medicine, Jan 42(1) pp35-46

Sega] A, Goldstein J, 1989, “Exploring the correlates of self-provided health care
behaviour”, Social Science Medicine, Vol. 29 (2) pp. 153: 161,

Shannon J, Kirkley B, Ammerman A, Keyserling T, Kelsey K, DeVellis R, Simpson R,
1977,”Self-efficacy as a predictor for dietary change in a low-socioeconomic-
status southern adult population,” Health Education Quarterly, Vol. 24(3)
35 7:368

Silagy C, Stead L, 2002 “Physician advice for smoking cessation”, Cochrane Review,
Vol 2, (2)

Simon L, Albright A, Belrnan M, Tom B, Rideout J, (1999), “Risk and Protective Factors
Associated with screening for complications of diabetes in a health maintenance

organization setting”, Diabetes Care 22(2)208-212

Skelly A, Marshall J, Haughey B, Davis P, Dunford R, 1995, “Self-efficacy and
confidence in outcomes as determinants of self-care practices in inner-city,
African-American women with non-insulin-dependent diabetes”, Diabetes
Educator 21(1) 38-46

Sorlie P, Backlund E, Keller J, 1995, “US. mortality by economic, demographic, and
social characteristics: The National Longitudinal Mortality Study”, American
Journal of Public Health, Vol. 85 ( 7) pp949-956

Suchman E, 1962, “Stage of illness and medical care”, Journal of Health and Human
Behavior. pp] 1 4: 128

Stewart A, Napoles-Springer A, Perez-Stable E, 1999, “Interpersonal processes of care in
diverse populations”, Milibank. 77: 305-39, 274.

Szasz, T., and Hollender, M, 1956, A Contribution to the philosophy of Medicine: The
Basic Models of the Doctor-Patient Relationship. American Medical Association
Archives of Internal Medicine

Taveras E, Capra A, Braveman P, et al., (2003), “ Clinician support and psychosocial risk
factors associated with breastfeeding discontinuation” Pediatrics 112(1) pp. 108-
115

Terry P, Healey M, 2000, “Physician’s role in educating patients: A comparison of
mailed versus physician-delivered patient education”, The Journal of Family
Practice, Vol. 49(4), 314:318

The Diabetes Control and Complications Trial Research Group, 1993, “The effect of
intensive treatment of diabetes on the development and progression of long-term
complications in insulin-dependent diabetes mellitus”, The New England Journal
of Medicine, 329:977-986

The Writing Group for the Activity Counseling Trial Research Group, 2001, “Effects of
physical activity counseling in primary care: The Activity Counseling Trial: A
Randomized Controlled Trial”, Journal of the American Medical Association,
Vol. 286, (6) pp 687-687

140

Tillotson L, Smith S, 1996, “Locus of control, social support, and adherence to the
diabetes regimen”, The Diabetes Educator, March/April Vol. 22 No 2, pp.
133: 139

Tuomilehto J; Lindstrom J, Eriksson J; Valle T; Harnlainen H, Ilanne-Parikka P,
Keinanan-Kiukanniemi S, Laakso M, Louheranta A, Rastas M, Salminen V,
Uusitupa M, Finnish Diabetes Prevention Study Group, 2001, “Prevention of type

2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose
tolerance”, New England Journal of Medicine, 344(18): 1343-50

United Kingdom Prospective Diabetes Study Group, 1998, United Kingdom Prospective
Diabetes Study, Lancet, 352:837z863

United Kingdom Prospective Diabetes Study Group, 1998, United Kingdom Prospective
Diabetes Study 24, Annals of Internal Medicine, 128(3)]65-175

US. Department of Health and Human Services, Centers for Disease Control and
Prevention, 1998,”National Diabetes Fact Sheet”

US. Department of Health and Human Services, Centers for Disease Control and
Prevention, 2000, “Healthy People 2010”. Atlanta, GA. National Center for
Chronic Disease Prevention and Health Promotion

United States General Accounting Office, 1997, “Medicare: Most beneficiaries with

diabetes do not receive recommended monitoring services”, March 1997,
GAO/HEHS-97-48

Vinik A, Vinik E, 2003, “Prevention of the complications of diabetes”, American Journal
of Preventive Care, Vol 9(3) Supplement pp 863:80

Von Korff M, Gruman J, Schaefer J, Curly S, Wagner E, 1997, “Collaborative
management of chronic illness”. Annals. 127: 1097-102.

Wagner E, Austin B, Von Korff M, 1996, “Organizing care for patients with chronic
illness. [Review] Milbank; 74:511-44.

Wagner E, 2000, “The role of patient care teams in chronic disease management”, [see
comments]. [Review] British Medical Journal; 320:569-72.Notes: COMMENTS:
Comment in: BMJ, 2000 Feb 26; 320(7234): 526-7

Wagner E, 2001, “Improving care for people with chronic conditions”, Thirteenth Annual
National Forum on Quality Improvement in Health Care, Dec 9-12, 2001,
Orlando, Florida

Waitzkin Howard, 1991,The Politics of Medical Encounters, Yale University Press, New
Haven and London

Waldron, Ingrid, 1988, Gender and Health-related behavior, Health Behavior, Emerging
Research Perspectives Edited by David S. Gochman, Plenum Press, New Your
and London

Walker E, 1999, “Characteristics of the adult learner”, The Diabetes Educator, Vol. 25(6)
ppl 8-24

141

Wang C, F enske M, 1996, “Self-Care of adults with non-insulin-dependent diabetes
mellitus: influence of family and friends”, The Diabetes Educator, Sep/Oct. 1996,
Vol. 22(5) pp. 465:470

Wee C, 2001,“Physical activity counseling in primary care, The challenge of effecting
behavioral change”, Journal of the American Medical Association, Vol. 286(6)

Weinberger M, Kirkman S, Sarnsa G, Cowper P, Shortlife A, Simel D, Feussner J, 1994
“The relationship between glycemic control and health-related quality of life in

patients with non-insulin-dependent diabetes mellitus.” Medical Care, Vol.
32(12): 1173-1181

Wiggers J Sanson-Fisher R, 1997, “Practitioner provision of preventive care in general
practice consultations: Association with patient educational and occupational
status.” Social Science Medicine, Vol. 44(2): 137-146

Wiggers J, Sanson-Fisher R, 1997, “Duration of general practice consultations:
association with patient occupational and educational status”, Social Science
Medicine, Vol.44(7) pp. 925-934

Williams D, 1990, “Socioeconomic differentials in health: A review and redirection,
Social Psychology Quarterly, Vol. 53(2) 81 :99

Williams G, Freedman Z, Deci E, 1998, “Supporting autonomy to motivate patients with
diabetes for glucose control” Diabetes Care; Vol.2]: 1644-51

Wolinsky F,1980, “The Sociology of Health, Principles, Professions, and Issues. ”
Boston, Little, Brown and Company.

Wright P, Teacher A, 1982, “The problem of medical knowledge: Examining the social
construction of medicine”, Edinburgh: Edinburgh University Press

Zgibor J, Songer T, Kelsey S, Drash A, Orchard T, 2002,”Influence of Health Care
Providers on the development of diabetes complications”, Diabetes Care,
25:1584-1590

Ziff M, Conrad P, Lachman M, 1995 “The relative effects of perceived personal control

and responsibility on health and health-related behaviors in young and middle-
aged adults”, Health Education Quarterly, Vol. 22(1), 127-142

142

 

   
 
 

ARIE

lililiilijilljjiifll[[111 1111

09