DOCTORAL DISSERTATION SERIES cm nm titleah htiicd M m w m Mimi ms m HsuAces m m u c w m tm AUTHOR umob l m&dM university DEGREE HICHI6AH STATS CUL, PL b. d a te PUBLICATION NO. s \ UNIVERSITY MICROFILMS A N N ARBOR • M I C H I G A N /9 5~/ AN OPERATIONAL METHOD FOR MEASURING MEDICAL HEEDS AND RESOURCES IN RURAL COIIMUNITIES by Linwood L. Hodgdon A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF FHILOSOPKY Department of Sociology and Anthropology 1951 D E D I C A T I O N This thesis is respectfully dedicated to my father, 0. V/. Hodgdon, II. D., of l/arren, New Hampshire, who in 1951 has completed 50 years of service as a general medical practitioner in rural communities in Maine, Now Hampshire, and Vermont. His professional competence, devotion to duty, and high ideals of service have been a constant source of inspiration and encouragement. A C K N O W L E D GEM ENT This author is indebted to many persons for their coop­ eration in making this research project possible. Specific acknowledgement and thanks are hereby made to Dr. Charles F. Loomis, head of the Department of Sociology and Anthropology, for his foresight in establishing the Social Research Service in the department, thereby setting up the machinery through which this and many other projects have been planned and financed; to Dr. Charles R. Koffer for his patience and skill in supervising field work and other aspects of the lilchigan Health Survey, and for constructive criticism of this manu­ script; to Miss Fay Blakely, Clarence Jane, Jean Boek, Walter Boek, and Sheldon Lowry for assisting in the collection of data; to Mr. Hugh Brenneman, Public Relations Counsel for the Michigan State Medical Society, who played an invaluable coordinating role between the Social Research Service and the Michigan State Medical Society; to the Survey Advisory Committee of the Michigan State Medical Society (Drs. H. 3. Zemmer, J. E. DeTar, J. R. Rodgers, R. J. Hubbell, and W. K. Huron) for splendid cooperation with the Social Research Service in planning, financing, and carrying out the state-wide Michigan Health Survey; to Mr. Frank Martin, head of the Tabulating Department at Michigan State College, for valuable advice and assistance in coding and tabulating procedures; and to Drs. J. F. Thaden and D. L. Gibson, members of the joint committee for the Michigan Health Survey, for participation in planning the the survey and in carrying out specialized studies. Dr. Gibson also critically examined the manuscript and offered many con­ structive suggestions. Special acknowledgement and appreciation is expressed to Dr. Edgar A. Schuler, former Associate Professor of Sociology and Anthropology at Michigan State College, and Associate Director of the Social Research Service, for permission to employ the operational method (developed jointly with the author) in this dissertation, and for his untiring assistance and absorbing interest in this project from its beginning. His patience, ability, and unusual understanding of social problems have been a constant source of inspiration and en­ couragement. My wife, Candis Uorris Hodgdon, assisted in the collection of data and has given unselfishly of her assistance and encouragement throughout the entire study. Appreciation is also extended to Dr. H. C. Fryer, Profes­ sor of Mathematics at Kansas State College, lor assistance with statistical problems in Chapter VII, and to Professor George Montgomery, head of the Department of Economics and Sociology at Kansas State College, who provided trained typists to pre­ pare the final draft of this manuscript. 1 TABLE CF CONTENTS FART I - HEALTH AND MEDICAL CARE IN THE UNITED STATES CHAPTER I. NATURE OF THE PROBLEM Health and Medical Care as a Social Problem Seeking an Approach to the Study of Social Problems Social Problems and the Role of Social Research CHAPTER II. THEORY OF THE OPERATIONAL METHOD The Development of Operationalism An Operational Definition of "Law" (Illustra­ tion #1 ) An Operational Definition of "Social class status" (Illustration #2) CHAPTER III. STATEMENT OF THE FROBLEM AND OBJECTIVES Statement of the Problem Thesis Objectives Page 2 p > 7 12 ILJ / ■ > _ L lA ly 1C It 21 PART XI - COMPONENTS OF THE OPERATIONAL METHOD CHAPTER IV. CHAPTER V. CHAPTER VI. THE NATURAL AREA CONCEPT Definition of the Concept Delineation cf Medical Service Areas Testing the Natural Area Concept as it Relates to Medical Services Testing the Natural Area Concept as it Relates to Hospital Services Conclusion AO MEASURING PHYSICIAN RESOURCES Measuring the Service Capacity of General Practitioners Measuring the Patient Load of General Practi­ tioners AA MEASURING THE NEED FOR MEDICAL CARE The Medical Needs Schedule Validation of the Medical Needs Schedule Methodological Problems Reliability of Data Furnished by the Housewife for Other Adults in the Family -3 2C at at >2 TABLE OF CONTENTS (Cont'd) Page CHAPTER VII. VARIABLES AFFECTING NEED FOR MEDICAL CARE Introduction Conclusions Regarding Correlation Analyses A. Age as a Factor in Medical Need B. Frequency Distributions Within Age and Income Levels C. Income as a Factor in Medical Need D. Other Variables Affecting Medical Need Conclusions Regarding Variables Affecting Need for Medical Care PART III.- MEASURING MEDICAL NEEDS AND RESOURCES IN RURAL COMMUNITIES CHAPTER VIII.MEASURING MEDICAL NEEDS AND RESOURCES IN PLEASANTVILLE AND ELMWOOD Application of the Operational Method to Communi­ ty Situations A. Steps in Determining the Need for General Medical Practitioners in the Community B. Steps in Determining the Supply of General Medical Practitioners in the Community Measurement of Medical Needs and Resources in Pleasantville A. Measurement of the Need for General Medical Practitioners in the Community of Pleasantville B. Measurement of the Supply of General Medical Practitioners in Pleasantville Comparison of the Operational Method with the Traditional Method of Measuring Medical Needs and Resources Measurement of Medical Needs and Resources in Elmwood A. Measurement of the Need for General Medical Practitioners in the Community of Elmwood B. Measurement of the Supply of General Medical Practitioners in Elmwood Comparison of the Operational Method with the Traditional Method of Measuring Medical Needs and Resources CHAPTER IX. VALIDITY OF THE OPERATIONAL METHOD Margins of Error Involved in the Operational Method A. Validity of the Natural Area Concept B. Margins of Error Involved in the Measurement of Medical Needs 1. The Medical Needs Schedule TABLE OF CONTENTS (Cont'd) Pa ?6 2. Reliability of Data Furnished by the Housewife for Other Adults 3. Seasonal Variation in Need for Medi­ cal Care *+. Estimation of the Number of Fatient Calls Which Would be Required to Treat all Individuals Having Fositive Symp­ toms C. Measurement of General Practitioner Resources Other Factors in the Estimation of Medical Needs and Resources Appraisal of the Over-all Validity of the Oper­ ational Method CHAPTER X. SUMMARY AND CONCLUSIONS The Need for a Conceptual Framework for the Study of Social Problems Aspects of the "Value” Approach A. Objective and Subjective Aspects B. The Natural History of a Social Problem The Theory of the OperationalMethod An Operational Method for Measuring Medical Needs and Resources in Rural Communities A. Thesis Objectives B. Components of the Operational Method C. Validity of the Operational Method D. Comparison of the Operational Method of Measuring Medical Needs and Resources with the Traditional Method Conclusion BIBLIOGRAPHY 93 9*+ 9 *+ 95 96 97 99 99 101 101 10*+ 10*+ 10? 105 107 110 110 112 113 APPENDIX A - SCOPE AND OBJECTIVES OF THE MICHIGAN HEALTHSURVEY ll6 APPENDIX 3 123 AFFENDIX C - METHODS AID PROCEDURES EMPLOYED IN TIE MICHIGAN HEALTH SURVEY - THE SAuPLE FOR THE MICHIGAN HEALTH SURVEY APPENDIX D - THE MIC HI GA.. IfEALTH SURVEY SCIEDULE 1?3 136 LIST OF FIGURES FIGURE 1 PAGE . Communities - their centers and service areas; medi­ cal service areas in Clinton County, Michigan 2. General practitioner medical service areas in Michigan, 19^8 O* 27 3. Ratio of doctors to population in 30^,medical ser­ vice communities in Michigan, 19^8 30 b. Stephenson-Daggett medical service community in Menominee County, Michigan, 19*+7 3V 5. Per cent services adjacent place of 36 6. of 6 3 1 families seeking general medical within their medical service area, in an area, or in a non-adjacent area from their residence, Michigan, 19*+8 Primary and secondary hospital service areas in Michigan, 19*+9 7. Symptoms page of the Medical Needs Schedule employed in the Michigan Health Survey, Michigan, 19^8 8. Percentage distribution of 3 5 778 individuals in the Michigan Health Survey sample population reporting designated number of positive symptoms, by age groups, Michigan, 19^8 9. Correlations between age and number of positive symp­ toms within eight income levels, Michigan Health Sur­ vey, 19^8 39 53 66 CS 10. Percentage distribution of 3,786 individuals in the Michigan Health Survey sample population reporting designated number of positive symptoms, by income groups, 19*+8 72 11. Correlations between income and number cf positive symptoms within nine age levels, Michigan Health Sur­ vey, 19*+8 7'£ 12. County highway map of Muskegon County, Michigan, show­ ing the four types of sample areas used in the Michigan Health Survey, 19^8 (APPENDIX B) 13. Percentage distribution of the households and individuals in the sample for the Michigan Health Survey, by 13C LIST OF FIGURES (Cont'd) FIGURE residence, 19^8 (APPENDIX C) l^f. Distribution of households in the sample for the Michigan Health Survey, by counties, 19^8 (APPENDIX C) 15. Village map of Concord, Michigan, showing method of assigning sampling units to blocks, and selection of block to be used in the samnle for the M'chigan Health Survey, 19^8 (APPENDIX C) LIST OF TABLES TABLE PAGE I. Number of families seeking medical care within their general medical practitioner service area, in an adja­ cent area, or in a non-adjacent area in Michigan, by counties, 19*+8 II. Number of families seeking hospital services within their hospital service area, in an adjacent area, in a non-adjacent area, or in two different areas in Michigan, by counties, 19^8 III. Service capacity of all physicians expressed as a pro­ duct of decimals showing their service capacity at different age levels, United States, 19^0 IV. Per cent of 153 individuals needing and not needing medical care according to Medical Needs Schedule and Medical examination in three Michigan c ounties, 19*+6 V. A comparison of the reporting of need for medical care by 182 housewives on other adults in the household, with reports by these adults on themselves, Michigan Health Survey, 19^8 VI. Correlations between age and number of positive symp­ toms within eight income levels, Michigan Health Sur­ vey, 1 9 ^ 8 VII. Frequency distribution of 3 j*+98 individuals in the sam­ ple population for the Michigan Health Survey, by income and age levels, Michigan, 19^8 VIII. Correlations between income and number of positive symp­ toms within nine age levels, Michigan Health Survey, 19^8 IX. Estimated average change in number of positive symptoms for each additional year of age, by income level. Michi­ gan Health Survey, 19^8 (regression coefficients) X. Chi square analysis of association between place of resi­ dence "and number of positive symptoms within designated age and income categories, Michigan Health Survey, 19^+8 XI. Number and per cent of 3 5786 individuals in the sample population for the Michigan Health Survey reporting designated number of positive symptoms, by sex, 19^+8 LIST OF TABLES (Cont'd) TABLE XII. Distribution of households and individuals in the sample for the Michigan Health Survey, by residence, 19^0 (APPENDIX A) PAGE 122 PART I HEALTH AMD MEDICAL CARE IN TEE UNITED STATES CHAPTER I NATURE OF THE PROBLEM Health and Ifefl3,c.aA Care £S £ Social Problem The maintenance of good health is a matter of direct concern to all segments of the population. Because of this fact, it is not surprising that few social problems1 in re­ cent years have aroused greater public interest and concern than the problem of obtaining adequate medical care. Three factors, in particular, have played an important role in focusing the attention of the public upon this prob­ lem. First, the widespread publicity which has been given to the draft statistics for World War II inductees has called the attention of the public to the fact that the level of health care, as indicated by the relatively high proportion of in­ dividuals who were rejected because of physical or mental dis­ abilities, is not as high as it should be. Secondly, attempts on the part of various groups and agencies in recent years to Fuller and Myers define a social problem as "an actual or imagined deviation from some social norm cherished by a considerable number of persons." According to these authors, social problems do not arise full-blown but go through a natural history involving the phases of 1 ) awareness, 2 ) policy determination, and 3) reform. It is apparent that the problem with which we are concerned is now in the very active stage of policy determination, and that widespread reforms relating to health and medical care are imminent. For a more detailed dis­ cussion of the theory of social problems see the following articles by Richard C. Fuller and Richard R. Myers: "The Natural History of a Social Problem," ^m. Soc. Rev., Vol. VI, (June, 19*+l), pp. 320-329; "Some Aspects of a Theory of Social Problems," Am. Soc. Rev.. Vol. VI (Feb., 191!-!)} pp. 2h-32. 3 enact some type of health legislation on the national level have resulted in widespread publicity being given to the whole problem of medical care— in all of its manifold aspects— by both the opponents and proponents of the various Federal health proposals. Thirdly, the relative inadequacy of medical facilities and services in rural areas has resulted in growing concern on the part of physicians and laymen alike, particu­ larly the rural families living in these areas. This situation has been attributed to the increasing tendency on the part of younger physicians to specialize and to locate in urban areas, thus leaving in the rural sections of the country an insuffi­ cient number of doctors to meet the medical needs of the people. Furthermore, many general practitioners have found it to their advantage to locate in the larger centers of population, thus making the shortage of physicians in rural areas even more critical. The advances in medical science during the past few decades have been tremendous, yet these potential advantages have not accrued to all the people because of the lack of progress in social organization and related social activities. Economic barriers, lack of health education, inadequate dis­ tribution of medical facilities and services, and many other factors, have all played their part in preventing large seg­ ments of the population from realizing the benefits of the many achievements in the field of medical science. There is a growing awareness that steps must be taken to protect health, combat disease, and provide adequate medical care for all who need it. Regardless of the means by which these objectives are attained, there is little reasonable doubt but that these measures have social as well as medical significance. Extensive social research, carried out by competent research personnel in cooperation with existing medical organizations and agencies, must be undertaken to supplement the efforts of medical personnel before the many problems relating to health and medical care can be properly evaluated or ameliorated. Seeking an Approach to the Study of Social Problems The meaningfulness of social research is measured to a considerable extent by the degree to which the data obtained from research projects are related to some theoretical or conceptual system. In the absence of such a system, collected data remain a mass of unrelated and relatively meaningless facts. For this reason, it becomes apparent that any science, particularly one which is relatively young and in the process of development, should concentrate upon the early development of meaningful and valid theoretical frameworks. In this connection, the author suggests that the so-called ’’value” approach to the study of social problems, initially developed by Richard C. Fuller 2 of the University of Michigan, Richard C. Fuller, "The Problem of Teaching Social Problems,” American Journal ,s£ Sociology. Vol. W (1938), pp. bl5-b25. 2 5 and further elaborated by Fuller and Myers^, Cuber and 1+ Harper , and others, can be profitably employed as a basic, theoretical framework for the study of social problems. Ac­ cording to the value approach, social problems are defined as such only when a condition arises which is believed by the people in a given culture to represent a deviation from some social value or objective which is cherished by a considerable number of persons. Moreover, social problems arise and are sustained because people do not share common values and ob­ jectives. A dual conflict of values actually exists with re­ gard to social problems; first, disagreement among the people as to whether or not a condition constitutes a threat to the welfare of the group, and second (assuming agreement that a condition is a threat), disagreement regarding what should be done about it. Social problems, when viewed from the "value" approach, are also seen to have both their objective and subjective as­ pects; the objective aspect requiring factual demonstration that a given condition actually exists in the society, and the subjective aspect involving the manner in which social values perform the dual role of creating the problem as well as preventing its solution. •3 Richard C. Fuller and Richard R. Myers, "Some Aspects of a Theory of Social Problems," American Sociological Review. Vol. VI (Feb. 194-1), PP- 24-32. L. John F. Cuber and Robert A. Harper, Problems of American Societvs Values in Conflict. New York, Henry Holt & Co., 194-8. Using this approach to study the problem of health and medical care, it becomes apparent that the inability of a sizeable portion of the population to obtain adequate medical care— because of low purchasing power, unavailability of physicians, or other reasons— constitutes a definite threat to a considerable number of persons and families in the society. Thus, in terms of the values held by these people regarding the desirability of 11good” health and "adequate” medical care, this situation is defined by the people them­ selves as a social problem. This is further evidenced by the increasing concern which is being shown over this situation by many groups in the population, for reasons already stated. Regarding the dual conflict of values involved in social problems, an investigation of the problem with which we are concerned clearly shows that cultural values which are widely accepted in our society concerning this problem, e.g., the "right" of physicians to practice in whatever location they choose, the pecuniary emphasis so characteristic of our cul­ ture which motivates physicians (and those in other businesses and professions as well) to establish themselves in areas where ability to purchase medical services is relatively high, etc., can unquestionably be considered as causative factors involved in the problem with which we are concerned. Even more apparent, however, is the manner in which con­ flict of values actually retards the solution of this situation which many people have defined as undesirable. First of all, there is not universal agreement among the people that the present situation regarding health and medical care in this 7 country is actually a problem, and secondly, among those groups which do acknowledge the presence of a problem, there is widespread difference of opinion— even violent and bitter antagonism— concerning remedial measures which should be adopted. One needs only to read the senate hearings on bill S-1606, introduced into Congress in 19*+6 and described by its sponsors, Senators Wagner (N.Y.)-Murray (Mont.)-Dingell (Mich.), as "a bill to provide for a National Health Program11 to verify the fact that determined and widespread conflict of values exists among the many groups whose representatives testified at these senate hearings. Social Problems and the Role of Social Research What, then, one might ask, are the functions of social research as related to social problems? In terms of the con­ ceptual scheme outlined above, it becomes apparent that a major role of social research relates to the so-called "objective1' aspect of social problems; i.e., one important function of research is to demonstrate whether or not a given condition actually exists, to what degree, and to study the functional inter-relationships of the many variables which may be found to be causally significant. The contributions which social research can make in the broad area relating to the social aspects of medical care are limitless, but appear to be of two general types— theoretical and practical. The ideal type of research project, however, would be one in which the two types are combined, and in which an attempt is made not only to contribute to the theoretical framework of sociology in this area but also to add to the scientific knowledge already available and thus facilitate the amelioration of the many problems relating to health and medical care. Much valuable research by sociologists has already been done in this area, and there are indications that it will be greatly expanded in the future. Extensive investigations, involving many different phases of medical care, have been carried out by Edgar A. Schuler and Charles R. Hoffer of Michigan State College, A. R. Mangus of Ohio State University C . Horace Hamilton and Selz C . Mayo of North Carolina State College, and Charles E. Lively of the University of Missouri, 0. D. Duncan of Oklahoma A & M College, William Mather of Pennsylvania State College, Paul Landis of Washington State College, Carl Kraenzel of Montana University, to mention only a few of the outstanding research men and programs 5 which have been undertaken. For a more detailed description of some of these projects see "Symposium on Objectives and Methods of Socio­ logical Research in Health, 11 Rural Sociology,XIV (Sept., 1 9 *+9 )j pp. 1 9 9 -2 1 9 , consisting of the following articles: (1) C. E. Lively, "Objectives and Methods of Rural Sociological Research in Health at the University of Missouri pp. 1 9 9 -2 0 6 . (2) Edgar A. Schuler, Charles R. Hoffer, Charles P. Loomis, and Paul A. Miller, "Objectives and Methods of Rural Sociological Research in Health at Michigan State College," pp. 2 0 6 -2 1 2 . (3) A. R. Mangus, "Objectives and Methods of Rural Socio logical Research in Mental Health at Ohio State University," pp. 212-219. 9 In nearly all eases attempts have been made to coordinate the research work with the programs of state and local medical societies, and with other interested groups and agencies. For example, the extensive program in mental health at Ohio State University under the chairmanship of Dr. A. R. Mangus is car­ ried out in cooperation with the Division of Mental Hygiene in the State Department of Public Welfare and the Ohio Agri­ cultural Experiment Station, and the state—wide Michigan Health Survey^ which was recently completed was a joint pro­ ject of the Michigan State Medical Society ana the Social Research Service (in the Department of Sociology and Anthro­ pology) at Michigan State College. Another significant development in recent years has been the employment of full- or part-time extension specialists in rural health at many of the state colleges and universities. Extension rural health services are carried on in at least fifteen states, and others have been setting up extension health committees to explore Extension*s potential contribution to rural health, or have participated actively in state-wide health planning committees concerned with improving rural 7 health services and facilities. These fifteen states include ^ C. R. Holler, and others, ’’Health Needs and Health Care in Michigan," East Lansing, Michigan, Agricultural Experiment Station Special Bulletin 3^5 (June, 1950)* 7 For additional information on this interesting develop­ ment see: Elin L. Anderson, ’’The Extension Service’s Responsi­ bility in Aiding Rural People to Improve Their Health and Medical Services," United States Department of Agriculture Extension Service, Washington, D.C., (July, 19^7). Arkansas, Florida, Illinois, Indiana, Kansas, Michigan, Mississippi, Montana, Missouri, Nebraska, New York, North Dakota, Ohio, South Carolina, and Wyoming, With regard to the culturally defined problem of health and medical care, an important function of social research consists in developing useful tools and concepts whereby medical needs and resources can be objectively measured, in actually measuring these needs and resources, and investi­ gating the relationships which exist between the need for medical care and such important variables as age, income, etc. In other words, a thorough understanding of any social problem, which can best be gained by objective and unbiased investi­ gations, is a necessary condition to the prediction or con­ trol phases of the problem. The type of action which follows this fact finding process is, in the last analysis, again dependent upon the values held by the people with regard to it. In summary, then, the necessity of relating social research to some theoretical or conceptual system, and of broadening and improving the framework within which social problems are analyzed, should be recognized by all social scientists. The "value" approach to the study of social problems is, in the opinion of the author, a most useful and valid theoretical system in the light of which all social problems can be analyzed objectively, and will be employed in this research to critically study the problem of health and 11 medical care as it exists in the United States at the present time. The specific nature of the problem to be considered, the specific objectives toward which this investigation is oriented, and the method by which these objectives will be sought by means of operational methods and techniques, will be described in the two chapters immediately following. 12 CHAPTER II THEORY OF THE OPERATIONAL METHOD The Development of Operationalism One of the greatest contributions to scientific think­ ing in recent years, in the opinion of many qualified observ­ ers, has been the development of the so-called operational method, or operationalism. Harry Alpert made one of the clearest pronunciations of operationalism when he stated; That symbols and concepts are amenable to manipulations or operations which cannot be per­ formed on the things symbolized or conceptualized was clearly revealed in physics with the estab­ lishment of the Einsteinian theories of relativity. One of the significant contributions of Einstein was, in fact, a *final disentanglement of that part of any physical event which is contributed by the observer from that which is inherent in the nature of things and independent from all observers. 11 2 Alpert stated that Bridgman , while working out the methodological implications of the revolution in the physical sciences wrought by Einstein, tried to develop a method of conceptualizing which would to a large extent eliminate the need for, and hence the recurrence of, similar drastic changes in the conceptual framework of physics. His method 1 Harry Alpert, "Operational Definitions in Sociology," Am. Soc. Rev.. Vol. Ill (Dec., 1933), pp. 855-861. 2 P. W. Bridgman, The Logic of Modern Physics, Macmillan, 1 9 3 2 ), p. 37. 13 was nothing more than an insistence that concepts be made operational, i.e., that they be defined not in terms of metaphysically conceived "properties" but rather in terms of observable physical operations. Thus was developed the operational point of view in physics. The introduction into sociology of the operational point of view originally developed in physics has perhaps been most vigorously urged by Dodd and Lundberg , the latter of whom supports his position by stating: That sociologists exhibit only slight agree­ ment even in the use of the most common terms is a matter of common knowledge. The same sociologist frequently uses the same term in various senses in the same article. This state of affairs is not surprising, because the only way of defining any­ thing objectively is in terms of the operations involved. Concerning this important question of definitions in sociology Dodd^ states that the criteria of definitions which formed the basis for his work Dimensions of Society were as follows: first, his working hypothesis was that a scientific definition of a sociological concept should be operational, i.e., it should state the procedure and the See S. C. Dodd, Dimensions of Society. Macmillan (19^-1), and G. A. Lundberg, Foundations of Sociology., Macmillan (1939) for basic postulates and methods. L. G. A. Lundberg, "The Thoughtways of Contemporary Sociology," Am. Soc. Rev.. (Oct., 1936), p. 709. (Words underlined by the author.) Stuart C. Dodd, "A System of Operationally Defined Concepts for Sociology," Am. Soc. Rev.. Vol. IV (Oct., 1939), p. 619. Ilf materials to be used to obtain the entity defined, just as a kitchen recipe defines a cake or a Binet manual defines a mental age; secondly, a system of definitions covering a field of knowledge should be parsimonious in number; thirdly, it should conform to the canons of classification, namely, a single basis for classes which are mutually exclusive and totally inclusive of the field classified; and lastly, that a definition should have its reliability measured by some statistical index experimentally derived from repeated in­ dependent applications of the definition to the class of entities that it defines. In addition to operationalism, a second methodological tenet which Alpert believes should be stressed is that of probability. The task that sociology should undertake, he believes, is the setting up of definitions that combine the operational and probability points of view. In other words, social phenomena should be defined both in terms of the probability of occurrences and in terms of socially meaning­ ful, verifiable operations. Alpert presents the following probability-operational definitions of (1 ) ‘'law” and (2 ) •‘social class status" as illustrative examples. An Operational Definition of "Law" (Illustration No. 1) The following probability-operational definition of "law," which draws its inspiration from the writings of Holmes, Cardozo, Frank, Cohen, and others, is both clear and verifiable, as well as being socially relevant: 11Law is the probability that a rule of conduct will be enforced by, the courts According to Alpert, Holmes* prediction or probability theory of the law, which is here summarized, has become a classic. He wrote: If you want to know the law and nothing else, you must look at it as a bad man, who cares only for the material consequences which such knowledge enables him to predict... what constitutes the law? You will find some text writers telling you that it is something different from what is decided by the courts of Massachusetts or England, that it is a system of reason, that it is a deduction from principles of ethics or admitted axioms or what not. But if we take the view of our friend the bad man we shall find that he does not care two straws for the axioms or deductions, but that he does want to know what the Massachusetts or English Courts are likely to do in fact. I am much of this mind. The prophecies of what the courts will do in fact, and nothing more pretentious, are what I ---— Eauns’is:7 An Operational Definition of "Social Class Status" tration No. 2.) (Illus­ "Social class status is the probability that an indi­ vidual will reap the highest rewards available in a given • * ..8 society." In terms of the above definition, the operational test £ Alpert, op. cit.. p. 8597 Holmes, quoted in Alpert*s "Operational Definitions in Sociology," p. 859. ^ Alpert, .op. .cit., p. 36l. 16 of social status is this: What are the life chances of an individual, or the probabilities that the actual operations will take place whereby he will enjoy what the culture esteems highest? In a society in which the highest values are pe­ cuniary, for example, the test becomes that of the chances for a given individual to become extremely wealthy. From this point of view, a social class would be a group of persons having equal probabilities of reaping social rewards. Alpert is aware of the fact that the calculation of status probabilities will obviously vary as we move from the one extreme of a caste system in which vertical mobility is practically zero (and where one's chances are determined by birth) to the other extreme of a highly competitive and rela­ tively wide open class structure in which the vertical mobil­ ity is relatively great. It will also vary as the objects of social valuations vary, but in all cases the phenomenon of stratification will be found to be socially relevant only as it bears on the comparative life chances of the individuals in a given society. As might be expected, operationalism has not been met with open arms by all physical or social scientists. Contro­ versies still rage over what it is or is not, what it pur­ ports to do, e t c C o n c e r n i n g this situation, Boring at least attempted to clear the atmosphere concerning the operational ^ See George A. Lundberg's "Operational Definitions in the Social Sciences," The Am. jJ. of Soc.« Vol. XLVII (March, 1 9 ^ 2 ), pp. 727-7^5. 17 method in psychology when he stated: Operationalism is not a technique for the formation of concepts or theories, nor a system of philosophy, but a set of regulative or critical standards in the light of which the meaningfulness and fruitfulness of scientific concepts can be ap­ praised. The outstanding requirements which operationalism has justifiably stressed are as follows: concepts which are to be of value to the factual sciences must be definable by operations which are logically consistent, sufficiently definite (quan­ titatively precise), empirically rooted, technical­ ly possible, repeatable, and aimed at the creation will function in laws of greater In summary, then, we might say that if the fundamental tenet of the operational view in physics is that physical concepts must be defined in terms of actual physical opera­ tions, then the operational postulate in sociology should be that social concepts must be defined in terms of social operations, or operations which are socially meaningful. 1 0 Edwin G. Boring, et al.. "Symposium on Operationism," Psychological Review. Vol. LII (194-5)? P* 24-8. 18 CHAPTER III STATEMENT OF THE PROBLEM AND OBJECTIVES Statement of the Problem In recent years there has been growing concern on the part of both physicians and laymen over the relative inad­ equacy of medical facilities and services in rural sections some time Although this problem has been recognized for/ of the country. there is a lack of reliable information concerning the pro­ portion of the rural population which actually has insuf­ ficient medical care, or the extent to which the need for medical service varies among different communities. There is no very satisfactory method which will show the extent of need for medical care among the people and the number of physicians required to care for their needs. This thesis describes a method which is designed to provide such infor­ mation and outlines the steps or operations by which they may be obtained. 1 1 Edgar A. Schuler and Linwood L. Ilodgdon, 11An Operational Method for Appraising the Supply of, and Need for, General Medical Practitioners in Rural Communities .*• Unpublished manuscript, processed, East Lansing: M. S. C., Dept, of Soc. and Anth., (Feb. 21, 19*+9), 15 PP. The method here referred to was developed jointly by Schuler and Ilodgdon and will be employed in this dissertation. In connection with the development of the operational method, the writers wish to acknowledge their gratitude for encouragement and indebtedness for critical assistance re­ ceived from professional colleagues, both social scientists and medical doctors. Among these are: Elin L. Anderson, 19 It has frequently been assumed by physicians and others working with the problems of health and medical care that one doctor for every one thousand people is an adequate physicianpopulation ratio in any given area. This crude measure has traditionally been accepted by health authorities as an in­ dication that adequate medical services are present in a given area to meet the needs of the people. The central hypothesis upon which the operational method employed in this dissertation is based, however, is that the present crude method of measuring the adequacy of medical services, as stated above, does not represent a true picture of medical service adequacy, for the following reasons: 1. The area within which a given population resides is not defined by those who employ the crude method. 2. A simple counting of doctors may be highly misleading because it is based on the assumption that all doctors have equal capacities for render­ ing medical services. 3. There is increasing evidence which indi­ cates that need for medical care is not a constant, as those who employ the crude method assume, but rather a variable which is affected by such factors as age, income, etc. Regarding the first of these factors, one would assume that those who employ the crude method use township or other artificial political units as the basis for their population Odin W. Anderson, 0. D. Duncan, Wm. U. Form, Maurice H. Fried­ man, M.D., Duane L. Gibson, C. Horace Hamilton, Louis P. Heil­ man, Chas. R. Hoffer, Roelof Lanting, M.D., C. E. Lively, Henry B. Makover, M.D., A. Ray Mangus, Wm. G. Mather, Selz C. Mayo, Robert L. McNamara, Fred D. Mott, M.D., Milton I. Roemer, M.D., J. F. Thaden, Richard E. Weinerman, M. D. 20 estimates since data for these areas are most readily avail­ able from census materials. The use of such data for this purpose ignores one of the basic tenets of operationalism, namely that they are not socially relevant. People do not necessarily seek medical services within the limits of the township or other political unit in which they reside, but rather within medical service areas (natural areas), the boundaries of which overlap, and are independent of, arti­ ficial political demarcations. The boundaries of the latter are not fixed, but are determined by such factors as topography, transportation facilities, nearness to other service centers, number and type of services available in these service centers, etc. Concerning the appraisal of physician resources, there is considerable evidence available which indicates that the service capacity of a physician varies significantly with several variables, one of which is the age of the physician. It becomes apparent, therefore, that any reliable estimation of physician resources must consider these available resources not as a constant but rather as a variable because the service capacity of all physicians is not the same. The third factor, that of medical needs, must also be considered as a variable rather than as a constant in any de­ termination of medical needs and resources in a given area. In view of these three factors, which those who employ the crude ratio of one doctor per thousand population ignore, it becomes apparent that any operational definition of d 21 the number of physicians needed to meet the needs of the people in a given community must include the following basic components: first, a definition of the area within which medical facilities and services are most likely to be ob­ tained; secondly, a method of measuring physician resources which takes into consideration the relevant variables in­ volved; and thirdly, a method of measuring medical needs which recognizes and makes allowances for the socially rele­ vant variables involved in this component. The three com­ ponents of the operational method are discussed in detail in Part II of this dissertation (Chapters IV-VII, inclusive), and Chapter VIII is devoted to an explanation of how they can be combined in actual community situations to determine the relative adequacy or inadequacy of general practitioner medi­ cal personnel in relation to the measured medical needs of the community. Thesis Ob.iectives The main objectives of this thesis are four-fold: first, to outline the need for, and explain the steps in the develop­ ment of, an operational method for measuring medical needs and resources in rural communities; secondly, to define each of the three main components of the method operationally and demonstrate how each may be measured quantitatively; thirdly, to illustrate the procedures involved in the over-all method by applying the operational method irs actual community situa­ tions; and fourthly, to refine further the techniques and 22 measuring instruments employed in the operational method by rigorously testing the basic hypotheses and concepts upon which the method is based. This testing and refinement is made possible by the great amount of factual data relating to health and medical care problems which was obtained in the state-wide Michigan Health Survey. 2 2 The Michigan Health Survey was a state-wide project in which information was gathered concerning the medical needs, experiences, etc. of 3»786 individuals. Because of the importance of these data in testing procedures, hy­ potheses, etc. the Michigan Health Survey is discussed in detail in Appendixes A, 3, and C in terms of (1) scope and objectives, (2) the sample, and (3) methods and procedures. PART II COMPONENTS OF TIE OPERATIONAL METHOD 23 CHAPTER IV THE NATURAL AREA CONCEPT Definition of the Concept One of the basic hypotheses in the operational method is that if the natural area concept is valid as a conceptual tool for the study of services in general it is valid for medical services. Since one of the essential steps in the experimental design of this dissertation involves the test­ ing not only of the measuring instrument used but also the basic hypotheses inherent in the operational method, it was necessary to design a crucial test of the concept of natural area as applied to medical and hosjjital services. The pur­ pose of this chapter is to describe what this test was, and the results obtained by its use. Proper sampling implies, as does the use of statistical techniques in general, the use of some kind of universe in which these devices have validity. The use of the measuring devices employed to measure both the needs for medical service and the physician resources available to meet these needs likewise requires that the area or universe in which they are employed be defined. The term "natural area" as used here means a trade center and the trade area contiguous to it. In other words, it is a town-country community— a population center rendering a sufficient variety of services to satisfy most of the needs of its people both within the trade center 2b and in the area dependent upon it for services. (Figure 1) A community or trade area tends to take the most natural shape in terms of available means of transportation. The core of the community is the town or village in which the density of population is greatest, and in which most of the goods and services, including the medical services, required by the inhabitants are to be found. The tributary area sur­ rounding the trade center usually lias as large a population as that of the trade center, and in most instances has an even greater population. If there is a resident physician, he is usually a member of the community: his children go to school there, his family attends a local church, and he patronizes to a large extent the local stores, business establishments, professional people, etc. for most of the goods and services needed or desired by the members of his family. Similarly, the trade center tends to be the most efficient location for his office, since it is in the trade center that the members of the community seek most of their medical services. The boundaries of a trade area are determined by to­ pography, transportation facilities, nearness to other trade centers, etc., rather than by artificial township, county, or other political boundaries. Other things being equal, the larger the trade center and the more extensive the serv­ ices offered in the center the greater will be the 11drawing power” of the trade center. This principle applies equally well to medical facilities and services, and in communities 2? iwa’pr* T rtapids hi s i e S t.J o h n s Oy iJ Fowlej* W es tp h a l i a D e iV it Ba th E2 C h IN TON COUNTY “" ^ ( tRa a T (ToJiTl'Y 241 Lans Ing Figure 1. Communities--Their Centers and Service Areas; Medical Service Areas in Clinton County, Michigan. Source: J. F . Thaden, "The Lansing Region and 11,s Tribute] Town-Country Corununitles," Michigan State College, Special bullet In 302, Agricultural Mxperiment Station, Sectiion o: '~ iology (liarch, 195hS ) . ..Ranted fros Figure 11, n . 26 where there is no physician the adjoining medical service areas tend to overlap'1’ to supply the deficient community. Delineation of Medical Service Areas One of the specific phases of health and medical care which was selected for intensive investigation in the Michigan Health Survey was the delineation of local medical service (general practitioner) areas. This ecological study was conducted by Dr. J. F. Thaden, and resulted in the de­ lineation of 30*+ general practitioner medical service areas throughout the state of Michigan, 2 (Figure 2). The manner It has been suggested that the overlapping of medical service areas would invalidate to some extent the data relat­ ing to physician resources within a given community or trade area. Concerning this methodological problem, it is recog­ nized that the boundaries of trade areas are not permanent or precise and that some overlapping, or interflow between areas, does occur. Assuming, however, that two adjacent trade areas are similar in size and in the quantity and quality of medical facilities and services that are available, the flow back and forth will usually be compensatory. On the whole, these trade areas can be delineated with a considerable degree of accuracy, and to the extent that the natural area concept is valid at all, it is equally valid for medical services. An examination of Figure 1 will show that the trade centers with the greater populations, and with more extensive medical services and facilities, exert a correspondingly greater draw­ ing po\*er than do the smaller ones. It should again be em­ phasized that we are concerned primarily with general practi­ tioners and with rural communities. For the services of specialists the trade area concept becomes less valid, and the more specialized the type of medical service which is re­ quired the greater will be the area served by it. 2 J. F. Thaden, "Delineation of medical communities in Michigan and determination of their estimated population." Unpublished manuscript, processed, East Lansing; M.S.C., Social Research Service, July l1*-, 19^3, 6 pp. 27 GENERAL PRACTITIONER CO M M U N ITIES Figure 2. General practitioner medical service areas in Michigan, 19*+3. Source: Unpublished material of Dr. J. F. Thaden, Department of Sociology and Anthropology, Michigan State College. 28 in which these medical service areas were delineated will now be briefly described. The first step in the delineation of the general practi­ tioner medical service areas in the state of Michigan, accord­ ing to Dr. Thaden, was the listing of all population centers in the state having 500 19^0 census figures. or more inhabitants according to the Normally, a village with 500 inhabitants would have a tributary trade and service area of additional people, or a total of 1,000 people. 500 or more Presumably, then, such an area would have sufficient population to support a doctor. There were a few exceptions to this general rule, as might be expected. In Oscoda and Montmorency Counties, for example, there were no communities with as many as 500 in­ habitants, nevertheless the villages in Mio, Atlanta, and Hillman were considered to be centers for medical service in these counties. Also, some villages with more than 500 in­ habitants were not considered as individual medical centers if they were located within ten miles of large population centers. Thus the many small communities within a ten mile radius of the city of Lansing, including Holt, East Lansing, Haslett, Dimondale, and Potterville, were considered as be­ longing to the Lansing service area rather than making up several smaller areas contiguous to the city of Lansing. The reasons for this are rather obvious; for the closer we approach to the large centers of population, the more numerous and varied are the medical and other services v/hich we find there, and the greater is the "drawing power" over the surrounding areas. 29 The second step in the delineation of these medical serv­ ice areas was the delineation of the boundaries between the selected community centers. The mid-points between a commun­ ity and all of its neighboring community centers were first located, plotted, and adjusted for relative drawing power. These adjusted mid-points between community centers are then elongated and connected, thus locating the approximate bound­ ary of the composite trade and service area surrounding a community center. The third step was to compute the present population of the community center, and the population within its surround­ ing service area. Population estimates for the 125 urban centers of the state, and for the remaining rural population of each county, were obtained from the State Department of Health for July 1, 19^6. These figures, together with school census figures for some 6,000 school districts for the years 19^-0, 19^6, and 19*+7, constituted the basis for arriving at population estimates for community centers and tributary areas as of July 1, 19^7The fourth and final step involved the computation of the ratio of physicians to population for each of the communities. This was made possible through records kept by the Michigan State Medical Society and the various county medical societies throughout the state. The ratio of doctors to population in the 30*+ medical service areas in Michigan is presented pictorially in Figure 3. One will observe that the physician-popula- tion ratios for the 30*+ medical service areas in Michigan vary NUMBER OF PH Y S I C l AN - P O P U L A T I O N COMMUNITIES 25 70 58 jH jM RATIO A M = 1 p h ysician = fOO people each r. 42 17 .„ p -.p, * -r P ' ,■ r' V. ‘ .- ^ . A -1 j. • .. . Lv. : c: • ‘: : it c: i r v,' "c ■’ '■ ra:* c-l 1 c . - ni 4 - aaaa jt*~AAAAM jb-AAAAAAM i-AAAAAAAAA MMMMMMMMUlx r - r..'c; - 70 r- ' " t *i- ■ • T 'rr'. / •* 'V' * *P-^ £ ‘1 1 / ^ ^' * r*^'' r'C" ’ " ^^; C h' V’ 0 Oli 'i - -•;-•* - n '- ' ■ 7 ■ > ' /' ^ r 4 n ' - ‘■ y / * ‘ * ^^ • 31 widely. At one extreme were the 25 service areas with ratios ranging from 1 /5 0 0 to 1/1,000; while at the other extreme were the 17 service areas whose ratios ranged from lA-,000 to 1/10,000. In addition to the 282 medical service areas whose physician-population ratios are shown in Figure 3, there were 22 communities in the state of Michigan, with populations ranging from 1,000 to 5»000 people,which had no resident medical doctors at all. According to the traditional crude method of determining medical service adequacy (one doctor per 1,000 people), up to 92 per cent of the medical service areas in the state are subnormal. As a matter of fact, 151 (50 per cent) of the service areas in Michigan have ratios which are more than double this generally accepted ratio. It is undoubtedly true that many areas in the state of Michigan are not adequately staffed with doctors, but one should not necessarily conclude that this deficiency in doctor personnel is directly pro­ portional to the above ratios. Thetraditionally method of measurement actually does usedcrude notdefine or "measure" any of the relevant factors involved— the area, medical needs, or medical resources— thus rendering the physician-population ratios per se relatively meaningless. Testing the Natural Area Concept as it Relates to Medical Services In any research project it is essential that the basic assumptions on which the investigations are based be critical­ ly examined and tested, rather than taken for granted or 32 blindly accepted on an £ priori basis. Thus, in this dis­ sertation a great deal of consideration is given to the man­ ner in which these basic hypotheses are tested, and to the results obtained from these tests. The natural area concept has long been a useful concep­ tual tool to the sociologist. This chapter is concerned primarily with the description of how the concept of natural area was tested; one of the basic hypotheses being that if the natural area concept is valid for services in general, it is equally valid for medical services. The first step in the testing procedure was to select one specific service area^ from among the 3 0 *f general medical practitioner service areas in the state of Michigan. This would serve as a test case and would permit an intensive investigation of the specific question which is being inves­ tigated, namely, "Where do families go for their medical services?" The second step was to select, from all of the families interviewed in the community, those families who reported that they had a certain doctor to whom members of the family went for most of their ills. The specific wording of the question was, "Do you have a certain doctor to whom you (and 3 The medical service area which was used for this testing procedure was the Stephenson-Daggett community in Menominee County in the upper peninsula of Michigan. An intensive health survey of this community had been conducted in 19^7» prior to the Michigan Health Survey, but the methods and techniques employed were the same. 33 members of your family) go for most of your ills?11 The third step was to record the place of residence of each of these families, and the office location of the doc­ tor to whom family members went for medical services. The fourth and final step in the testing procedure was to trans­ fer the above data on a large map of the area, on which had been delineated the boundaries of the medical service area. Out of the total of 99 families on whom this information was available, 85 families (85 percent) reported that medical services had been obtained within the medical service area which had previously been delineated for this community, and the remaining l1* families (15 percent) reported that these services had been obtained in an adjoining area. Of these fourteen families going outside of the area, four were living near its borders and would have been placed in a different area according to a later and more accurate delineation of this area. (See families numbered V, 5» 7> and 36 in Figure b.) Fifty-two families in the sample used for the test lived in the villages of Stephenson and Daggett, and the remaining forty-seven families lived in the contiguous service area. The location of the non-village families and the place where each family obtained its medical services is shown in Figure b. A companion study, using data obtained from the Michigan Health Survey, showed similar results. Of the 631 families on whom this informs.tion was available, 502 (8 0 percent) d 5L d 35 sought services within their medical service area, 90 families (1*+ percent) had gone to an adjacent area, and the remaining 39 families (6 percent) had sought these services in a nonadjacent area from their place of residence. These data are shown graphically in Figure 5» and by counties in Table 1. Testing the Natural Area Concept as It Relates to Hospital Services The natural area concept was also tested as it relates to hospital service areas. The same method was used for both testing procedures, and the testing results were surprisingly similar. Of the available, 72 82 families on whom this information was (8 2 . 8 percent) reported that hospital services had been obtained in a town or city which was located within the hospital service area in which they resided, 7 families (5-7 percent) reported that hospital services had been sought in more than one service area. (Figure 6). The data given above refers to the use of the small (secondary) hospital areas. When the major hospital areas were used as the basis for delineation, 88.5 percent of the families were found to have obtained these services within their designated hospital service area, and an additional 9.2 percent in adjacent areas. On the basis of the major hospital service areas, therefore, fewer than 3 percent of the families reporting use of hospital services travel further than to an adjacent hospital service area. (See Table 2). A An.TAr-^NT 'TT^TN Tom i1 i ps 37 able 1 Number of families seeking medical ser­ vices in designated areas Number of families seeking medical ser­ vices in designated areas C ounty Within Adiacent lcona llegan renac araga arry 5 7 2 0 2 0 2 0 1 4 0 0 0 0 0 Keewenaw Lenawee Livingston Luce Macomb 1 4 2 1 14 0 0 0 0 11 0 0 1 0 4 ay errien ranch alhoun ass 16 7 5 14 1 1 7 1 2 1 0 0 0 1 Marquette Menominee Midland Missaukee Monroe 5 1 21 1 18 0 0 0 0 0 0 0 0 0 3 heboygan hippewa linton elta ickinson 4 6 0 9 10 0 0 0 3 0 0 0 1 0 0 Montcalm Montmorency Muskegon Newaygo Oakland 8 2 11 3 39 1 1 8 0 10 0 0 1 0 4 aton enessee ladwin ogebic r. Traverse 13 39 4 7 11 1 1 2 2 0 0 2 0 n J. 1 Cntonogan Osceola Ottawa Saginaw Sanilac 1 2 3 42 9 0 0 5 2 3 0 0 0 0 0 2 0 c, c. 2 2 0 1 2 1 4 0 2 2 Shiawassee St. Clair St. Joseph Tuscola Van Surer 10 15 2 6 q 0 1 1 1 1 0 0 0 0 0 1 0 1 0 0 0 1 1 7 Washtenaw 6 5 — — - 1 - ‘ -n O M County Humber of families seeking medical cere within "their general medical practitioner service area, in an adjacent area, or in a non-adjacer.t area in Michigan, by counties, 194-8. 90 ... 39 80 14 6 onia ron ackson alanazoo ent 3 2 24 14 42 1 — — Adi a cent Non-ad.i,, — 2 TOTALS Per Cent i 15 Within ---- 1 ratiot illsdale oughton uron ngham Non. ad.i Table II Number of families seeking hospital services within their hospital service area, in an adjacent area, a nonad jacent area, or in two different areas in Michigan, by counties, 194-3. County Alcona Bay Berrien Calhoun Cass Number of families seeking hospital s orvices in designated areas Within j Adjacent Non-adjacent 0 j 0 1 42 0 0 1 0 0 1 1 2 0 0 Two areas 0 0 0 0 0 Chippewa Delta Eaton Gennessee Gladwin 1 2 1 9 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Gr. Traverse Gratiot Hillsdale Huron Ingham 2 1 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 0 2 3 0 T_ 0 0 0 0 0 0 0 0 1 0 1 1 0 3 0 0 0 0 Jackson Kent Keewena Macomb Marquette Midland Missaukee Monroe Montcalm Muskegon Oakland Ottawa Saginaw Sanilac Shiawassee St. Clair St. Joseph Tuscola Van Buren Washtenaw Houghton Gogebic 0 1 3 1 6 1 1 0 0 c 0 1 0 0 0 1 0 0 c c 0 0 0 0 0 0 0 0 1 O 1 o 2 2 72 32.8 0 1 c 0 0 0 7 0 0 0 0 0 o • CO TOTALS PERCENT 1 1 0 1 i I L c 0 0 1 0 0 0 0 0 0 0 0 0 0 0 3 ,_3.5..... . r 5 5.7 j i 39 HOSPITAL SERVICE AREAS bO Conclusion On the basis of the above tests of the natural area concept, it was concluded that the use of the natural area concept for the determination of both medical and hospital service areas is not only a useful conceptual tool but also a valid one in the field of health and medical care research. It is probable that if medical service areas were de­ lineated with optimal accuracy, and proper consideration given to such factors as size of service center, nearness to other centers, relative "drawing power" of adjacent centers, etc., approximately 90 percent of the families within an area would be found to have obtained medical services within the area. If, on the other hand, the boundaries of the natural area are either under- or over-extended, this proportion would be correspondingly lowered, thereby decreasing both the reliability and the usefulness of the concept. In Figure *+, for example, the southern boundary of the service area was too far extended and therefore includes several families (numbers *f, 5> and 7) theoretically would obtain (and actually did obtain) medical services at a different service center. Proper allowance was not made for the "drawing power" of the two cities to the south, each of which has a population in the neighborhood of 10,000 people. The crude method which has been used in the past does not even define the area within which physicians function or in which people reside who are their potential patients, and b-1 therefore any attempts to measure either medical needs or resources have neither statistical reliability nor practical utility. CHAPTER V MEASURING PHYSICIAN RESOURCES Measuring the Service Capacity of General Practitioners In studying the problems of medical care in a given community, it is often desirable to make some appraisal of the physician resources of the community. It has long been the popular conception in this country that one physician for every one thousand population constitutes adequate physi­ cian personnel for a given area. The most basic hypothesis underlying the operational method, however, is that one physician for every one thousand population does not neces­ sarily constitute adequate medical personnel for an area, for not only does the need for medical services vary according to age, income, and other factors, but the ability of physicians to meet these needs is also affected by several significant variables, one of which is the age of the physician. Elliott H. Pennell, statistician for the United States Public Health Service, holds the same point of view and states that recent findings by various groups working with problems of medical care in this country indicate that a plain numerical count of physicians does not provide a true picture of avail­ able resources.First of all, there is considerable evidence Elliott II. Pennell, "Location and Movement of PhysiciansMethods for Estimating Physician Resources," Pub. Health Rep.. Vol. 59 (March 3, 1 9 W , pp. 281-305. to indicate that concentration of physicians is correlated with financial ability to purchase care rather than with the need for care. Secondly, Pennell states that existing data indicate that the number of patients seen by the private practitioner declines steadily on the average after age *+0. Thus it becomes necessary to interpret physician resources not in terms of the number of physicians alone, but in terms of the "service capacity1' of these physicians.2 Pennell, by applying life-table techniques to data concerning the patientcarrying capacity of physicians at different age levels, has been able to evaluate in a quantitative manner the cumulative effects of changes in professional capacity associated with aging and other factors. He has developed constants to take into account the retirement of physicians as age advances and has translated the resulting adjusted physician totals in terms of service capacity. Aided by data gathered by Ciocco and Altman^ concerning the average weekly patient load of physicians, Pennell determines these service equivalents in the following manner: 2 "Service capacity" and "Service equivalent" are used interchangeably by Pennell and are defined as the decimal fraction obtained by dividing the average number of patients seen by a physician of designated age by the corresponding number seen by a physician at the peak of his career; i.e., between the ages of 38 and *+0. 3 Antonio Ciocco and Isidore Altman, "The Patient Load of Physicians in Private Practice— a Comparative Study of Three Areas," Pub. Health Ren.. Vol. 53 (Sept. 3, 19^3), pp. 1329-13 51. ••.By reducing values for average weeklypatient load at each 5-year level to relatives, with the maximum- - 1 7 0 patients at age *+0--as unity, a series of adjusted decimal fractions are provided which reflect change in activity with age. When these data are plotted over an age scale and a smooth curve is fitted thereto, approximate measures of a physician's capacity for service at each year of life may be read from the curve Decimal fractions have been worked out for each year of age of a physician's life. Based upon these service equiva­ lents, a physician reaches his peak of efficiency, in terms of number of patients seen, between the ages of 38 and J+O. At age 57> for example, the average physician's capacity is approximately two-thirds as great as at age *+0, at 65 it is one-half, and at age 71 the average number of patients is only one-third as great as at age *+0. Table 3 shows service equiva­ lents of physicians at each age level.^ Measuring the Patient Load of General Practitioners The determination of the patient load of general practi­ tioners is an essential step in the development of the opera­ tional method, and is concerned with the second component of the problem, namely, the appraisal of physician resources. Extensive studies by Pennell have demonstrated that, in general, the service capacity of a physician varies with his age. For our purposes, however, the mere determination of a physician's service capacity, expressed as a decimal fraction, has ^ Elliott II. Pennell, up. cit. p. 298. ^ .Ibid., adapted from Table 5> P« 298. *+5 'able III. Service capacity of all physicians expressed as a product of decimals showing their service capacity at different aye levels, United States, 1940. Year of Age Service Capacity Year of Age Service Capacity 25-26-26-27-27-28-28-29--29-30-- 0.10 .24 .37 .48 .59 65-66--66-67-67-68-t'"—69— — 67-70-- .47 .45 .42 .40 .38 30-31-31_32-32-33-33-34-34-35-- .68 .76 .83 .88 .92 70-71-71-72-72-73--73-74-74-75-- .35 .33 .31 .29 .27 35-36--36-37-37-38-38-39-39-40-- .95 .97 .99 1.00 1.00 75-76--76-77-77-78-78-79-7<--80- .25 .23 .21 .19 .17 40-41-41-42-42-43-43-44-44-45-- 1.00 .99 .98 .97 .95 81-82-82-83-‘3-^4-86-85-- .15 .14 .12 .11 .j8 45-46-46-47-47-48-48-49-49-50-- .93 .'91 .99 .67 .84 •7.-87-4 7-c S__ Sc-8 9-8 9-90-- .08 .07 .06 .05 .04 5u-51-51-52-52-53-53-54-54-55-- 4 1 .79 .77 7C • /*0 * 3— — 91-92-92-93-93-94-94-95-- .03 .02 .02 .02 .01 55-56-56-57-57-58-58-59-59-60__ .7u .68 .' 6 .63 .61 r-5-96__ 9c-97-95-98--98 -99-uf _ 1 ^ 0 ____ .01 .01 .01 .ol 6o-61-c1-62—— 62-63-63-64-64-( 5-- .59 .56 .54 .52 .50 * Elliott K. Pennell, op. cit., Adaptec from Table 5, p. £9? * b6 relatively little significance unless this capacity can be expressed in terms of patient load. Recent studies by Ciocco and Altman furnish valuable data on this aspect of the prob­ lem.^ Their data on the patient load of physicians were derived from successive studies with the Procurement and Assignment Service for Physicians, Dentists, and Veterinarians made in the District of Columbia and in the states of Maryland and Georgia in 19^2. These studies were conducted by means of questionnaires sent to all physicians in private practice. On the basis of the data gathered by Ciocco and Altman from these three areas— District of Columbia, Maryland and Georgia— the authors concluded that physicians apparently reach their peak of activity (in terms of number of patientcalls) between the ages of 35 to W . The general practitioners under 35 years of age were found to have on the whole a slightly lower average, but nevertheless one that is higher than the average for the ages b5 and above. Physicians 65 years of age and older showed the lowest load in terms of patient calls, from one-third to one-half that of the men between 3 5 and Mf. The average weekly patient load for both rural and urban physicians between the ages of 35 and in these three areas was found to be nearly identical— 160 patients. The physicians in Baltimore, in the District of Columbia, and in the six counties of Georgia containing 50,000 population or more were & Ciocco and Altman, oja. cit. Also, by the same authors, see: "Statistics on the Patient Load of Physicians in Private Practice," J. Am. Med. Assoc.. 121:506-513 (Feb. 13, 19^3)• *+7 classified as "urban.'1 The physicians in the 23 counties of Maryland and in the remaining 153 counties in Georgia were classified as "rural." According to these studies by Ciocco and Altman, the average weekly patient load for rural physicians was found to be nearly identical with that of urban practitioners. One might assume that less favorable traveling conditions, greater territory to be covered, and other conditions generally asso­ ciated with the practice of medicine in rural areas, would have the net result of lowering the average patient load of rural practitioners. But this assumption is apparently not supported by the facts in the Ciocco and Altman studies. The successive steps in the determination of these measurements— the need for medical care on the one hand, and the general practitioner medical personnel available to meet these measured needs on the other hand— will be described in detail in Chapter VIII. The data in Chapter VIII were obtained from two communities in Michigan and demonstrate how the opera­ tional method can be employed in actual community situations. There are undoubtedly other variables which would affect the service capacity of medical doctors, such as partial retirement because of illness or age, part-time participation in other vocations or avocations, etc. As a matter of fact, most of these variables have already been taken into consid­ eration by Pennell in his studies which led to the determina­ tion of the "service equivalents" described earlier in this chanter. It is probable, however, that further refinements A *4-8 could still be made in this component of the operational method, thereby increasing the validity of the operational procedure. Although additional methodological refinements could still be made, the chief purpose of this study has been to sharpen the exceedingly crude methods of measurement used heretofore. ^9 CHAPTER VI MEASURING THE NEED FOR MEDICAL CARE The Medical Needs Schedule The function of this component of the operational method is that of developing a statistical measure of need for medi­ cal care in a given community cr trade area. It was assumed by those who developed the Medical Needs Schedule that a method to measure medical needs, both in terms of their pres­ ence and the extent to which they are met or unmet, would have to be relatively simple and inexpensive if it were to be used with large numbers of families. The ideal procedure would be to give each individual in the community a thorough medical examination, but this would be both expensive and impractical. Consequently, an attempt was made to devise a method that would make it possible for qualified lay interviewers, trained in the use of the Medical Needs Schedule, to obtain the desired information by the survey method. The essential elements of the Medical Needs Schedule are as follows 1. A list of questions concerning the presence or absence of certain specified symptoms, ailments, Charles R. Hoffer and Edgar A. Schuler in cooperation with Drs. Neligh and Robinson, "Determination of Unmet Need for Medical Attention Among Michigan Farm Families," J. Mich. State Med. Soc .. Vol. XLVI (April 19^7), PP. ^> + 3 . (Des­ cription of elements of Medical Needs Schedule quoted from above). 50 or conditions. These questions are based on those commonly used by doctors in taking a patient's medical history. 2. The answers to the questions are recorded by sym­ bols to show whether a person reporting a positive symptom has seen a medical doctor about it, has seen a non-medical practitioner, or has done nothing about it except possibly apply home remedies. 2 Thus it is possible to determine the number and the proportion of persons in a com­ munity having unmet medical needs. 3. The questions are of sufficient medical importance that if any one is reported positive for an indi­ vidual, that person is regarded as having a need for medical attention, at least to the extent of consulting a physician for a thorough diagnosis. if. These questions are phrased in simple, every-day language, so that the informants can readily under stand them. 5. Simple formulations of questions about symptoms are used to enable a person without medical The question has been raised that no recognition has been given to the medical services performed by such medical practitioners as osteopaths and chiropractors. The reader is reminded that the Medical heeds Schedule was constructed in such a manner that the type of medical service received would be reported for each positive symptom: 1 -home remedies, 2-practitioner other than M.D. and 3-medical doctor. The Schedule was constructed in this manner on the assumption that, in general, the medical doctor (M.D.) is the best qualified to diagnose and prescribe, other practitioners less qualified, and the individual patient the least qualified. In any continuum there is danger of over-simplification, but this is unavoidable, and until a better system is devised the use of the present method will be continued. There is probably a tendency to underestimate the services of non­ medical practitioners for the following reason: informants may report that they have seen a "doctor11 (meaning an osteo­ path, chiropractor, etc.) for a certain ailment in which case the interviewer might interpret and record the answer as meaning a medical doctor (M.D.). The reverse situation, however, (i.e., reporting of an M.D. for some other type of practitioner) would occur very seldom. A constant error would thus be intro­ duced in which the services of non-medical practitioners would be underestimated. 2 d 51 training, if properly instructed and supervised, to obtain the desired information. A set of instructions has been developed to help stand­ ardize the procedure. 6 . The procedure calls for one adult member of the family, preferably the housewife, to give the information about herself and other members of her family for the six-months period preceding the date of the interview. The development of this method, called the Symptoms or Medical Needs Approach, has been taking place during the past five years. In the summer of 191*1* Dr. Edgar A. Schuler, then a staff member of the Division of Farm Population and Rural Welfare, Bureau of Agricultural Economics, in the United States Department of Agriculture, began the development of a series of questions designed to measure health conditions and health care needs of farm families. These questions constituted one section of a comprehensive exploratory family standard and level of living schedule, including all major phases such as food, clothing, housing, transportation, and education. In the development of the section of the schedule on health and health care a number of United States Public Health Service Officers, particularly Drs. Frederick D. Mott, Milton I. Roemer Henry B. Makover, and Charles L. Williams, Jr. gave invaluable technical assistance. The exploratory schedule was tested with a small number of families in a rural county in New Jersey. The schedule has undergone several revisions since that time, and has been used in one form or another in the states of Alabama, Georgia, Michigan, Mississippi, New York, North Carolina, Pennsylvania, 52 South Carolina, Tennessee, Virginia, and Washington. It has also been used in a Spanish translation in Cuba and Peru. The original schedule contained over forty symptoms questions, but the revised schedule being used at the present time contains twenty-seven. A number of genito-urinary symp­ toms which tended to cause embarrassment were dropped and other changes, designed to give maximum reliability to the measuring instrument, were incorporated in the schedule. It is recognized that the final list of symptoms in the Medical Needs Schedule does not cover all the significant points which would be included in a physician's questioning of a patient for his medical history. The questions which were retained, however, were regarded by the cooperating physicians as having unquestionable medical significance. Figure 7 is a reproduction of the symptoms page of the Medical Schedule. (See rows numbered 19 through *+5). Validation of the Medical Needs Schedule One of the essential characteristics of a measuring instrument is that its measurements be valid, thus it was necessary to design a crucial test of the Medical Needs Schedule before this approach could be used with confidence. Such a test was conducted in the spring and summer of 19^6 when, in a study of the medical needs of farm families as part of a project sponsored by the Michigan Agricultural Experiment Station, an experiment was carried out to test the ?3 (2) FART I M E D I C A L NEEDS (H,W,S,D.P.L.B.SS,4C.N) T " of Ho u s e hold Eolation Nade: to Hoad lieave B l a n k (Code for r e l a t i o n to head) 11 Indicate sex in e a c h column: 1-Male 2-Female 12 Indicate age in e a c h c olumn (01 - 99) 13-19 Leave B l a n k (Code for No. in family) 15 CidJT Tdtf i n f o rmant 0-1 -^-5 ~ l o 1 Enter "1" for e a c h p e r s o n present at interview; 17' Code for person's n u m b e r 18 1 2 — T h plained L o s s of Weight: Persons over 18: 10 lbs. or mo r e in 6 1 0 . Persons u n d e r 18: any u n e x p l a i n e d loss of w e ight 19 C o n t i n u e d Loss of App e t i t e 20 Un e x p l a i n e d Tire d n e s s : r e g u l a r l y 21 i R u n n i n g E a r or Ears: watery, bloody, pus 22 Poor Vision: for distant or close work, e.g. read i n g 23' R e p e a t e d N o s e b l e e d s Not Duo to B l o w or Injury 29 Persistent H e a d a c h e s 25 Toothache 26 U nable to C h e w Food: teeth "sore*' or m i s s i n g 27 Sore Mouth: due to p l ates or b r i d g e s 28 R e p e a t e d o r F r e q u e n t B l e d d l n g O u m s 29 ilstent Sk in R a s h e s or I t ching of Skin: -- " b r e aking o u t k (One w e e k or more) 30 L u m p s or D i s c o l o r e d Patches on Skin 31 Persisten t Pains in Chest 32 P e r s i s t e n t Cough: ^except colds '.n chest) 33 C o u g h i n g or Si it ting B l o o d 39 Severe Shortness of B r eath: after d o i n g light w o r k 35 A s t h m a or H a y f e v e r 36 R e p e a t e d or Per s i s t e n t B a c k a c h e 37 P ersistent ^alns in the J o i n t s 30 O p e n or R u n n i n g Sores or U l o e r s T h a t do not Heal 39 tated or P ersistent S w e lling of Ankles: T w o w e e k s or m o r e 90 Repeated Vomiting: (Several days or more) 91 sated or P rolo n g e d Pains in S t o m a c h or.Any w h e r e in A b d o m e n 92 R u p t u r e ", Hernia, o r W e a r i n g of T r u s s 93 ltlng Spells: Stuttering: Stammering: Nervousf B r e a k d o w n : Fits: C o n v u l s i o n s 99 Ldental Injuries: b r o k e n bones, head or severe injuries, accidental poisoning, snake bites, etc. 95 each p e r s on w i t h one or m o r e "l's" in column, riD o y ou think --- needs to see a doctor?" 1-Y, 2-N 96 each p e r s o n w i t h "1" in row 9&, ''What would y o u say is the n reason --- h asn't sean a d o c t o r ? ” 1-Lack of time 2 -Symptoms not thought serious 3-T00 e x p e nsive 9 - O t h e r (specify on r e v e r s e side) 97 each p e r s on w i t h one or more w3's'' in column, "Did the M.D. advise --- to go to a h o s p i t a l ? " 1-Y, 2-N 98 D id he (she) go? 1-Y. 2-N 99 L ea ve b l a n k (individual h e a l t h care code pa L eave b l a n k (total No. of positive symptoms 51 E n t e r "1" if not w o r k i n g n o w due to illness or inju r y 52 T o t a l days off in last 6 m o n t h s due to illness 5 3-59 N u m b e r of times --- has seen a d o c t o r in last 6 m o n t h s (a) at d o c t o r ' s office 55 (b) at y o u r home 56 57 T i m e s --- has seen dentist in last 6 m o n t h s V a c c i n a t i o n or I m m u n i z a t i o n for: (a) S m a l l p o x (all over 1 ye a r old) 58 (b) D i p h t h e r i a 1.(> m o n t h s to 16 years 59 (c) W h o o p i n g cough (6 m o n t h s to l5 years) 60 > 1 1 2 }' 9 6 1 5 8 7 ' ! _ 1 | I .i 1 1 ' - 1 fJS- 5 6 7 .8.J.-2- COPYRIGHT by SOCIAL R E S E A R C H SERVICE of I;ICl''IC-AN STATE COLLEGE. -Tune, I9US. Permission to reproduce any portion of this- schedule must be obtained in writing. Figure 7. Symptoms page of the Medical Needs Schedule employed in the Michigan Health Survey, Michigan, 1948. 4 in­ validity of information obtained by the use of the Medical Needs Schedule.3 A carefully selected sample of 306 families was chosen and interviewed in the three Michigan counties of Shiawassee, Kent, and Cheboygan. The counties themselves had been selected with due regard for such factors as density of population, physician-population ratio, purchasing power, and other factors, so that they would represent as nearly as possible typical areas with reference to medical service adequacy. Validation was accomplished by having a sample of the families who were interviewed come to a clinic to receive a medical examination. Then the findings of the medical exam­ ination were compared with the information obtained by the use of the Medical Needs Schedule. The medical examination was a typical general examination except for the fact that it was not possible to make a detailed check of eyes and teeth. The State Department of Health cooperated by providing facili­ ties for chest X-rays, blood test for syphilis, and laboratory services for analyzing urine specimens for albumen and sugar. The Medical School of the University of Michigan made avail­ able the services of physicians to do the examinations, and local public health departments assisted with details. In the validation study 153 persons, representing dif­ ferent families, received medical examinations, and in eight 55 out of ten cases there was agreement between the findings of the medical examination and the Medical Needs Schedule data regarding need for medical attention. The 80 per cent correspondence between the findings of the medical examina­ tions and the Medical Needs Schedule refers to the per cent of individuals having medical needs, and does not refer to a point-by-point correspondence on specific symptoms. That is, the schedule data were confirmed by medical examinations for 80 per cent of the individuals C^fO/S for persons needing medical attention, and for persons not needing medical attention), and not confirmed by medical examinations for per cent of the individuals 20 for persons needing medical attention according to the Medical Needs Schedule but not according to medical examinations, and l^'a for persons not needing medical care according to the Medical Needs Schedule but needing medical attention according to medical examina­ tions). Thus of the 20 per cent of cases not confirmed, for every person improperly classified as having need there were two persons improperly classified as having no need. (See Table 1+) It should be emphasized that the method does not presume to provide a diagnosis of the person's illness; that is con­ sidered to be the function of the doctor. Rather, what the responses to these questions are expected to provide is a statistical basis for measuring medical needs in a group of individuals which is sufficiently accurate to be useful in planning community programs of medical care and health education. A 56 It was concluded from the validation study that the correspondence between the data obtained v/ith the Medical Needs Schedule and the medical examination was sufficiently accurate to warrant use of the Schedule as a reasonably valid measure of unmet needs for medical attention among groups of families. Table IV Percent of 153 individuals needing and not needing medical care according to Medical Needs Schedule and medical examination in three Michigan counties, 19*+6* CONFIRMED: Per Cent Schedule data confirmed by medical examination: 80.0 Persons needing medical attention *+0.0 Persons not needing medicalattention ^0.0 HOT CONFIRMED: Schedule data not confirmed by medical examin­ ation: Persons needing medical attention according to Medical Needs Schedule but not according to medical examination 20.0 6.0 Persons not needing medical care according to Medical Needs Schedule but according to medical examination needing care l^-.O Methodological Problems The authors of the operational method are aware of numerous methodological problems connected v/ith the determin­ ation of medical needs by the use of the Medical Needs Schedule. ♦Adapted from Koffer and Schuler, ojd. cit., Table 1. 57 First, it is recognized that the schedule does net cover all symptoms which might be of significance in the complete diagnosis of a patient's ills. As h s been stated, it was deemed advisable to omit certain questions pertaining to genito-urinary symptoms, etc., to avoid possible embarrass­ ment to both the lay interviewer and the interviewee. more­ over, it is possible that a constant error of undermeasurement exists in the determination of symptoms relating to acute illnesses, since it is logical to assume that symptoms re­ lating to chronic illnesses and current acute conditions would be most readily recalled whereas symptoms relating to acute conditions which had existed in the past would bo less readily remembered. Secondly, the schedule does not have the flexibility to follow up diagnostic cues in the same manner as would a physician in his examination of a patient. Standardization of the schedule, however, seems unavoidable and is in fact essential in the interest of greater reliability. To the extent that flexibility and complete coverage of symptoms are sacrificed in the interests of greater reliability the symptoms approach as here employed in all probability yields an underestimation of medical needs amounting to perhaps A third problem is essentially one of definition. 8 p>. The question has well been raised, "When does a need exist?" This question is of basic importance and must be give a con­ sideration. Does a medical "need" exist for an unvaccinated 58 child, who obviously would have no symptom indicating "need" for the preventive service of vaccination? If one or more untreated positive symptoms is reported, does a need exist at the preventive, the diagnostic, or the remedial level? In other words, if an individual reports a symptom— be it real, functional, or purely imaginary— does a "need" exist equally at all levels? Concerning this problem the writers acknowledge that a "need" may exist at any or all of the levels mentioned. Because of the nature of the measuring instrument, however, need for preventive services cannot be measured: functional illnesses are included only insofar as the symptoms related to this type of illness are listed on the symptoms schedule; and, since need for treatment can be appraised only on the basis of appropriate medical examinations, tests, and so on, extent of remedial need cannot be determined by the symptom which is reported for any individual is regarded primarily as a diagnostic "need", the diagnosis being performed in a com­ petent manner by a properly qualified physician. The writers are aware of a fourth methodological problem, namely, the reliability of data, involving a memory factor, furnished by one adult informant for the entire family. Symp­ toms are obtained for the six-months period preceding the date in the interview on the assumption that errors due to the mem­ ory factor will be considerably less than if the informant were to be asked to recall symptoms over a period of a year. Also, it appears reasonable to assume that the informant might re­ call his own symptoms more readily than those symptoms involving other members of the family. The Michigan Health Survey 59 throws considerable light on the latter problem. In this project the information given by the housewife concerning her husband’s symptoms will be compared with the information given by the husband concerning his own symptoms. Analysis of these data should provide some measure of the reliability of data furnished by one adult informant (usually the housewife) for the entire family, and thus aid in the development and refine­ ment of the symptoms approach. To the extent that the memory factor can be shown to be significant, even for the six-months period, the Medical Needs Schedule probably represents an under­ estimation of the need for medical care. Reliability in the Family Data Emvolafagfl by t&e .ho.ugewiX.e XflX 0.ther Adults One of the major objectives of the Michigan Health Survey, from the point of view of certain personnel of the Social Research Service, was the collection of reliable data which would make possible the testing of basic hypotheses and tech­ niques employed in the operational method. In order to test the reliability of data provided by the'rousewife for other family members, a random sample of 182 adults, other than the housewife, were interviewed regarding their medical needs, and these data then compered with those furnished by the housewife for these 182 adults. A comparison of the reporting of need for medical care by the female head of a household on other adults, with reports given by these adults on themselves, has been made by Dr. Duane L. dibson of the Social Research Service. T’ ^e following quotation from the above rerrort summarizes the d results of the testing procedure:^ "Any attempt to compare the reporting of need for medical care which is given by the female head of a household on other adults, with reports given by those adults on themselves, obviously needs to consider-(The purpose for which the Medical Needs Schedule is to be usecQ. Our focus of attention hss been restricted almost exclusively to evidence of "need" for medical attention rather than upon individual symptoms or combinations of symptoms. For purposes of our research in this area, every item of the symptoms page is a '.stop item;' that is, the presence of anv untreated positive symptom is evidence that a person needs medical (or dental) attention. By this criterion, any tendency on the p rt of the female head to underenumerate symptoms on another adult was not considered serious, for our purpose, unless she failed to report anv untreated symptoms which were actually present. With this purpose in mind, the table shown below represent the indication of need for medical care expressed by the female head on another adult in the family (in the majority of cases this would be the male head) as compared with that person's report on himself. It should be noted that in all 182 cases, the "other adult" was not present when his symptoms were being reported b'r the female head. It should also be pointed out that the comparison here provided refers only to the reporting uJnpublished data of the Social research Service, Jept. of Sociology and Anthropology, Michigan ^tate College, East Lansin Report prepared by Dr. Duane L. Gibson. 61 Table V. A comparison of the reporting of need for medical care by 182 housewives on other adults in the household, with rejorts by these adults on themselv< Michigan Health Survey, 194-2 High level of agreement: (A)* Female head reports no rositive symptoms for other adultj reporting corroborated by other adult on self (A) Female head reports one or more positive symptoms and indicates that all symptoms were treated by an I!.D., a non— K.D., or both, reporting corroborated (E)* Female head reports one or more positive symptoms none of them treated; reporting corroborated (E) Female head reports one or more positive symptoms untreated and one or more treated; reporting corroborated Number Percent 69 38 22 15 25 H 128 Sub-total, high level of agreement 70 Substantial agreementI (A) (A) (B) (B) Female head reports jgo positive symptoms for other adult; other adult reports one or more symptoms which have been treated Female head reriorts one or more symptoms which have been treated; other adult reports jag positive symptoms Female head reports one or more untre ated symptoms; other adult reports one or more untreated symptoms and one or more symptoms which have been treated Female head reports one or more untreated symptoms and one or more symptoms which have been treated; other adult reports only one or more untre ated symptoms 5 3 5 3 2 U Sub— total, substantial agreement High level of disagreement: (?)* Female head reports jqo positive symptoms for other adult; other adult reports one or more untreated symptoms 15 (?) Female heed reports no rositive symptoms for other adult; other adult reports one or more symptoms untreated and one or more symptoms which have been treated (?) Female head reriorts all symptoms treated; other adult reports one or more symptoms untreated as well es one or more symptoms treated (?) Female head reports one or more untreated symptoms; other adult reports rjg untre.r.ted symptoms (?) Female head reports one or more untreated and one or more treated symptoms; other adult reports jQfi untre ated symptoms (?) Female head reports one or more untreated and one or more treated symptoms; other adult reports only tree ted symptoms Sub— total, high level of disagreement Total (a ) Evidence of no medical need (b ) Evidence that there is medical need 2 4- 4-0 182 22 100£ 62 of female heads on adults (persons 21 years of age and over) and the reporting of those adults on themselves. Comparison of female heads and minors is not available. The items classified "substantial agreement" have been placed in that category because, for our purpose, there is corroboratory evidence of need or lack of need for medical care between reporting by the female head on another adult and the reporting by the other adult on himself. The items under "high level of disagreement" are classi­ fied because the record of symptoms and their care or lack of it provide no corroboratory evidence of need or lack of need for medical attention. types. The disagreement is of two basic The first three items (27 cases or 15/'* of the total) represent instances in which data provided by the female head indicates rjQ need for medical attention while data provided by the other adult would indicate that there jLs medical need. The second type of disagreement (13 cases or 7/* of the total) is made up of instances in which the female head provided data on another adult which indicated need for medical care although interviews with that adult did not corroborate the need. If this situation held true for larger samples it would seem logical to conclude that dependence on the fema.le head as informant on symptoms for all adults would result in about an 8 /j (15/* - 7 h) under-enumeration of need for medical care." d 63 CHAPTER VII VARIABLES AFFECTING- NEED FOR MEDICAL CARE Introduction The primary function of thin chapter is to demonstrate by means of quantitative data obtained by a state v/idc survey of the medical needs of 3,786 individuals in the state of Michigan, that the need for medical care is not a constant, but varies in relation to several significant variables. The data on the following pages demonstrate the fallacy of merely counting "people", and assuming that the medical needs of different segments of the population are constant and un­ changing, as those who have employed the crude method of determining medical service adequacy have falsely assumed. Rather than replace the Medical Needs Schedule, an analy­ sis of the principal variables affecting medical need merely demonstrate conclusively the need for on unbiased, easily ad­ ministered measuring instrument by means of which the medical needs of a group of people can be reliably estimated as a basis for more intelligent social planning in the field of medical and health care. Moreover, it is conceivable that an anlysis of the variables affecting medical need in a given community situation could supplement the medical needs schedule by providing an additional facet of quantitative information regarding the medical needs and characteristics of the people in a given stratum or group. 6MA second function which such an analysis could perform is that of prediction. Table IX, for example, page 77» shows the estimated average change in number of positive symptoms per each additional year of age, within eight income levels. The regression coefficient (b) in all cases is positive, and declines consistently from the "0" income level (under $1,000) where it is -f 0.039 to income level H6 H ($5,000-$7 »^39) where it is only +• 0.006. Because of the large size and representativeness of the state-wide sample used for the Michigan Health Survey, the data relating to the medical needs of the 3»7S6 individuals on whom thie information wae obtained was felt to be ideal for the purpose of analyzing need ip. relation to such factors as age, income, residence, etc. The first step in such an analysis was to construct charts showing the relative effect of each of the independent variables (age, income, etc.) upon the dependent variable (the need for medical care), when other factors are uncontrolled. Following this initial investigation, these variables can then be subjected to more detailed statis­ tical analysis to determine the effect of each factor when the influence of other variables is eliminated. Conclusions Regarding Correlation Analyses The following evidence, basea upon a statistical analysis of the data obtained in the Michigan Health Survey, is presented in support of the author's hypothesis that the need for medical care is not a constant, but varies with respect to such signifi­ cant variables as age, income, etc. For the sake of clarity, 65 these variables will now be considered independently as they affect the need for medical care. A. Age as Factor in Medical Need The factor of age as a variable affecting the need for medical care is apparently an important one. The first step in the analysis of this relationship war to construct a chart showing the effect of age alone, when none of the other varia­ bles was controlled. (See Figure 8) The need for medical care among different age groups, as measured by the number of positive symptoms for each of the 3>776 individuals on whom this information was obtained, appears to increase consistently with increasing age. in Figure 8. This relationship is shown graphically We must be cautious, however, in interpreting this chart, and in assuming that such a convincing trend as indicated by this chart rctually exists. In other words, this trend appears to be present, but we do not actually know how much of it may be due to income, residence, or other variables. This can be determined by statistical analysis of the lata from which the chart was constructed. It was first thought that: this objective could be achieved by multiple correlation techniques, which, it was believed, would show the relative importance of age, income, etc. upon the need for medical care. These techniques proved to be unsuitable, however, because of the non-homogeneity of the groups with which we were working. Thus it was found necessary to use partial correlation techniques. d CL F r e i u e n p 1es 57^ 74 5«1 BR7 4 33 751 36 3 1P1 100 Sly or morel Five our Two | r~l Q -4 - 3 o o .x*VI None TTnder 1 Yr. 1-9 10-19 20-99 30-39 40-49 50-59 5 0-69 70 "n Ovpr Ape proups !•i -uro :;ut:on o ‘ •','77 ‘. :n ::v : Je :o’i:1 4 67 The importance of age as a factor affecting the need for medical ca^e, when income war controlled in eight levels,^- is shown in Table VI. All linear correlations are positive and generally larger than when age was controlled. VI and VIII for this comparison). (See Tables In other words, the number of positive symptoms tends to increase with Increasing age in all income levels. With one exception, all correlations are statistically significant, and in this one instance, (that for income level six) it may be due to sampling error. Another feature regarding the correlation between age and number of positive symptoms is illustrated in the graph below. (Figure 9) The strength of the linear relationship between these two variables decreases cuite noticeably with increasingly higher income except for the last level. It was shown statistically, however, that the difference between the last two correlations (for levels six and seven) is reasonably assignable to chance when one considers the sizes of the samples (76 end 27 respectively). ■^These eight income levels are as follows; .Level 0 1 Range under .>1,000 2 hi,000 *12,000 - n2 , 0 0 0 > 3,000 3 m3, 000 - ,,lf,000 b >^,000 - >5,ooo 5 R5,ooo - v7,5oo 6 > 7 7,500 -*> 10,000 >10,000 or more 68 Table VI. ncome roup 0 1 2 3 4 5 6 7 N-3 D/F N 299 594 1094 816 395 198 75 27 , -3498 — Significant * Significant ** Significant Correlations between age and number of positive symptoms within eight income levels. Michigan Health Survey, 1948. Z r .438** 296 .323** 591 1091 .245** .271** 813 .178** 392 .148* 195 72 .124 N .S. 24 .434* ..3474, , at 5/S level at 1% level at 0.1?' level (N-3) Z (N-3) Z 2 139.12 197.985 272.75 234.144 70.56 29.25 9.00 11.16 .47 .335 .25 .288 .18 .15 .125 .465 2.263: 65.3864 66.3250 68.1875 67.4335 12.7008 4.3875 1.1250 5.1894 290:7351 — Average Z = .277 X2 : 267.0194 23.7157*** ifference between 6 and 7 ^weighted Z * 20.16 * .210 96 (20 .16 ) (.210) = 4.2336 £ ( N - 3 ) Z2 = 6.3144 X 2 - = 2.0808 (ldf) average Z N.S.(Not Significant) ifference between 3 and 4 ■^weighted Z ■ 304.704 - .253 1205 and 7 75 27 average Z (304.704) (.253) = 77.0901 ^(N-3) Z 2 = 80.1343 X2 = 3.0442 (ldf) r Z l/(N-3) .0139 .124 .125 .0417 .434 .0556 .340 U.S. sd = .236 .340 = 1.44 N.S, 123& d .50 OORR L LATIONS •4-5 .4-0 .35 .30 .25 .20 .15 .10 3 4 INCOME Figure 9. LEVELS Correlations between age and number of positive symptoms within eight income levels. Michigan Health Survey, 194-8. (See accompanying Table VI). 70 B. Frequency Distributions Within Age and Income Levels It was thought that the higher income levels might in­ volve a significantly greater proportion of older persons, which then would need to be considered when interpreting the correlations discussed above. The following table (Table VII) presents the necessary distributions. Except for the lowest and highest income levels, the median age for the income levels ranges from 2b to 30, so considering this fact and the distributions, there does not appear to be any serious distortion in the distributions due to age differences. As a matter of fact, the only serious lack of homogeneity as regards either age distribution within income level, or income distribution within age class is caused by the 127 older persons lowest income level. (60 years or older) in the However, an analysis of the correlations fails to uncover any distortion due to that fact. C. Income as .a Factor in Itedlcal Need The first step in analyzing the importance of income with respect to the need for medical care was to construct a simple chart showing the relationship between Income and the need for medical care, when other variables were not controlled. This relationship is shown graphically in Figure 10. As was the case with age, this relationship appears to be a very sig­ nificant one. Here again, 1owever, one must be cautious in assuming that such a uniform relationship as indicated by the chart actually exists in all groups. The use of partial correlation techniques is thereiore necessary to deter..ine the Frequency distribution of 3*492 individuals in the sample o . p lation for the Michigan Health Survey, ly income and ore levels, Michigan, 1942'. Income i—1 G ro u p 2-5 6-12 0 1 2 3 4 5 to 7 7 34 47 30 1J 4 0 0 12 57 11? 94 53 20 0 s 4 23 61 12 3 149 66 32 19 1 27 56 93 94 39 25 10 3 361 474 2 .9 3 .2 — 152 3,498 i-iedian income 2.5 le v e l M e d ia n i n c o m e , ^ c v a lu e 1 299 594 1094 216 395 198 75 27 i—l No. Age Groups 0-1 "■ 50-59 ■ - j n r 2 0 -2 T 16 5 1 18 57 161 162 74 37 15 7 23 63 131 75 48 37 10 347 528 2.9 G / 29 S2 103 59 24 14 9 1 127 85 88 49 21 13 4 1 531 396 321 388 2 .4 3.2 2.1 2.5 1.2 2,900 43,200 $ 2,900 02,400 v3,200 42,500 41,200 > > 99 210 104 6o O O Table VII. Median age grnnp0 Median age 50-59 20-29 20-29 2 .-2 9 20-29 30-39 30-39 3o-39 57 29 27 24 75 25 3o 37 eouenclea [Five I isjy or morel 100 Four ThreeI Percent of income group One L None .7 , 000499 *9 mirr- :r . ' r cf r.:■ 'c li ex r L’ o u ;-ior. of f •:.] th ‘1 :.:-vo;r 2e • o u.l • inoo~o rrov • r, r r-CT 1 t V; CYT : t OTT:T » ,c--o , Vf>r du: i :n n ' ■•■c.-im r. 73 effect of income when the influence due to age, for example, is eliminated. The relationship between age and the need for medical care is, as one might expect, positive. In other words, as age increases there is a corresponding increase in the need for medical care, as measured by the number of positive symp­ toms which people have. The relationship between income and the need for medical care, r.ov/ever, appears to show a definite negative trend. This trend, also, night be expected. In other words, it appears from the chart that as incone in­ creases at all levels, the need for medical care decreases. A correlation analysis of this relationship shows that there is generally a negative linear correlation between the number of positive symptoms and the income level; i.e., the higher the income the fewer the number of positive symptoms. However, the correlations are low (none larger than -.2^5) and are statistically significant only for infants (where the correlation is positive), for adolescents (13 to 19 years), and for ages at and above *+0 years. T'or ages 2-12, and 20-39 there is no evidence of a relationship between income and number of positive symptoms. One might think that since even the statistically sig­ nificant correlations are low there are only sampling differ­ ences among the correlations for the different age levels. That such is an incorrect conclusion is shown by the signifi­ cant X2« 17.05, 8 degrees of freedom, obtained from a test of the hypothesis that the age level samples were drawn from a 7b conunon bivariate population.-*- These correlations, in tabular and graphic form, are presented in the accompanying table VIII and Figure 11. Table IX (regression coefficients) shows the estimated average change in number of positive symptoms per each addi­ tional year of age, within nine income levels. For example, at the 0 income level (under T1,000), the increase (C is positive) is estimated at .039/ year or 0.39/ decade. D. Other Variables Affectina Medica,! Need A chi square test of association between place of resi­ dence and number of positive symptoms, with age and income controlled, showed that there was some tendency for a higher proportionate number of symptoms to be present in rural and village areas. This was true particularly for ages at and above bo years, and appeared also for the age group 13-39 within the h*gher income group (income levels 3-7)• Because of the large number of cases required for an analysis of this kind, it was necessary to combine certain age and income groups in order to have enough cases to get statistical reliability. Accordingly, the age groups were combined into three groupings as follows; 0-12 years, 13-39 years, and bo years and over. The income categories wore similarly grouped into two categories; 0-2 income levels, and 3-7 income levels. Specifically, the chi square test of association between ISee Snedecor, G. W., Statistical hethods. (Uth L.dition) C oll­ egiate Press, Inc., Ames, Iowa (19*+6), p. 15*+. Table VIII* Correlations between income and nurber of positive symptoms within nine age levels* Michigan Health Survey, 194-8. Age group 0-1 2-5 6-12 13-19 20-29 30-39 40-49 50-59 60 N N-3 D/F r 152 361 474 347 528 531 396 321 388 149 358 471 344 525 528 393 318 385 .195* - .060 - .039 - .126* - .052 - .058 - .164** - .145** - .245** 3498 3471 * Significant at 5% level ♦♦Significant at 1% level 2 .200 .060 .040 .127 .052 .058 .166 .146 .250 (n -3)a (N-3)22 29.8 21.48 18.84 43.69 27.30 30.62 65.24 46.43 96.25 5.9600 .2888 .7536 5.5486 1.4196 1.7760 10.8298 6.7788 24.0625 379.56 . Averr ge 2 = .109 X2= 17.0457* 58.4177 41.3720 D/F= 8 Difference between 6-12 and 13-19 ^ weighted 2 = 62.53 = .077 Z815 6.3022 - 4.8148 = 1.4874 D/F = 1 U.S. D/F = 1 N.S. D/F = 1 K.S. Difference between 13-19 and 20-29 weighted 2 = 70.99 869 £ = .082 6.9682 - 5.8212 = 1.14-70 Difference between 30-39 and 40-49 ^ weighted 2 = 95.86 921 = .104 12.6058 - 9.9694 = 2.6364 76 £ _ . 27C £ ..2 5 C £-.23C £ _ .2 2 C o z £ ..1 9 C ' Z£ _ .1 7 C =c J £-.15C iJ c ^£^130 D -> iL-nc £..09C L-07C jL.osc *£_.03C 0 5 0-1 2-5 10 6-12 15 20 13-19 25 20-29 AGE 30 35 4-0 30-39 A5 40^9 50 55 50-59 GROUPS 0 Significant • Not Significant Figure 11. Correlations between income and number of positive symptoms within nine age levels, Michigan Health Survey, 194-o. 60 65 Table IX. Estimated average change in number of positive symptoms for each additional year of age, by in­ come level, Michigan Health Survey, 194-3 (re­ gression coefficients).* INCOME LEVEL_________b INCOME LEVEL________b________ 0 /0.039 U /0.010 1 /0.020 5 /0.008 2 /0.016 6 /0.006 3 /0.016 7 /o.oio * The regression coefficient is the estimated average change in number of positive symptoms per each additional year of age. For example, at the 0 income level, the increase (c is /) is estimated as .039/year or 0.39/ decade* 78 ' place of residence and number of positive symptoms shoved no association for the age sroup gories 0-2 or 3-7. age category 1 3 -3 9 0 -1 2 within the income cate­ The chi square test of association for the (income category 0 -2 ) lilccwise shoved no association. Beginning with age category 13-39, and income category there appears a definite tendency for a higher 3 -7 , proportionate number of symptoms to be present in rural and village areas. chi square was In the category just mentioned, for instance, 2 9 .2 6 with a probability of *£ _.0 2 with 12 degrees of freedom. For the age category *+0 years and over, within the incorae group 0 -2 , the chi square test was 21+.5 l with a probability of .02, and 12 degrees of freedom. For the same age group, at income group 2 2 .M+ 3 -7 , chi square was of Z_.o5 and 12 degrees of freedom. with a probability The three chi square tests just mentioned are ail significant at the ?,-> level. A table of chi square, shoving the e;-:tent of association between place of residence and number of positive symptoms, within the specified age and income categories, appears in Table X. Concerning sex as a factor in medical need, no definite relationship is apparent eycept that the females have a slightly greater number of symptoms above the two level i.e., the number of females reporting three, four, five, and sir or more is consistently higher than the number of males reporting this number. This relationship is shown in table XI. 79 Table X. Chi square analysis of association between place of residence and number of positive symptoms within desig­ nated age and income categories, Michigan Health Survey, IS)AO. Group D/F 5.38 .30 9 age group 3-7 income group 7.77 >.20 6 13-39 age group 0-2 income group 17.21 V • o 0-12 age group 0-2 income group X2 12 13-39 age group 3-7 income group 25.26* ^ .02 12 'AO age group 0-2 income group 2/5.51* .02 12 Remarks No association 0-12 1 * Significant at 5% level. 12 1 22.AA« ! N ! 1 WoT 'A0 age group 3-7 income group Some tendency for higher proportion­ ate number of symptoms in rural and village areas, 80 Table XI. Number and per cent of 3,786 individuals in the sample for the Michigan Health Surveyreporting designated number of positive sym­ ptoms, by sex. Michigan, 1 9 *+8 . Number of Males Females ____________________________________ positive svmrtoms_____ No. Fer cent No. Per cent None 6 1 .0 1122 llM-8 58.9 836 2 3 .2 372 1 19.5 178 2 8.5 157 9.3 5.8 1 1 0 6b 3.3 3 1 . 8 2 .8 b b-5 33 5 .9 17 1.7 ?? bb 6 or more _ 2A 2.5,. . 1.3 1 0 0 .0 1880 1 0 0 . 0 _________ 1 906 Conclusions Regarding Variables Affecting heed For MeAical Care The function of this chapter was to demonstrate that the need for medical care is not a constant, as those who have employed the traditional ratio of one doctor per 1 ,0 0 C people have assumed, but that need varies according to several significant factors. The preceding analysis of the variables affecting the need for medical care demonstrates the need for an unbiased measuring instrument by means of which the medical needs 01 a community can be reliably estimated as a first step in planning community programs in medical and health care. PART III MEASURING MEDICAL REEDS AID RESOURCES I RURAL COMMUNITIES 81 CHAPTER VIII MEASURING MEDICAL NEEDS AND RESOURCES IN PLEASANTVILIE AND ELMWOOD * Apcllcatlon ££ the Operational Method to Community Situations By combining the components of the operational method, l.ey, the natural area concept, the measurement of physician resources, and the measurement of need for medical care, it is possible to make a quantitative appraisal of the physi­ cian personnel needs of a community or other population group. The two aspects of the problem which can be measured by this operational method have already been specified, namely, (1 ) the extent of- medical needs in the community, as determined by the presence of symptoms which, according to the opinion of medical doctors, should receive the attention of a physician at least to the extent of obtaining a thorough examination and diagnosis, and (2 ) adequacy of physician personnel in the community expressed in terms of: (a) number of physicians practicing in the community; (b) physicians1 potential patient­ load capacity, as measured by each physician's age; and (c) the need for physicians' services, based on measures of both met and unmet medical needs. Since both of the above factors- the need for medical care and the capacity of local physicians to meet the need - are expressed quantitatively and in terms of the same unit, this method makes it possible to comp.'re different communities or trade areas with regal’d to their medical service adequacy. *The names of these communities ere fictitious, but the data were *K+* o -i ^ rn n o l r> r i m i n i i m ' f v d h l l n t l i n T s if 1V ' l i c r o ,n - 82 For clarity it should be emphasized that this operational method applies primarily to rural communities or trade areas rather than large urban centers, and that the measurement of physician resources refers to general practitioners. It is not possible at the present time to ap^ly the method to urban areas because of numerous methodological problems which relate to the evaluation of the services of specialists, etc. The successive steps in the determination of these measurements will now be described, followed by an illustra­ tive example the data for which were taken from an actual community situation. A. Stens in Deter mi ninr: the Need for General Medical Practitioners in the Community: (1) Determine the number of persons in the sample population who have had all positive symptoms (as reported on the Medical Needs Schedule) treated by a physician. (2) Determine the number of persons in the sample popu­ lation who have reported positive symptoms, but who hove not seen a physician about all these symptoms. (3) Total the two quantities found in (1 ) and (2 ) and multiply this total by the number of physician calls per person with one or more symptoms, as reported by those individuals who had all positive symptoms treated by a physician. The resultin' total is an approximate measure of the number of physician calls which wo1 Id be required to treat all individuals reporting one or more positive symp­ toms on the Medical Needs Schedule, assuming that the untreated will need no more calls than the treated cases. (b) Multiply the quantity found in (3) by 2 to obtain the number of physician calls required to treat all individuals reporting positive symptoms for one year. (The Medical Needs Schedule obtains informatio on the symptoms of the sample population for a sixmonths period only.) 83 (5) Multiply the total found in (*f) by the recip­ rocal of the sample rate to obtain information on the medical needs of the total population of the community; i.e., if the sample rate is one-tenth, multiply by ten, etc. The resulting total is an approximate measure of the medical needs of the community, measured in terms of physician calls. B. Steps in Determining the Supply of General Medical Practitioners in the Community: (1) Determine the age of each physician (K.D.) practicing in the community. (2) For each of the figures in (1), obtain the deci­ mal fraction corresponding to each of the specific ages. (Table 111). The resulting figures are the approximate service capacities of the physicians in the community. Total these figures. (3) Multiply the total found in (2) by 7 8 8 0 to obtain the approximate total annual patient-call load which physicians in the community can carry. ( 7 8 8 0 is assumed to be the maximum yearly patient-call load which a physician at the peak of his career can carry, based upon an average weekly patient-call load of l6o for a period of 1+8 weeks.) By comparing the two totals obtained in A and B it will be apparent either that there are enough doctors in the com­ munity to meet the general medical needs of the people, or that additional general medical personnel is needed. This measurement is possible because both of the crucial factors-the need for physicians’ services and the adequacy of physician personnel to meet these needs — have been reduced to the com­ mon denominator of physician calls. Two illustrative examples which represent actual com­ munity situations in the state of Michigan, will help to clar­ ify the procedures involved: Measurement of Medical Needs and Resources in Pleasantville A. Measurement ifcft QggX Xgr General Medical Practitioners in the Community of Pleasantville. Village 0pen (1) ^5 persons in the village sample popuCountry lation (plus 31 in the open country) have had all positive symptoms, as reported on the Medical Needs Schedule, treated by a physician. SAMPLE PERSONS: b5 31 (2) 37 persons in the village sample popu­ lation (plus 8 0 in the open country) have reported positive symptoms, but have not seen a physician about all of these symptoms. SAMPLE PERSONS: 37 60 (3) 82 village persons (*+5 / 37) and 111 (31 / 8 0 ) open country persons have reported positive symptoms. The b5 village persons who have had all positive symptoms treated by a physician reported, for the preceding six-months period, an average of 3*2 doctor calls. By multiply­ ing 8 2 by 3*2 we have an approximate measure of the number of physician calls which would be required to treat all village persons who reported one or more positive symptoms on the Medical Needs Schedule. The comparable figures for the open country are 111 and 2.5, which make a total of 277-5 physician calls for the sample population in the oren country. PHYSICIAN CALIS IN SIX MONTHS: 82 x 3-2 = 26 2 26 2 111 x 2.5 = 277.5 277.5 (b ) Multiply the above totals by 2 to obtain the number of physician calls required to treat all individuals reporting positive symptoms for one year. (The Medical Needs Schedule obtained this information for a six-months period only). PHYSICIAN CALLS IN 12 MONTHS FOR SAMPLE POPULATION: 262 x 2 = 52^._ 52*f 277.5 x 2 = >?5 (5) Multiply the quantity found in (*+) by the reciprocal of the sample rate to obtain data for the entire population, e.g., if 85 Open Village Country the sample rate is i multiply by b. (If the sample rates in the village and open country areas are not the same, the totals in step (5 ) will have co be determined separately and the resulting figures added to deter­ mine the data for the entire commun­ ity). PIIYGICIAi: CALLS IA 12 MONTHS FOR TOTAL COMMUNITY: Village sample rate 1/5: 52b x 5 = 2 , 6 2 0 2 ,6 2 0 Open country sample rate 1/25: 555 x 2 5 m 1 3 ,8 75 13,875 In this community the sample rates for the village and open country areas were different, therefore, the totals for step (5) were determined separately. These totals were as follows: Village.......................... 2,620 Open Country TOTAL COMMUNITY PHYSICIAN CALLS FOR 12 MONTHS: 16,^95 B. Measurement of the Supply of General. .Medical. Practi­ tioners in Pleasantville. (1) The ages of the physicians in the community are as follows: bl, m-7 , 5 8 , and 8 6 . (2) The decimal fractions corresponding to these specific ages are .9 9 ) .8 9 , .6 3 , and . 0 7 res­ pectively. These fractions represent the ap­ proximate service capacities of the four physicians in the community. The total of these decimal fractions is 2.58. (3) The combined annual patient load capacity of the four physicians in Pleasantville, expressed in potential patient calls rer year, is 2 . 5 8 x 7 ) 6 8 0 or 18,91^. The data obtained by using the Medical Needs Schedule in­ dicate that at least 1 6 ,^ 9 5 patient calls will be required to treat all individuals in the community who require medical 86 examination and general care. The patient-carrying capacity of the four physicians in the community is approximately 18,91*+ patient calls, leaving a surplus of 2,*+19 physicians calls per year over and above the estimated need. This represents a slight excess of physician personnel in the community equivalent to .3 1 . (2 ,1+19 is 3 1 /J of the maximum annual patient-call capacity of 7,680.) This excess is equivalent to one physician of about 72 years of age. Hence it is apparent that the resident general practitioner medical personnel in the community are adequate to m^et the obvious needs for diagnosis and general care of the people living in Pleasantville. Comparison of the Operational Method with_ the Traditional Method of Measuring Medical Needs,and Resources This appraisal of need and resources in Pleasantville compares with the conventional approach of one medical doctor per 1 ,0 0 0 population as follows: The population of the Pleasantville medical service area is approximately 6 ,3 0 0 , thus, requiring a minimum of six physicians according to the traditional method of esti­ mating physician resources. The data obtained through the use of the operational method indicate, however, that the present physician personnel in the community (four physi­ cians aged *+1 , *+7 , ?8 , and 86 at the time of the survey) are adequate to meet the measured needs of the people. The three investigators who conducted the survey in this community all received the general impression, after 87 interviewing nearly 200 families, that the great majority of the familes considered the community adequately suprlied with doctors. This evidence, although admittedly of a subjective nature, tends to substantiate the findings ob­ tained by the Medical Needs Schedule and the operational method- namely, that the physician personnel in the community, as measured by the operational method, are adequate to meet the measured needs of the people. 88 Measurement of Medical Needs and Resources in Elmwood A. Measurement of the Need for General Medical Practitioners in the Community of Elmwood. Open Village Country (1) 37 persons in the village sample population (plus 22 in the open country) have had all positive symptoms, as reported on the Medical Needs Schedule, treated by a physician. SAMPLE PERSONS: 37 22 (2) 11 persons in the village sample population (plus 9 in the open country) have reported positive symptoms, but have not seen a physician about all of these symptoms. SAMi-LE PERSONS: 11 (3) ^ village persons (37+11) and 31 (22+9) open country persons have reported positive symptoms. The 59 persons who had had all positive symptoms treated by a physician reported, for the previous sixmonths period, an average of 3.6 doctor calls. By multiplying 59 by 3«6 we have an approximate mea­ sure of the number of physician calls which would be required to treat all persons who reported one or more positive symptoms on the Medical Needs Schedule. PHYSICIAN CALLS IN SIX MONTHS: **8 x 3.6 = 173 173 31 x 3.6 = 112 (V) Multiply the above totals by 2 to obtain the number of physician calls required to treat all indi­ viduals reporting positive symp­ toms for one year. (The Medical Needs Schedule obtained this in­ formation for a six-months period only). PHYSICIAN CALLS IN 12 MONTHS FOR SAM LE POPULATION: 3^ 173 x 2 = 3^6 112 x 2 = 22 ^f 9 112 89 Open ■ZillSLKS (5) Multiply the quantity found in (V) by the reciprocal of the sample rate to obtain data for the entire population, e.g., if the sample rate is i multiply by 2. (If the samule rates in the village and O’en country areas are not the same, the totals in step (5) will have to be determined separately and the resulting fig­ ures combined to determine the data for the entire community.) PHYSICIAN CALLS IN 12 MONTHS FOR TOTAL COMMUNITY: Village sample rate 1/27• 3*f6 x 27 * 9,3*+2 9,3^2 Open country samnle rate 1/50: C a u g tg y ________ 11«200 22V x 50 =11,200 TOTAL 20,5^2 TOTAL ESTIMATED PHYSICIAN CALLS FOR 12 KuNTHS:20,5V2 B. Measurement of the Sunrlv of General Medical Practitioners in Elmwood. (1) The ages of the physicians in the community are as follows: 30, Vo, Vl, 58, 58, and 57. (2) The decimal fractions corresponding to these specific ages are .68, 1.00, 1.00, .68, .68, and .66. These fractions represent the approximate service capaci­ ties of the six physicians in the community. The total of these decimal fractions is V.70. (3) The combined annual patient load capacity of the six physicians in Elmwood, expressed in potential patient calls per year, is V.70 x 7 , 6 8 0 or 36,096. The data obtained by using the Medical Needs Schedule indicate that a minimum of 20,5V2 patient calls will be re­ quired to treat all individuals in the community who require medical examination and general care. The patient-carrying capacity of the six physicians in the community is approxi­ mately 36,096 patient calls, leaving a substantial surplus of physician personnel over estimated need. 90 Comparison &£ t.frg 0per^tjLojnaJ. Mg.thod with t±e Traditional Method of Measuring Medical Needs and Resources This appraisal of need and resources in Elmwood com­ pares with the conventional approach of one medical doctor per 1,000 population as follows: The population of the Elmwood medical service area in 1950 was 8,000 (^,0C6 in the village and 3,99*+ in the con­ tiguous service area, thus requiring a minimum of eight physicians according to the traditional method of estimating physician resources. The data obtained through the use of the operational method indicate, however, that the present physician personnel in the community is more than adequate to meet the measured needs of the people. The physician potential in this community is unusually high because all of the practicing physicians in the community are at or near the optimal age at which the greatest service can be rendered in terms of service capacity. 91 CHAPTER IX VALIDITY OF THE OPERATIONAL METHOD This dissertation would be incomplete without c. con­ sideration of the margins of error involved in the three components of the operational wet hod, and an estimation of the over-all validity of the operational method based upon these marcine of error. Any adequate evaluation of the operational method would need to consider the folicwine Questions: First, what are the mar pins of error in each of the three components of the operational method (the natural area concept, the lnencurenont of me ’ical needs, and the measurencnt of physician resources), and how do they affect the validity of the over-all method? and second; What factors in the total medical situation in a community or medical service area ore .*ot measured by the operational n;e11:od , and 1.cw can they be ovaluated? Margins of Error Involved in the Operational Method A. Validity of the natural. Area Concept The natural area concept has I m p been a useful con­ ceptual tool for the sociologist, and yet few attenpts have been made tc test its validity. iron the o]o^aticnal point of view, which stresses anon,-; other things the necessity not only of clearly 1e fi ni :iq concepts end -r t'r 's , but also of critically tes tin.; thar, the do termination of the validity of eacli concept is essential. A 92 A second reason for determining the validity of the natural area as a conceptual tool is basically a statistical one. The use of statistical methods and techniques in the analysis of research data presupposes the definition of the area or statistical universe from which these data were ob­ tained, and in which they have validity. This is important, because any generalizations which are made concerning the relationships involved among the variables being studied (in this case health and medical care) are valid only for the universe from which the data were obtained. Those who employ the crude ratio of one doctor per thousand population as the criterion of medical service adequacy have defined none of the factors involved - what the area is, what constitutes a medical need, or what medical services are included. The results of the test of the natural area concept, which were reported in Chapter IV, sup.ort the author's hypothesis that the natural area concert as used in this research is not only a useful conceptual tool, but one which is also socially relevant and valid. Cf the 99 fami­ lies living within the boundaries of the Fleasantviile medical service area who reported that they had a doctor whom most members of the family employed, 85 families, (86$) consulted a physician laving within the area, ana the re­ maining lb- families (15/*) sought services in an adjacent area. It is irnrortant to know the degree of validity of the natural area concept in order to be able to make generalizatio V. / - /* i- ■■ ✓ / iJZLC- £ / kahsa s state college Deportment of Economics & Social: February 4, 1952 ALL STAFF MEMBERS: Subjoct: Dopartnont Senlnar Meatine3 Programs have beon orranged for the dopartnont seminar as foilowe: (a) Tuesday, February 5, 3 p.m. Mr. Russell Ives, Economist for the American Moat Institute, will discuss tho work of the Institute0 Mr. Ives is a graduate of Iowa State and hoadod the analysis unit of the war=-time Moat Board. Ho has also held othor positions in the USDA. His prosont position is the one formerly held by Bob Eggert who successfully led tho fight for decontrol of neat prices 0 Mr. Ives will be in Manhattan to speak on the Farm and Ilome Week programo We will moot in Grimo3 Memorial Seminar Roon0 (b) Thursday, February 21, 3 p.m. Mr* Ray Johnson, of the Kansas Industrial Development Commission, will discuss tho work of the KIDC. This will give many of us an opportunity to bocoma better acquainted with Mr. Johnson and the KIDC. Wo v/ill meet in the Grimes Memorial Soninar R^om. Anyone from othor departments who caros to attend will bo volcano. Sinceroly, William Barber, Secrotary Seminar Committee V/Bs jk 93 regarding health and medical care from data obtained within these areas. B. Margins of Brror Involved in the Measurement of Medical Needs 1. The Medical Needs Schedule Based upon a special validation study of the measur­ ing instrument employed to measure medical needs (which was reported in detail in Chapter VI), we know that the degree of correspondence between data obtained by the schedule and health findings based upon an actual physical examination, is 80 per cent. Of the 20 per cent which were not confirmed, we know that for every person improperly classified as having need for medical care, there were two persons improperly class­ ified as having rjo need. study, therefore, The net result of the validation showed that the Medical Needs Schedule under­ estimated ’’actual" need by approximately 8 per cent (assuming that the medical examination was a true indication of actual need ) . 2. Reliability of Na_ta Furnished by the housewife for Other Adults A detailed report of a second v a l i h t i c n study de­ signed to test the reliability of health data furnished by the housewife for other adults was also made in Chapter VI. These data were based cn a sample of 1 8 2 cases in which in­ formation was first given by the housewife for other adults, and later given by these same adults on themselves. The margin of error involved was an i pa:’ cent underestlnation of the needs of the other adults in ohe family for w: c r the :.cuse wife furnished symptoms data. 4 9^ 3« SeasoiqgJL Variation in Need For Medicaj Care The reader will recall that symptoms data were obtained from informants for a six months period only. This was done to minimize the factor of forgetting, and thus in­ crease the reliability of the data. In order to obtain data on a yearly basis one must multiply the measured needs for the six months period by two. on the assumption The above procedure is based that seasonal variation in medical needs would not be a serious factor. been conducted to test this hypothesis, thus the margin of error involved in this procedure, not ’-nown. U-. No experiment has as yet if any, is Estimation of the Number of Patient Calls Which Would Be Required to Treat All Individuals Having Positive Symptoms A fourth margin of error is to be found in the method which is used to determine the total number of pat­ ient calls which wrculd be required to treat all individuals in the sample population who reported one or more positive symptoms. This is done by using the average number of calls (3.2 in the Plcasantville community) which were actually required to treat those individuals v:ho reported that a l l their positive symptoms had been treated by a medical doctor, and multiplying the total number of people with either met or unmet need by this factor. The resulting figure represents the total number of calls which would be required to treat all those individuals who reported anv positive symptoms during the six months period. Involved in the above procedure, of course, is tho assumption that these individuals with any 96 established norm. Since the number of physicians in most rural communities is invariably small, it seems logical to assume that some relatively simple but accurate method of estimating the actual patient carrying capacity of these physicians is needed (as compared to Fennell*s averages) in order to make the operational method reliable and useful. No method of making this adjustment is now known. Other Factors in the Estimation of Medical Needs and Resources In addition to the margins of error in the operational method itself, at least two other factors should probably be considered in any community program designed to measure med­ ical needs and resources by the use of the operational method. First, the proportion of total need which is actu­ ally measured in a community is not known with any degree of certainty. gories: This unmeasured need falls into two main cate­ (l) preventive services such as vaccinations, in- noculaticns, etc., and (2) medical needs requiring the ser­ vices of specialists of various kin's (obstetricians, pedia­ tricians, etc.) which arc .not normally found in the type of community with which we are concerned. Secondly, one sheul ' bear in n i n 1 that the operational method is concerned with the? measurement of oris tin" nodical needs rather than with the effective de;and for medical ser­ vices. Effective demand for medical services in a given community is undoubtedly a function of many variables, as is the need, for these services, and the relationship which these two v a r i a b l e s bear to each ether in specific cormunity 97 situations has not as yet been determined. Annrdsal o^ the Over- all Validity of the Operational Ilet3od It is the opinion of the author (an opinion which is supported by that of a professional statistician whom the author consulted ) that an estime.tif. n of the over-all validi­ ty of the operational method with any decree of a ccur•cy at the present time is not possible. This conclusion is based upon the fact that each component of the operational method is subject to sampling and other types of errors, the varia­ tion of which in some instances is not known. The determina­ tion of the validity of the metJiod, therefore, is contingent upon the further refinement of those techniques and procedures herein indicated, and the determination of their respective margins of error. The formula for the determination of the estimated medi­ cal needs of a rural community by the use of the "symptoms" approach, at the present state of development, is as follows: 2 (Xp + Xo)fe- _____________ r . E X = medical needs X ■ number of persons with all positive 1 symptoms treated 7.2~ number of persons with -ositive symptoms not all treated X -3 _i- X, = number of calls per person in Xj group (= 3.2) E - adjustment for fact that Iledical Heeds Schedule does not measure all needs (preventive services, etc.) r = car.'le rate 98 Since the operational method for measuring medical needs and resources is still in the rrocess of refinement, and has been used in only a few communities, it is too early to determine its practical value with any degree of certainty. The practical value of the "symptoms'* approach to the study of medical needs, however, is demonstrated by the willingness of the Michigan State Medical Society and other cooperating groups to spend several thousands of dollars on the Michigan Health Survey and other related studies during the past few years. These studies have proved valuable as a basis for social planning in the field of health and medical care, and will probably be used even more extensively in the future. It is hoped that equally valuable conceptual tools will derive from the present dissertation. 99 CHAPTER X SUMMARY AND CONCLUSIONS The Need for .a Conceptual Framework for the Study of Social Problems The meaningfulness of social research is dependent in the last analysis upon the degree to which research projects are related to some theoretical or conceptual system. In the absence of such a conceptual framework, collected data constitute little more than a mass of unrelated and relative­ ly meaningless facts. For this reason, it becomes apparent that any science, particularly one which is relatively young and in the process of development, should concentrate upon the early development of useful and valid theoretical systems. Prior to the development of the so-called "value" approach by Richard C. Fuller and others, a conceptual system, in the light of which all social problems could be objectively anal­ yzed, had not been developed. Fuller had long been concerned over the lack of a uniform methodological system in the gen­ eral field of social problems, and set himself to the task of developing a systematic approach tc the study of social problems which would meet the following criteria: first, any approach in order to be useful as a conceptual tool must be applicable to all social problems, regardless of the soc­ ial context in which they might be found; second, an adequate approach should be socially relevant, and third: such an 1UU approach should permit the sociologist to perform the role of analyst rather than therapist. According to the "value" approach, social problems are defined as conditions which are a real or imagined deviation from some social norm cherished by a considerable number of persons. One should observe that there are three important factors involved in this definition: the concept of group norm, the concept of deviation from the norm, and the idea of group concensus. Applying this definition to the medical and health care situation as it exists in the United States at the present time, the following facts emerge: 1. A basic value or norm in the culture in which we live is that each individual is entitled to "good" health and "adequate" medical care. (group norm) 2. The inability of a sizeable portion of the popu­ lation to obtain "adequate" medical cvre because of low purchasing power, scarcity of medical per­ sonnel and facilities in many sec tic ns of the country, and other factors, leads to an increasing awareness on the part of the people that such a condition constitutes a definite threat to a soc­ ial value or norm which they cherish. (deviation from norm) 3. The fact that there is widespread concern over this condition is attested to by recent pei’sirtent attempts by various groups to initiate health leg­ islation at the national level, and the equally persistent attempts by other groups to prevent passage of such legislation. It should be noted that the conflict of values occurs, not in recog­ nizing that the problem exists, but In the establish­ ment of policies by means of which the condition may be alioviated. (gre ur cono ensus) In terms of the above definition, then, the present medical and health care situation in the United Ctctes is clearly a social problem in the cultural context in which it 4 101 exists. Aspect^ A. ”Value” Approach Objective and Subjective Aspects When viewed from the ^value” approach, social problems are seen to have both objective and subjective aspects. The objective aspect consists in demonstrating whether or not a given condition actually exists, and to vhat extent it exists. This can best be accomplished by means of scientific investigations carried out by skillful and impartial observ­ ers. The subjective aspect concerns the manner in which, social values, per se T enter into social problems - both in causing them, and in preventing their solution. The operational method for measuring medical needs and resources in rural communities is concerned almost exclusive­ ly with the objective aspects of the problem of health and medical care in the United States, e.g., the operational method has as its primary purpose the development of a method for measuring medical needs and resources which will more adequately determine the rroportion of 2'e° ■1° in a given groin- which actually hrs Insufficient medical care, and the number of physicians required to care for their needs. The present crude method of measurement, which arunes that one physician for everpr 1,000 people represents "adequate” medi­ cal coverage, may be misleading and inaccurate since it assumes that both the need for medical care, and the resources required tc meet the need, are constants. Considerable evidence is available, on the other hand, which indicates that both of I 102 these factors are not constants, but vary considerably according to several significant criteria. Concerning the subjective aspect of social problems, one of the significant aspects of the "value" approach in­ volves the conflict of values which characterize such problems. This conflict assumes dual roles: first, that of actually creating social problems because individuals in a given culture do not share common goals or objectives, and secondly, that of retarding or actually preventing the amelioration of these problems by a failure to agree on the means by which certain goals or objectives may be attained. Concerning the health and medical care situation in the United States as it exists today, it becomes apparent that the inability of a sizeable portion of the population to obtain "adequate" medical care because of low purchasing power, unavailability of physicians, etc., constitutes a definite threat to a considerable number of individuals and families in the society. This condition is brought about by conflicting values in our society which Under the achieve­ ment of the cultural goal of "good" health and "adeauate" medical care. The emphasis placed on individual freedom, for example, grants to each physician the prerogative of practicing in whatever area he may choose; the consequence being the relative concentration of physicians in urban areas where cultural advantages, ease of reaching patients, and other factors are more appealing to the average physician. A second value, that of financial success, likewise contributes A 103 to the maldistribution of physicians because the distribu­ tion of physicians has been shown to be much more highly correlated with the ability of people in a given area to purchase medical care rather than with the actual need for medical care as measured by the prevalence of positive symptoms. In summarizing this particular aspect of the problem, then, it can be stated that values actually cause social problems because they frequently run ot cross pur­ poses. Of equal importance with the causative factor just discussed is the manner in which conflict of value s actually retards or prevents tho amelioration of social problems. First of all, there is not universal agreement among the people either that a given condition exists, or that its presence constitutes a threat to the group's welfare, and secondly; (assuming agreement concerning the undesirability of a given condition); there exists widespread differences (conflict of values) concerning v:hat remedial measures should be adopted to ameliorate these conditions. With regard to the specific social problem with winch this dissertation is concerned, the introduction in 19^> of Senate Bill S-1606 (the so-called 'Nagner-hu ray-Singell Bill) into the United States Senate, became the catalyst which speeded the activities of both the proponents and opponents of this legislation. The loch of positive reme Hal action in the intervening years, following the Introduction of this bill "to provide for a Notional Health Program", is 10^ indicative of the widespread conflict of values which exist among the many groups whose representatives testified at the senate hearings for and against the bill. B. The Natural History of £ Social Froblem The natural history concept is based on the red.iza- tion that social problems do not arise full-blown but go through certain developmental stages including those of awareness, policy determination, and reform. Although the problem with which this dissertation is concerned is in the stage of policy determination, the principal emphasis, as previously stated, is on the objective aspects of the problem; namely, the development of an operational method of measuring both medical needs and resources in rural communities, with the objective of determining more accurate­ ly the extent of the need and the physician personnel reruired to meet the measured need. The nrtural history concept is included here solely for the purpose of complet­ ing the theoretical frarnevor'; within v.ddc1 the operational method itself is developed. The Theory of the Operational Method The methodological implications for tho social scien­ ces of a major theoretical advance in the physical sciences is recently becoming recognized. This significant develop­ ment is operaticnalism, which is based on the realization that symbols and concepts are amenable to manipulations which cannot be performed on the things being conceptualized. The operational point of view was first developed in physics, 105 and had as one of its major objectives the development of a method of conceptualization which would to a large extent eliminate the need for frequent drastic changes in the con­ ceptual framework of physics. This me tied was nothing mere than an insistance that concepts be made operational; that they be defined not in terms offnetaphysically conceived "properties1' but rather in terms of verifiable physical operations. A second methodological tenet which operation­ al! sn has stressed is probability. In other words, social concepts or social phenomena should be defined both in terms of the probability of occurrence and in terms of socially relevant, verifiable operations. An Goeraticnal hotho Measurinc medical '.Toeds and Resources in Rural Communities A. Thesis Objectives The general objective of this investigation has been to study the problem of health and medical care in the Tnited States within the framework of the "value" approach to the study of social problems which was outlined in an earlier chapter. The specific objectives have been the following: a) To outline the need for, and tho steps in the development of, an operational method for measur­ ing medical needs and resources in rural ccmunities. b) To define each of the tliree components of the operational method, and demonstrate hew each may be measured quantitatively. c) To illustrate tine procedures involved in the over-all method by applying it to actual community situations. 106 d) To further refine the techniques and measuring instruments employed by testing the basic hypo­ theses upon which the method is based. The need for the development of the op-rational method arises primarily from the increased awareness on the part of physicians and laymen that medical facilities and services in many rural areas are inadequate. Important factors in this increasing awareness have undoubtedly been three-fold: first, the widespread publicity which has been given to the draft statistics for borld bar II inductees has focused the attention of the public on the health problemf secondly, attempts on the part of various groups and agencies in recent years to enact some typ-e of health legislation on the national level have spurred both the proponents and opponents of such; legislation to unprecedented levels of activity; and thirdly, rural people themselves have tahen the initiative In attempting to improve conditions in rural areas, and the publicity and activity which have resulted from these efforts hove heightened the public awareness of the whole problem of health* and medical care. Thvre is, then, a growing av/areness that steps must be tahen to protect health, combat disease, and. provide adequate medical care for all who need it. the role of social research, froi the "value11 point of view, appears to be that of demonstrting whether or not such a condition actually exists, and to what extent it e: ists. It becccr.es necessary, therefore, to develop a method of measuring both* medical needs and resources; a method which is objective, 107 socially relevant, and subject to verification by other re­ search workers. The operational method was developed to perforin this function. Components of the Operational Method The central hypothesis which the operational method intends to test is that the present crude method of measuring medical service adequacy (assuming that one physician per one thousand population is sufficient) does not necessarily present a true picture of either health needs or physician resources, for the following reasons: 1. The area within which a given population resiles and obtains medical service is not defined by those who employ the crude method. 2. A simple counting of doctors may be highly mis­ leading because it is based on the assumption that all doctors have equal capacities for rendering medical services. 3. A simple counting of people may be very inaccurate because it assumes that the need for medical care is a constant, and does not vary between groups of people. In view of these factors, which those who employ the crude ratio of one doctor per thousand people ignore, it be­ comes apparent that any operational definition of the number of physicians needed given to meet thw needs of the people in a community must include the following basic components: first, a definition of the area witlln which medical services and facilities are most likely to be obtained; secondly, a met} od of measuring physician resources which takes into consideration the relevant variables involved (e.g., age); and lastly, a method of measuring medical needs which recognizes 4 108 that the need for medical c?.re varies with age, income, and other factors. Chapters IV, V, and VI are devoted to an analysis of these three components, and of the techniques employed to test the validity of each. The use of statistical techniques in analyzing re­ search data presupposes the definition ofthe statistical universe from which the data were obtained, and in which the techniques employed have validity. In order to measure medical needs and resources with reliability, it therefore becomes necessary to delineate and define the areas within which medical services may most logically be obtained. It was demonstrated in Chapter 17 that people do not necessarily obtain medical services within the township or artificial political unit, but within "natural :T,easM the boundaries of which are not fixed or permanent, but nevertheless can be delineated with considerable accuracy by giving proper recognition to such factors as the size of the medical service center, quantity and quality of medical services and facilities available in the center, nearness to other centers, relative drawTing power of adjacent centers, and other factors. Tho results of a test of the validity of the natural area concept as applied to medical services in vfnich 99 families were interviewed within a single area, sl owed that 85/j of tho families had sought medical services within the medical service area which had previously been delineated for them, and the remaining 15/: had gone to adjacent areas. (See 109 Figure ^ in Chapter IV). Of the 15/j who had gone outside the area, approximately one-third lived right on the border* It Is probable that with optimum accuracy of delineation, between 8 5 and 90 percent of the families in any given area would be found to have obtained medical services within the area which had been delineated for them. V/ith regard to physician resources, several studies have demonstrated that, in general, the capacity of a doctor to render medical services varies with the age of the doc­ tor. In order to measure medical services with a reasonable degree of accuracy, therefore, it becomes obvious that merely to count the number of doctors in a given area, with­ out regard to their service capacities, may lead one to false conclusions regarding the adequacy of medical services in a given area. Service equivalents have been worked out showing the relative capacities of physicians at different age levels. these data were based on all physicians in the United States, and reflect the cumulative effects of aging, retirement, and other variables, upon the capacity of physicians to render service. The use of these service equivalents makes possible a more accurate determination of rhysician resources than that of merely countin : doctors, and assuming that all doctors have equal capacities for rendering such services. The third component, that of measuring the need for medical care, is based on the assumption that merely counting reorle and assuming ‘ :hat one group has the same amount of 110 need as another group; is as fallacious as merely counting doctors and assuming that all doctors have the same capacity to render service. In other words, the need for medical care is not a constant, but varies according to age, income, and possibly other factors. The instrument by which medical needs are measured is the Medical Meeds Schedule, which is a list of 27 medical symptoms, any one of which indicates a need for medical attention in the opinion of the panel of physicians were consulted in its development. who Meed, therefore, is de­ fined operationally, as the presence of one or more positive symptoms for a given individual. C. Validity of the Operational Method A detailed consideration of the margins of error in­ volved in each component of the operational method, and of the resultant validity of the over-all method was made in Chapter IX. It will suffice here to reiterate that at the present time it is not possible to determine' the over-all validity. This conclusion is based upon the fact that each component of the operational method is subject to sampling and other types of errors, the variation of which in some instances is not known. The determination of the valiiity of the method is therefore contingent ut on the further re­ finement of those techniques and procedures herein indicated, and the determination of their respective margins of error. D. Comparison of the Operational Method of Measuring Medical Needs and Resources with the Traditional Method Ill After the development of the operational method, and the testing of the basic hypotheses and concepts employed, the method was employed to actually measure the medical needs of two rural communities in Michigan, as well as the physician resources which would be required to meet the measured need. The findings in both instances varied con­ siderably with the traditional met1od of appraising these two factors. A summary of these two community studies fol­ lows : COMMUNITY OF FLEASANTVILIE A. Operational Metfagd 1) Petermination of Medical heeds - The measured medi­ cal needs of the community, in terms of the patient calls which would be re uired to meet these needs, was estimated to he l6,*+95. 2) deterr:1nation of Physician Fersonnel Heeded - The patient-carrying capacity of the four physicians in Fleasantville (whose ages are ^1,^7,5c, and 86) have a combined patient loa.d capacity of approxi­ mately le,glh. 3. Traditional Method The population of the medical service area of Fleasantville is 6,300, thus requiring a minimum of six physicians according to the traditional method of estimating physician resources required in a given area. COM--UNITY OF ELI MOOD A. Operational Method 1) Determina 11on of Medic al Needs - The me asur ed medical needs of the community, in terms of pat­ ient calls, was est' ms ted to be 20, yh-2• 2) Deter minati on of Fhysician Personnel Needed - The patient-carry.ing capacity of the six physicians in E1r.iwec.d is 36 ,096-. 112 B. Traditional Bethod This appraisal of needs and resources 1n Elmwood compares with the conventional j.rjroac! of one medi­ cal doctor per 1,000 people as follows: the popula­ tion of the community is F,0C0, thus requiring a min­ imum of eight physicians according to traditional methods of estiir ting person: el requirements. Conclusion Although additional refinements in the egc etional method are indicated before its validity can he determined with accuracy, many studies have already been completed which have used tie "symptoms" approach :: id other concepts and tecl niques n.sed in the operational met’od, ir.clu 'ing the state-wide iic". igaw health Survey recently completed by the Gocial Research Service in coopc ation with the hichigan St ate Bedica1 3cciet;* arrd otlie:• agencies. Tho r e have proved valuable as a basis for more intelligent and scien­ tific social planmhng in ti e fiel1 of he s'"td and medical care, and it is hoped that ether contributions, both theoretical aid practical, will ensue from- this research. 113 BIBLIOGRAPHY (Books) Bridgman, P.W., The Logic of Modern PhysicsT Macmillan (1932). Cuber, John F., and Harper, Robert A., Problems of American Society: Values in Conflict. Nev.1 York: Henry Holt &■ Company, 19®+8. Dodd, Stuart C., Dimensions of Society,. IIew York: The Mac­ millan Co mpany, 191*-!• Lundberg, G.A., Foundations of Sociology. New York: Macmillan Company, 1939. The Snedecor, G.W., Statistical Methods. Fourth Uj.itien; Ames, Iov/a: Collegiate Fress, Inc., 19*+£ . (Bulletins) Anderson, Elin L. , "The E:-tension Service 1s Restonsibility in Aiding Rural People to Improve Their Health and Medi­ cal Services", Washington D.C.: Unite1 Sta tes Deportment of A ;riculture Intension Service Bulletin, (July, 1 'Ml). Hoffe r. Charles R ., et al, "Health and Health Care in Michio East Lancing, Michigan: Michigan State College, o ,n% Agricultural Ik pcrinent Station S; ccial "'ulletin 352) (Sept., 1920). ;*+7, the first official meeting between representatives of the Michigan State hedical Society (the Health Survey Ad­ visory Committee) and the Social Research Service of Michigan State Colle ge war. hold for the purpose of U scussin g the scope, methods, and objectives of the project and planning the initial phases of the survey. By Juno, 19*+?, the necessary preliminary A 118 arrangements had been completed, and the field work was begun. The general objectives of the Michigan Health Survey, as stated in the official contract between the Michigan State Medical Society and the Social Research Service include the following;2 1. To determine the extent of medical needs of a repres­ entative sample of families in the state of Michigan (exclusive of Wayne County), the coverage of the sam­ ple to be a representative cross section of the state (exclusive of Wayne County) as agreed upon by the Social Research Service of Michigan State College and the Health Survey Advisory Committee of the Michigan State Medical Society. 2. To determine the extent of medical needs among a representative sample in one or more com. unities of Michigan to be agreed upon as above. 3. To obtain data on attitudes, opiruons, and practices of members of the sample population pertaining to health and health care as related to objectives de­ termined jointly by the Health Survey Advisory Com­ mittee of the Michigan State Medical Society and the Social Research Service of licliigan State College. b. To prepare, or have prepared, such materials (marked maps, aerial photographs, etc.) as may be necessary to yield efficient and economical samples of house­ holds of Michigan outside Wayne County. 5. Estimate the population of the various trade center areas of Michigan and determine the ratio of popula­ tion per physician (M.D.'s and Osteopaths) in each one. These general objectives clearly indicate the broad scope of the survey. As the project developed, however, many spec­ ific phases of the problem were delineated for more detailed analysis, and members of the Social Research Service staff jgreed to undertake the execution of Lhese special studies. 2. Quoted from ""Exhibit A” which was approved at a jo meeting of the above groups on February 25, 19*+8. 119 The following specific phases of the over-all problem of health and medical care were investigated: 1* A study of the medical needs and experiences of 1,113 families. 2. A study of opinions and attitudes regarding health services and health care issues of 717 adults re­ presentative of those in the 1 , 1 1 3 families. 3. An ecological study of the state involving the delin­ eation of 30*+ local medical service (general practi­ tioner) areas. *+• A demographic study yielding current population esti­ mates for each of the 3 0 *+ medical service areas. 5. A study of the distribution of general medical per­ sonnel in relation to the medical service areas. 6. An array of these medical service areas in terms of physician-population ratios. 7. Intensive studies in two communities which represented the two extremes in terms of adequacy of general medical personnel. The first major task confronting the planners of the Ilichigan Health Survey was the determination of broad ob­ jectives and the selection of specific areas in which factual data was desired. This task was accomplished through the splendid cooperation between the Health Survey Advisory Com­ mittee and the Social Research Service in joint meetings be­ tween the two groups; the doctors for the most part suggesting the areas or questions which they felt should be explored in­ tensively, and the personnel of the Social Research Service translating these questions into workable form. 'The skill and insight of both groups, and the resulting synthesis of the two specialized approaches to the problems at hand, con­ tributed greatly to the success of the survey. i 120 The second ma^or task was that of developing an adequate instrument for the systematic collection and classification of the data. This, of course, was a function of the Social Research Service. The interview schedule or questionnaire which was used for the collection of the data war divided into three parts. Part I was concerned primarily with the measurement of medical needs, and the measuring instrument used was the Medical Needs Schedule which was developed originally by Dr. Edgar A. Schuler and others (described in detail in chapter 1) and later validated by Dr. Charles R. Hoffer and others. The function of part II was to pro­ vide information re la ting tc the attitudes, opi ricr. s, and practices of the people regarding many phr; ses cf medical care including their own periences with doctors, group insurance plans, non-medical practitioners, community health problems, and other related aspects of medical care. Fart III consisted of numerous control items which would p-.-rcit intensive analysis of the data in terms of such significant factors as income, educational level, residence, and other relevant variables. Data relating to the ecologies! ~ rpects cf the investigation (e.g., the delincatio : c f m e dieel servico areas, the denegranLie study involving peculation rot'rates for the medical areas, the Ustributi n cf general medical jerscnnel in relation tc medical service areas in terms of physicianpopuls ties ratios; u:r: not obtained by "slug the schedule, but were go.there :• and c-gauisod by Dr. J. F. Tha '.on from *1 1 122 Table XII* Distribution of households and individuals in the sample for the Michigan Health Survey, by residence, Michigan 19^8. Sample Area ho. of Households ho. of Individuals Ho. of Adults Providing Ans wers About Practices and Opinions Open Country 369 1,317 2^-5 Village 12>b *+21 7^ hetro poll tan 153 5*+8 102 Urban ^57 1,500 296 1,113 3,786 TOTAL 717 The Michigan Health turvey is thus of vital importance in the present research project because it provided the data which made possible not only the extensive testing of basic hypotheses and concepts, but also an intensive analysis of the variables affecting the need for medical care. 123 METHODS AID FROCEDURES EHFLOYED IN THE HICHIGAH MEALTH SURVEY The objectives of the Michigan Health Survey were stated in Appendix A, and the manner in which the sample of 1,113 households was selected for this state-wide study will be explained in the following section (Appendix C). The present section is concerned with the over-all methods and procedures employed in the actual planning phases of the survey and in the field, work, together with an explanation of specific techniques which were employed. These methods and procedures can most logically be divided into five distinct categories, as follows: 1. The development of a schedule or questionnaire 2. Training field workers 3. The carrying out of the field work-collection of the data b. Editing, coding, and tabulating the data 5. Analysis and interpretation of the data As previously stated, Appendix A outlined the scope and objectives of the Michigan Health Survey, and the re­ spective roles cf the Social Research Service and the Survey Advisory Committee of the Michigan State Medical Society, In addition to its role of co-planner of the survey, the So­ cial Research Service was responsible for the carrying out of the actual research work in all of its phases. 12*f The Development of & Schedule or Questionnaire1 Once the broad objectives of the Michigan Health Sur­ vey had been outlined, and the specific areas in which in­ formation was desired spelled out in greater detail, it was then possible to proceed with the task of developing an adequate and reliable questionnaire for the systematic coll­ ection of the data. Since two basic types of data were de­ sired, the questionnaire was constructed accordingly, and a third section added for control items. The first section deals exclusively with the measurement of medical needs by the use of the Medical Needs Schedule. This measuring in­ strument, together with a report cf the validation study by means of which it was tested, has previously been described in detail in Chapter VI. Separate vertical columns are provided in the Medical Needs Schedule for each individual in the family, together with information about his age, sex, and certain etler per­ tinent facts. Information is obtained on the presence or absence of each of the twenty-seven symptoms for each in­ dividual, together with what has been done about each posi­ tive symptom reported; i.e., whether no treatment at all has been obtained, home remedies only, advice or aid by non­ medical personnel (osteopaths, chiropractors, or other persons engaged in the art cf treating illness) and lastly, whether the advice and care cf a medical doctor was sought. As stated previously, it is assumed that "no treatment" represent: 1A co; y of the complete quo stionnr.ir e is reproduced in Appendix D. r 125 the least adequate care, and "treatment by an K.D." the most adequate treatment with regard to the positive symptom re­ ported. Information was also obtained on vaccinations and immunizations, number of days which each individual had been kept from his regular activities due to Illness, and other factors relating to medical need. The information in this part of the questionnaire was of vital importance to the survey, because it was this data on medical needs which made possible the testing and refinement of many basic hypo­ theses which had previously not been possible because of an inadequate saiple, as well as a study of the factors affect­ ing need for medical care. (Chapter VII) The second section of the questionnaire was designed to obtain information on tire attitudes and opinions of a representative sample of adults regarding many aspects of medical care, as well as practices in regard to the use of medical facilities and services, etc. With the exception of the questions in this section which relate tc the use of medical and hospital facilities and services, which were used in Chapter IV (The hatural Area Concept), these data were not used in this thesis. Section three of the questionnaire consisted of many control items which, it was felt, would be helpful in analys­ ing and interpreting the data on medical needs. Since one of the basic hypotheses was that the need for medical care varies significantly according to age, Income* and other variables, these centre! items are necessary in testing this 126 hypothesis, as well as others. After the completion cf the initial draft of the questionnaire, trained interviewers visited several hones in the Lansing area for the purpose of testing the instrument, especially the wording and uniformity of interpretation of the questions in Fart II having to do with attitudes, opin­ ions, and practices of adults concerning such controversial subjects as Insurance plans for paying hospital and medical doctor expenses, the use of non-medical/personnel such as osteopaths and chiropractors, recent experiences with doctors and hospi­ tals, etc. After each such testing procedure, the inter­ viewers met with personnel of the Social Research Service for the purpose of revi sing and improving any cc ntroversial or ambiguous questions. points as many as 13 Only after many revisions, ( on some different versions we^c used) covering a period of several weeks, di^. the quc stionnaire meet the approval of all personnel concerned. Copies of the question­ naire were then prepared for the state-wide liichigan ;>slth Survey. Cne further point shoul’ be mentioned in connection with the levelopment of the que stionnr.ire . Because the tabulation of thx data at the completion of the survey would :!nv'lve ■ no sel'r 1r,?00 se ear ■; c ?r ’s Ca.be ’it 1 ,2CC famul” cam's and 2 , 6 0 0 individual carls), the use of IBk ecyir ment was "lanned fro:", the very ' m g i m i ng of the survey. Thi s y o c e ’urc entailed a few extra hours of labor in setting ra tlx inter vie schedule in the proper manner, but the ultimate saving in time 127 money, and labor is incalculable. To set the questionnaire up in this manner requires a thorough knowledge cf the IBM (Hollerith) card and the equipment to be used, which of course must be anticipated before the actual setting up of the cuestionnaire can be accomplished. Regardless of the skill with which a schedule or ques­ tionnaire is constructed, or the representativeness of the sample, the value of the data Spends to a large extent upon the energy, skill, tact, patience, per reverence, and Integrity of the field v:orkers who collect these data. It is of the utmost importance that the field workers be thoroughly famil­ iar with the objectives of the survey, the techniques of in­ terviewing , and the instrument which they will be using to gather the data. be reliable. Above all, however, the field workers must The importance of making every reasonable effort to contact all families in the sample, of interview­ ing only those families included in the s ample households, and integrity in recording all information accurately and systematically, all contribute immeasurably to the successful completion of any survey. The field workers were, with one exception, graduate students in the Department of Sociology and Anthropology at Michigan State College.2 The write- •s wife, Candis h. hod pdou. he field work was "ndor the direct seporvision of Dr. Charles h. Hoffer, aid conducted by hay ilakely, lean book, fal' 3oe k, C1 arenc e J ane , SlieIdon owr y, C andis h o o n, ~nd Linvo o■ ho 'gdcn. although not enrolled in the graduate school, had had con­ siderable interviewing experience with the Medical Meeds Schedule in the three Michigan communities of Me sick, Stephenson, and Concord, in connection v/ith health surveys conducted in those communities by the Social Research Service and the Michigan Department of Public Instruction. All interviewers were thoroughly indoctrinated in the use of the questionnaire, including the Medical Needs Schedule, prior to the time the state-wide survey got under way. This indoctrination included frequent conferences at which details and procedures were explained and clarified, several "practice" interviews ve?-e conducted in the Lansing area by each member of the field staff, and several manuals were prepared which related to all aspects of the survey. These manuals were pre­ pared for the use of personnel in the field, and carried de­ tailed instructions concerning all aspects of sampling procedures, reporting symptoms, etc. and were designed to insure greater uniformity in field work procedures. These manuals included the followingt 1. Operational and Sampling Instructions for the Field Work (I'm - pp.) 2. Special Sampling Instructions for Cities and Towns in Which Sampling Units must 3e Created and The Segments Selected in the Field (7 Pr.) 3. Manual For Interviewers; tenoral Instructions (8 rp.) b. Instructions for Reporting Symptoms ( U- pp.) Field Work - Collection o£ Data The field work was carried cut for the most part by teams of trained interviewers working in groups of two throughout 129 various sections of the state, although toward the end of the survey interviewers worked singly to complete the collection of data in sample areas not previously finished. The mechan­ ics of the field work procedure can rest conveniently be explained in terns of the steps involved, as follows: 1. Locating the sample areas The first step in the field work procedure was locating the sample area in wrich the interviews were to he taken. This task was greatly facilitated, as stated in Appendix C, by placing all the maps, aerial photographs, tracings, etc. pertaining to the sample areas within a given county in separate man!11aenvelopes. These envelopes were arranged alphabetically by counties, and were placed in a file where they were easily accessnble tc the field workers. -‘-'hose envelopes each contained a county highway map shoving the location of each of the sample areas within the county2, as well as more detailed naps, aerial photographs, etc. showing in greater detail the boundaries of each cf these areas, the identification number of the segment, the- sample rate, and other relevant data. The statistical laboratory which developed the sarrmle had also marked the boundaries of each sample segment in red. It was relatively easy, therefore not only to locate tie saw 1c areas but also to determine ^Figure 12, the county wag of huskegon County, shows the locaticn'of all sa: 'le segments in the county, as well as all four tyres of samrle areas, i.e., rural (fll), village (yfu), metropolitan (r/:3h), and urban CrH-u and 5u). fore detailed maps of these aro-s sup.losrnt' •• the county kigivay map. n 3 3 £ ♦ g e n e r a l h ig h w a y MUSKEGON map COUNTY F 131 their exact boundaries. A typical county map showing the location of all sample segments is shown in Figure 12. In cases where cruising was necessary to determine the sample segments, the cruising operation was usually carried out by a team of workers, and the interviewing done at a later date, thus allowing ample time for either the field workers or office personnel of the Social Research Service tc determine the sample segments which were to be used. This procedure, it was discovered, could not be readily accomplished in the field because of the many computations involved. The preparation of these county folders permitted a more efficient and econc aical planning of the field trips by the 3nterview teams, and eliminated the possiblity of dup­ lication of field work in the same area by different teams of interviewers. 2. Selection of households within the sample areas For the oren country segments, which were 4ncluded. in the state-wide sample, a complete enumeration of all house­ holds was made. For the other three types of segments, how­ ever, a sampling rate had usually been assigned which meant that only the designated fraction cf the total number of households war interviewed. In all sample areas where samp ling rat no had boon assign­ ed, the Trebles: cf determining which households to interview becomes of great importance, and here, as in so many areas of research, the main problem is that of eliminating biases as much as mos cible. The estabxish.'.Gnt o f a 'niform. ^ro-ce’isn c Id 13 the wife was not available, this information was accepted from the husband. At least three attempts were made to contact each family, usually at different times of day. In some instances individ­ uals were interviewed at their places of business, information on this point having been given by neighbors. In a few in­ stances, when families were visiting in other cities or were temporarily out of town, this information was given to in­ terviewers in the respective part of the state and the nec­ essary information obtained by them. Information in Fart II of the questionnaire, relating to attitudes, opinions, and practices of adults with regard to many different aspects of health and. medical care, was not obtained from every household but from a representative sample of approximately one-third of all adults in the sample. A separate table of random starts was used sc that a truly representative cross section of adult opinion could be ob­ tained. Data from Part II was not used in this dissertation, however, except that which related to the use of medical and hospital facilities, which war. presented in Chapter IV, THE NATURAL Ah TA CONCEPT. Editing. Coding. and Tabulating ii*e Data It was planned from the v^ry beginning of the Michigan Health Survey that all of the data obtained therefrom wo ".Id be transferred to 131. cards for tabulation. The savings in time and money, as well as In greater accuracy, which the use I3h of IBM equipment make possible are so great as to be well nigh incalculable. Before the final step of tabulation is possible, hov/ever, other important steps need to be consider­ ed. One may observe, after looking at the Michigan Health Survey questionnaire reproduced in Appendix D, that the questionnaire is constructed in such a manner as to greatly simplify the editing and coding procedures. At the comple­ tion of each day of interviewing, while the interview was still fresh in the mind of the field workers, each question­ naire was carefully checked to see that all answers were properly recorded and in general to see that ever3rthing con­ cerning the interview was in order. These completed inter­ views were then turned in at the Social Research Office where trained personnel performed the editing and coding, and later transferred the coded information to the IBM cards. After these procedures had been completed a ,fspot check” of the cards against the original interview sheet was conducted to ascertain the degree of error in this phase of the study. The margin of error found was approximately one tenth of one per cent, indicating that the error in this phase of the sur­ vey war so small as to be of no consequence. In order to further standardize the coding procedure and bring about uniformity in interpretation, coding instruc­ tions were prepared for each of the three parts of the questic mire. 135 AaalXSlS ££& Inter pretatlp.n o£ tj^e Pa.kfl Fewer steps in social research require greater skill and integrity on the part of the investigator than those in­ volving the analysis and interpretation of collected data. First of all, an adequate under standing of the theory and methods of statistics is a necessity, for without this mini­ mal background the proper analysis of data is im- o. sible, and the resulting interpretations not onl:r unconvincing but unreliable. Secondly, an undersLanding of the theory and methods of sampling are equally important, for without a represent;-- tive arid unbiased sarrle no valid comli’sicns can be arrived at. 1’hirdly, e::tren:& caution sho^l:'- be exercised in all phases of a research project to insure tha - facts are gathered to tort hypotheses rather than to substantiate them. A -rlori conclusions are all too frequent In research and of course contribute nothing but confusion a-.id. uisonde standing of social problems rather than an underafarding of them. Fourthly, and of fundamental ir.-port-r.nee, is the necessity of relating factual data, obtained by ccm.rot/ nt r ;•s<--arck workers to some theoretical or conceptual system. 'hithou'; a ccncoptua framework social research remains at best little wore than- a m *-i o' c (y t"* i ~)y p* 1° '1 0 fh'** Ill'll *- C'ii'lOH"7' 1- IT. — r ’11# AFFEKDIX THE SAMPLE FGR THE MICHIGAN HEALTH SURVEY Introduction The sample which was used for the Michigan Health Survey was a stratified random sample, on an area basis, of the population of the state exclusive cf Vv'ayne County. It was prepared by the Iowa State College statistical labora­ tory. Because one of the major objectives of the survey was to study the medical needs of rural families, and to compare the needs of rural people with the needs of people living in urban and other types of areas, the Health Survey Advisory Committee of the Michigan State Medical Society and the Social Research Service of Michigan State College decided to exclude Wayne County from the sample. Detroit (in Wayne County) and its environs include approximately M+ per cent of the population of the state. If Detroit were included, the resulting sample in rural areas would be too small for the desired types of statistical analysis. Thus the sample is a representative cross section of the households of the state of Michigan outside of Wayne county. The Area Method of Sampling The households were selected by a technique called the "area method" of sampling. Although not a new technique, the "area method" has only recently been developed and used extensively in this country. This method has been 137 developed jointly by the Bureau of the Census, the Bureau of Agricultural Economics, and the Statistical Laboratory of Iowa State College.1 The demand for accurate and timely information during the war intensified the need for, and hastened the development of, better sampling methods which could more adequately meet the many needs of these agencies. It was also felt that the "area method" would permit a greater degree of integration and thereby increase the effectiveness of the surveys carried out by these agencies. Experimentation with the "area method" led to the development of the so-called Master Sample of Agriculture. The Master Sample was originally designed to enable the accumulation of data relating to a representative group of 5,000 farms and farm families, but was later expanded to include 300,000 farms as well as a Master Sample of City Areas. These two projects were finally designed as integrated parts of an over-all sampling plan for the United States. Although it is beyond the scope of this paper to dis­ cuss in great detail the methods by which the Master Sample of Agriculture was obtained, it may be of interest to describe briefly the theoretical considerations under­ lying the "area method" of sampling, and to describe the -*-For a~detailed description of the "area method" and the Mas­ ter Sample of Agriculture see Jessen, R.J., "The Master Sample Froject and its Use in Agricultural Economics", Jour­ nal of Farm Economics. Vol. XXIX, No. 2, (May) 19*+7 and King, A.J., and Jessen, R.J., "The Master Sample of Agriculture," Journal of the American Statistical Association. (March, 19^0) Vol. *+0, pp. 38-56. 1 138 general plan of the present sample and some of the steps taken to carry it out. The Master Sample is a sample of small areas which average about 2.5 square miles in size, but these areas vary in size according to location and other factors. of these sample areas varies from 108 The size square miles in Nevada to 0.71 square miles in Indiana. The sample areas were selected from every one of the 3,070 counties in the United States and contain within their boundaries 1/lS of the land area of the United States, 1/lS of the rural popu­ lation, etc. The three primary strata were defined on the basis of incorporation and density of population. The total area of the United States was classified into: 1) incorp­ orated areas, 2) unincorporated areas relatively densely populated, and 3) unincorporated areas relatively sparsely populated. For convenience these area groups are called 1) incorporated, 2) unincorrorated, and 3) open country. In the type of sampling with which we are concerned the sampling units are small areas and every unit of ob­ servation is uniquely associated with one and only one such area. Furthermore, the total number of such area sampling units in the population is presumably known so that the proportion sampled (i.e., the sampling rate) is known. An unbiased selection of persons can thus be ob­ tained by taking all those who reside within the sample area. 139 Area sampling may also be used to provide the primary selection of units of observation which may then be done by subsampling of smaller areas or by making a prelist of individuals from which a small and possibly stratified sample may be drawn. At the present time the Master Sample consists mainly of materials. These materials are cf two kinds; 1) 3,070 county maps (one for each county in the United States) on which sample areas have been designated, providing, when enumerated, a sample of about three hundred thousand farms, and 2) materials useful for drawing samples so that the present sample- can be extended or supplemented for any sar.r le of agriculture or of persons. The Master Sample materials consist of large scale highway maps cf all counties of the United Stages; listing cf all towns, villages, cities and ether dense!:* pcaul-tod areas having an estimated population of 10C persons or mere; aerial rkoto prep hs for short f1 p'r rc.n c Lie 67,000 same le areas cf tic Mas her 2c:.r J.e; ma serials by which count-* reps shoving sample areas for a three hundred thousand farm sample can be reproduced, and tabulation cf fhe. cote ’ nnr’'.(pa of -’.veilings by small geographic ar-.ac and their curul tivc totals. These Master Sample materials, which are controlled by the Bureau of the Census, the Bureau of Agricultural Economics, and the Statistical Laboratory of Iowa State College, greatly facilitate the design and ere para tie.. cf 1*4-0 specific samples. The stratified random sample used in the Michigan Health Survey was prepared from the Master Sample materials, and adapted to the special needs of the Michigan Health Survey, by the Iowa Statistical Laboratory. Advantages of jthe Area Method si Sampling The area method of sampling has certain advantages over other types of sampling procedures, the most significant of which are the followings 1. The area method of sampling is readilyadapatable to such modern techniques as stratification, over sampling and subsampling for reducing the sampling errors. 2. The area method makes possible the use of expansion factors. An expansion factor applicable to the area method is known without error because the total number of sampling areas in the population may be determined by simply counting them on the maps. Hence, it is not necessary to use outside sources of information which may be inaccurate in order to make estimates of totals. 3. Shifts in population are readily measured and taken into account by the area method, thereby resulting In greater accuracy. The advantages of the area method of sampling make it highly desirable for a state-wide sample such as that employed in the Michigan Health Survey. A stratified type of sample, for example, makes possible an extensive analysis of medical needs by place of residence, thereby greatly increasing the utility of the sample. Moreover, when the expansion factors are used it is possible to obtain reasonably accurate estimates of the medical needs of all of the people throughout the state. 1*+! The significance of the data obtained by the use of these expansion factors for social planning in the field of health and medical care can readily be appreciated* Since the area method of sampling takes into account shifts in population, the accuracy of the resulting sample is increased considerably. In the Michigan Health Survey sample, for example, two "zero segments'’^ were encountered, and in several of the other segments or sample areas signifi­ cant increases in population had taken place during the past few years. Proper allowances for these shifts in population can very easily be made, resulting in a proportionately smaller sample from those areas which had lost population, and in a proportionately larger sample for those areas which had experienced sizeable gains in population. The expansion factor for any item in part I, the symptom or medical needs section of the schedule, is 977. For items in part II, the section on attitudes and opinions regarding medical practices, experiences, etc., the expansion factor is 3 (977) or 2931. One will recall that for part II of the schedule a random sample of only 1/3 of the adult members of the households was interviewed, compared to the complete enumeration which was obtained for part I. These expansion factors can be used, however, only for the three-fold ^A~*zer"o segment” is a sample area in which there are no households. Ib2 classification of areas as developed by the Iov;a State College statistical laboratory, and not for the four-fold breakdown. Because of this, if the three-fold breakdown is desired, it is necessary to combine the data for the ’’Villages11 and "Metro­ politan areas”, as explained previously, Tfog. Sample £pr the Michigan Health Survey The households for the Michigan Health Survey were drawn from three basic strata which are defined as follows: 1) Urban Zone-includes all incorporated places having 2,^00 or mere people according to the 19^+0 popula­ tion census. 2) "ural-r.lr.ee Zone-includes all incorporated places having less than 2,500 people, and all urine or t or a ted places. 3) Open Country Zone-consists of all areas outside Urban Zones and Rural-place Zones. The unincorporated ar eas cc usist of trr typo, s : a) name places, outside of incorporated places, v/ith 100 or more persons in 19*1-0, and b) 11 other ar e a s which have a ;ovula­ tion of 100 or more persons per square mile. The Urban Zones are equivalent to the urban areas defined by the Census bureau. The Rural-place Zones and Cron Country Zones to­ gether are equivalent to the Census' Rural Places. Although the three-fol-' classification is usually used for most samples drawn from the Master Sample Materials, the sample is designed in such a way as to permit a four-fold breakdown if desired. This is achieved by dividing the Rural- place Zones into two groups, as follows: a) all incorporated l*+3 places having 2,500 people or less, and b) all unincorporated places having 100 or r.ore people per square mile which are adjacent to urban centers of 50,000 or r.ore population. The former are designated as Villages and the latter as Heiropolitan Areas. Using the four-fold system of classification, the four types of sampling areas then become as follows: Urban, Metropolitan, Village, and Rural. Because one of the major objectives of the Michigan Health Survey was to measure the extent of medical needs by major types of areas, the four-fold classification was felt to be more desirable. The percentage distribution of households and individuals in the sample for the Michigan Health Survey, by place of residence, is shown in Figure 13. Distribution of the 1,113 sample households throughout the state, by counties, is shown in Figure 1^-. The sampling materials for each county which was in­ cluded in the sample for the Michigan Health Survey v;ere prepared and placed in separate folders by the statistical laboratory at Iowa State College. These materials consisted of county highway maps, aerial photographs, and city and township maps. The location of each sample segment was -clotted on these maps, together with information indicating the type of sample segnent and the sampling rate. In all of the rural sample areas each household was interviewed, but in most of the other sample areas samp ding rates were used, the rate being dependent upon the number of households in the 11^ RURAL VI L L A G E METRO­ POLITAN U RBAN Tor "•' r.r It f .:pl*n ip-v— , -^1 lk-6 particular sample segment. Special Sampling Instructions for Towns and Cities in Which the Sampling Units Must be Created and the Segments Selected in t M Fjeiq For certain of the cities and towns which were selected for the Michigan Health Survey sample no estimates of the distribution of households in the various blocks were avail­ able. It would have been possible to consider all blocks as having an equal number of households, and to have per­ formed the sampling accordingly. This would have been un­ desirable, however, because of the great variability in the actual number of households in the segments. It is much more efficient statistically to control this size of the sampling units. This is achieved by an operation known as "cruising", which consists of the following steps; ordering of blocks, cruising to obtain eye estimates of the number of households in each block, assignment of sampling units to blocks, and selection of the blocks to be used in the sample. This pro­ cedure will be explained and illustrated in the following paragraphs• a) Ordering of Blocks The blocks shewn on the nap (Figure 15) within the city limits are numbered serially in a contiguous, serpentine manner, beginning in the upper right hand corner of the map. This method of ordering has two advantages; 1) the blocks may later be combined, if desired, especially in those in­ stances where adjoining blocks have no households, and 2) this ordering makes it easy to distribute the segments through the city. In many instances maps for some of the smaller towns may not be available In situations of this kind, a sketch showing the street pattern will suffice, and it is relatively easy to make. b) Cruising to Obtain Eve Estimates of Number of Households In B_losj£ The purpose of the second operation, cruising, is to obtain estimates of the number of households in each block, and is most efficiently carried out by a team consisting of a driver and a recorder. When doing the cruising, the number of estimated households for each side of the block are re­ corded; later these can be totaled to get the number of households for the entire block. It should be emphasized that the unit is the household, and not the number of struc­ tures. Many houses contain several households, but these can usually be estimated by observing the number of mail boxes, entrances, size of the structure, etc. After the cruising operation has been completed, and the number of households recorded for each side of a block, the four sub-totals for each block are added up and recorded on the appropriate line of Form 1. This form makes possible an orderly arrangement of the data, and greatly facilitates the selection of the blocks which are to be used as sample areas. III. Assignment of Sampling Units to Blocks Through the cruising operation we now have an up-to-date estimate of the distribution of households by blocks. The chigan i'ATE LLEGE /3/A8 Bio cl No. MICHIGAN MEDICAL NEEDS SURVEY City No. of House­ holds Concord. Michi g a n No. of SampIing Units Cum. No. of Sampling Units Form I. lock No. __4_ IT « 31 10 11 12 aa 16 17 IS 20 21 22 n 66 12 27 28 30 31 32 70 71 72 26 27 76 28 78 30 80 81 86 21 87 88 89 .50 91 LL 100 No. of "| No. of HouseSampling holds Units SCCIAL RESEARCH SERVICE (CRH) Cum. No. of Sampling Units 150 next step is to combine these households into sampling units. By referring to Form 2, we know that kk sampling units are to be made for the town of Concord in Jackson county, and that only one sample segment is to be selected for the sample. In the illustration in Form I, the cruise estimate of the number of households is 2 3 8 . sampling units are to be made units averaging in size For this community, kk (Form 2). If we make samr-ling (i.e., in number of households) 22 8 . kk we should get kk sampling units containing on the average households. p.1* By rcunding to the nearest w h o 1e number we arrive at the figure 5 as the size of the sampling unit. The next procedure is to complete Form 1, thus creating kk sampling units of nearly equal size. If the number of households in any block divided by the average size of the sampling unit is less than 2, this block should be combined with the follow­ ing block in the list until this quotient for the combined blocks exceeds p. If blocks are combined, this should be so indicated on the sheet by means of a bracket. This quo­ tient should be entered in the column of Form 1 headed "Hum­ ber of sampling units." total this column. After this is done for all blocks, If it does not equal kk, it will be neces­ sary to make some slight modification in the size of the sampling unit, such as rounding to 5".2 instead of 5, or com­ bining two continuous blocks where the sampling unit total for the combined blocks is different from the sum of the in­ dividual block same ling units. The ratio of the cruise number of households tc the uni­ verse number of same lino uni t c wi 11 vary from city to city. The important rule, however, is that the number of sampling units indicated in Form 2 be created, and shat these units be nearly equal in number of households. IV. Selection of Sanrle blocks The column in Form 1 heeded "C ursula tive number of sampling units" shows us in which block any given same ling unit lies. In the case of Concord, Jackson county, the sampling unit which is to be included bn the state-wide sample is unit number 16, and by referrin- ay in to iorm. 1 it is apparent tbs t this uni t falls in block 12. The iden­ tification numbers cf the units which arc to be included in the samrlc wore :icked. by ? ranter: systematic scheme in such a manner that each unit has an equal chance of being chosen, thus minimising the possibility cf bias for those areas in which the samrlc ere a must be •oterr.ined by the cruising me t r:o 1 herein described. It is important tc note that the identification numbers given in Form 2 refer tc sam.rlinq units, and net tc blocks. The selection cf blocks is merely a separate stage of sampling, and only a means of locating 'she actual sr.mg ling segments themselves. After the sampling segment (or segments) has been located, the same instructions app y>" as for acl the other :HIOAM rATE LLE3E MICHIGAN 16/48 HEALTH S0C» L RESEARCH SERVICE (CRH) SURVEY All information in this schedulelT strictly confidential and under the exclusive control of the Social Research Service of Michigan State College. Names of all persons referred to in this schedule will not be quoted or made public in any way. 1-3 Schedule Number: Post Office Address: Sample area: 1-rtural ( ) l-3_ 4-5 County: 2-Village ( ) 3-Metro, area ( 4-5 ) 4-City ( ) 6_ Segment number 7-9. Interviewer's initials: Code for Row 16, Page 2: 10 Interview Typ* Check one - Not an informant 1 - Informant (female head or single male head), Part I (family) and Part III 2 - Informant Part I (self only), II and III 3 - Informant for Part I (family), II and III C Code for Row 18, Page 2: Call ( Number First Call ( ) (____ ) ( ) (Person No.)_____________ Inter view Completed Not completed ' Second Third My name is .... and I'm from Michigan State College. W e ’re making a health survey in homes all over the state of Michigan. Could you give me a few minutes of your time right now? For instance, we want to know if any member of your family has had any of these symptoms in the last six months... But first I need to know what people there are here, and their relation to the head of your family. What is the name of the head of the household? And may I ask his (her, your) age? Who else is there in the family? And his (her) age? What relation is he (she) to the head of the household? Code for symptoms: ' If negative, enter a dash( ) If positive and untreated (except for home care), enter If positive and Treated by non-M.D., enter "2" If positive and treated by M. D. or dentist, enter "3 1 Code for vaccination and Immunization: If not vaccinated andhas not had disease, enter a dash (__ If vaccinated and hasnot had disease, enter “1" If not vaccinated hutHas had disease, enter "2 If vaccinated and hashad disease, enter ' 3 If under minimum or over maximum age Indicated, enter 4 For each symptom reported as positive: ... "Did You (he, she) do anything about (for) it or not? (If other than home treatment or remedies ask:) "Was this an M . D. or not? COPYRIGHT by SOCIAL RESEARCH SERVICE of MICHIGAN STATE COLLEGE, June, 1948 Permission to reproduce any portion of this schedule must be obtained in writing. PART I MEDI CAL NEEDS (H.W,S.D,P,L,B,3S,GC.N) '— — .1........ ad cf Household Relation 11 Name: to Head 1 Leave Blank (Code for relation to head) 11 T Indicate Sox in each column: 1-Male 2-Female 12 i Indicate age in each column (01 - 99) 13-14 I 1 T Leave Blank (Code for No. in family) 15 Code for Informant c-l-2-3 1& I 1 1 Enter i(l" for each person present at interview: 17 4 Code for person'3 number 18 1 2 5 < 6 ? texnlained Loss of Weight: Persons over 18: 10 lbs. or more in s t b mo. Persons under 1 8 : any unexplained loss of weight 19 1 Continued Loss of Appetite 20 Unexplained Tiredness: regularly 21 1 1 Running Ear or Ears: watery, bloody, pus 22 t Poor Vision: for distant or close work, e.g. reading 23 1 Repeated Nosebleeds Not Due to Blow or Injury 2*+ J Persistent Headaches 25 Toothache 26 1 Unable to Chew Food: teeth "sore" or missing 27 1 i 1 Sore Mouth: due to plates or bridges 28 1 Repeated or Frequent Bledding Gums 29 1 1 :rsistent Skin Rashes or Itching of Skin: -- "breaking out" I j 1 (One week or more) 301 1 Lumps or Discolored Patches on Skin 311 ! Persistent Pains in Chest 32! ] Persistent Cough: (except colds in chest) 33! ! Coughing or Spitting Blood 3*+ Severe Shortness of Breath: after doing light work 35 Asthma or Hayfever 3bj Repeated or Persistent Backache 37! Persistent Pains in the Joints 3 8 ! ! Open or Running Sores or Ulcers That do not Heal 39 T scented or Persistent Swelling of Ankles: (Two weeks or more) 40 Repeated Vomiting: (Several days or more) 41 .-peated or Prolonged Pains in Stomach or Anywhere in Abdomen 421 "Ruoture". Hernia, or Wearing of Truss 431 ainting Spells: Stuttering; Stammering: Nervous Breakdown: Fits: Convulsions 44 tcidental Injuries: broken bones, head or severe injuries, accidental poisoning, snake bites, etc. 45 cr each person with one or more ''l's" in column, "Do you think -- netdj to see a doctor?" 1-Y, 2-N 46 or each person with ,fl" in row 46, "What would you say is the aln reason -- hasn't seBn a doctor?" 1-Lack of time 2-Symptoms not thought serious 3-Too expensive 4-0ther (specify on reverse side) 47 I or each person with one crv more '^'s" in column, "Did the M.D. advise -- to go to a hospital?" 1-Y, 2-N 46 Did he (she) go? 1-Y, 2-N 49 Leave blank (individual health care code) 5f Leave blank (total No. of positive symptoms) 51 Enter "1" if not working now due to illness or injury 52; Total days off in last 6 months due to illness 53-54 Number of times --- has seen a doctor in last 6 months (a) at doctor's office 55 (b) at ycur home 56 ) Times -- has seen dentist in last 6 months 57 1 Vaccination or Immunization for: (a) Smallpox (all over 1 year old) 58 (b) Diphtheria (6 months to l6 years) 59 (c) Whooping cough (6 months to 16 years) 60 11 * 2 i ! 1 1 e • 9 * ii 7 T 1 i : 1 1 ! 1 1 1 1 * ' r 1 .. 41 ... ... . i __i__ L 1 1 ..._ 4I ,,.. 1 . ... 1 1 )____ — I. ... | ___ i . 1 1 I 1 ____1 — i i I i ! 1 i . ; 1 I j *3 T _. . 1 ^ : 7 . i , ' 5 ' 6 '7 ! 8 * 9 (1 (3) PART II PRACTICES AND OPINIONS REGARDING HEALTH SERVICES Schedule N o . Now there are all kinds of doctors Just like there are all kinds of other 1-3 __ __ people. Seme people like what one doctor does. Others like what another doctor does. Suppose you think about the experiences you have had with different doctors. Just think of one or two you have liked best - we are not interested in their names - and tell me what you especially Tilted about them. A B 5 Do you have any other comments? C Now think of one or two you dldn't like so well What didn't you like about them? A __________________________________ ____ B Anything else? C A. You Just mentioned when we were talking about the doctors you liked that one thing you thought was important was that (A)______________ Do you feel that that is pretty typical of doctors in general, or not? 1-Typical ( ) 2-Not typical ( ) 3-Uncertain ( ) 10 B. Well, how about your statement that (B) Is it typical? 1-Typical ( ) 2-Not typical ( 11 You also said that (C) typical? 1-Typical ( ) ) 3-Uncertain ( ) Do you feel that that's 2-Not typical ( ) 3-Uncertain ( ) 12 I think that covers the things you said you liked about the one or two doc­ tors you liked best. Now here are one or two things you mentioned about doctors you didn't care as much for. You said (A) ___ ■ (Mention only items which are not clearly the opposite of those named In Questions 4, 5. and 6.) Do you feel that that is prettytypical of doctors in general, or not? 1-Typical ( ) 2-Not typical ( ) 3-Uncertain ( ) How about your statement that (B) typical of doctors in general, or nob? 1-Typical ( ) 2-Hot typical ( Y u also said that 1-Typlcal ( ( C) Do you feel that thattis 3-Uncertain ( ) ld_ ___________ . Is that typical? 2-hbt typical ( ) 3-Uncertain ( ) 15_ Do you have any other feeling about doctors in general, either one way or the ether? 16_ 17_ Or. the whole, have you benn satisfied with the help you have received from doc­ tors , or not? , , „ 1-Satisfied ( ) 2-Not satisfied ( ) 3-Uncertaln ( ) li-Had no help( ) 18 (If to 18) What sort of things aren't you satisfied with? . 19 3o far as you know are your friends and relatives satisfied with the help they have received from doctors, or have you heard them make complaints? l-3atisfled ( ) 2-Made complaints ( ) 3-Uncertaln ( ) 2 (If "made complaints") What sort of things have you heard them say? 2 Have you (and the members of your family) always been able to get a doctor’s help when you needed It, cr have you had trouble In getting a doctor's help? 1 -Always get one ( ) 2-Had trouble ( ) 3-Uncertain ( ) U-Haven’t tried( ) (If 2 is checked on 22) When was the last time this happened? 1-Year ________ 2-Month _____________ Would you mind telling me about it? (Probe for: A. Who needed a doctor. ____________ ____ B. Why couldn't a doctor come. C. What did you do about it. D. What were the results.) _____________________________ In your experience do you think that we have enough doctors or do we need more general M.D.'s or more specialists; or both? 1-Have enough { ) 2-General M.D.'s ( ) 3-Specialists ( ) 4-Both ( ) 5-Uncertain ( ) 6-Need more good doctors ( ) (If 2 or 3) Why do you feel that way? ____ ____________________ (If "more nedded" in 2d) Do you feel that this problem is so serious that some­ thing ought to bo done about it? 1-Yes ( ) 2-No ( ) 3-Uncertain ( ) How about other communities (towns): do you think they have enough or are more needed? 1-Enough ( ) 2-More needed ( ) 3-Uncertain ( ) If some community (town) needed more doctors do you have any Idea how it could get them? Don’t know ( ) In some communities doctors have organized into a group so they can work together In diagnosing and treating illnesses. Have you heard about such plans? 1-Yes ( ) 2-N.; ( ) 3-Unccrtaln ( ) Would you prefer such a group plan or would you prefer to gotoa doctor who practices alone? 1-Group ( ) 2-0ne doctor ( ) 3-Uncertaln ( ) Have you (or any members of your family) ever gone to an osteopath or other doctor who was not an M. D.? 1-Yes, self only ( ) 2-Yes, other member^ ) 3-Self and others ( ) A-No ( ) 5-UneertaIn ( ) (If "yes" to question 35 ) Was he an osteopath, chiropractor, or other kind of doctor? 1-Osteopath ( ) 2-Chiropractor ( ) 5-Othtr (specify) ______________________________________ _ _ _ _ _ When was the last time you (or some member of the family) went to him? 1-Within tho last year ( ) 2-Before the last year ( ) 3-Uncertain ( ) What kind of trouble did you (or members of your family) have the last time you went to him? How do you think the training of osteopaths compares with the training of M.D.’s? How do you feel abcut using doctors who are not M.D.'s, such as osteopaths? 1-Would use only M.D. ( ) 2-Would use only non-M.D. ( ) 3-Would use non-M.D. for certain things ( ) ^-Uncertain ( ) HOSPITAL SERVICES Now I'd like to ask a few questions about hospitalization. Have you (or any member of your family) been a hospital patient within the past year or two? 1-Yes, self only ( ) 2-Yes, other members ( j 5-Self and others ( ) 4-No ( ) (If "yes" to question 42) Would you mind telling happened? 41 me about how much that cost you the last time it ^2 Does this amou t include Doctor, hospital and nursing expense or not? 1-Yes ( ) k-wo t ) ^-Uncertain v f In general, how do you feel about the doctors gave you (or other members of 45 medical and surgical services which the the family)while in the hospital? 44 In general, how do you feel about the accommodations and services which were provided by the hospital? PAYMENT FOR MEDICAL SERVICE: ilext I’d like to ask seme questions about payment for medical service. Do you (or any members of your family) carry insurance to pay for all or part of: A, Hospital bills? 1-Yes ( 2-No ( 5-Uncertain ( ) 46 B , Fees for surgery? 1-Yes ( 2-No ( 5-Uncertain ( ) 47 C , Doctors' fees other than surgery? 1-Yes ( 2-No ( 3-Uncertain ( ) 48 (If "yes" to 46, 47, or 48) Which members are covered? A. Hospital: J-All members ( ) 3-Other (specify) __________ 4-Uncertain ( ) 49 B. Surgery: 1-All members ( ) 5-Other (specify) 2-Head only ( 4-Uncertain ( ) ) 50. C. Doctors' fees: 1-All members ( ) 5-Other (specify) 2-Head only ( 4-Uncertaln ( ) ) 51 3-Uncertaln ( ) 52 What is the name of the Insurance company? 1-Blue Cros3 or Blue Shield ( ) 2-Fraternal ( 4-Other (specify) ____________________________ Have you (or any members of your family) ever carried hospital insurance and dropped it? 1-Yes, self ( ) 2-Yes, other members ( ) 3-Self and others ( 4-No ( ) 5-Uncortaln ( ) (If "yes") Why? Was it Bluo Cross (or Blue Shield)? 1-Yts ( ) 2-No ( ) 3-Uncertaln ( ) 53 5^ (If informant does not have hospital Insurance) Would it be possible for you to get Blue Cross insurance if you wanted to? 1-Yes ( ) 2-No ( ) 3-Uncertaln ( ) 55 In general, do you think insurance plans for paying hospital and doctor bills are a good Idea, or not? 1-A good idea ( ) 2-Not a good idea ( ) 3-Uncertain ( ) 56 158 (6) Have ycu heard or read about a plan In which people would pay a certain percen­ tage of their Income to the government and In . return members of the family would have their doctor and hospital bills paid for by the government? 1-Yes ( ) 2-Yes, socialized medicine ( ) 3-No ( ) 4-Uncertaln ( ) even :gobd if "no", Idea, or If you haven't heard about It before,) do you think that It's a not? 1-A good Idea ( ) 2-A good Idea, with reservations ( 3-Not a good Idea ( ) 5-Uncertaln ( ) Why do you feel that way? Do you feel that there are any (good) (bad) paints? Good p o i n t s _______________________________ _____ 57_ 58_ 59 60 Bad points 61 '62 (If not mentioned above, ask) Have you heard or read about the Murray, Wagner, Dingell Bill? 1-Yes ( ) 2-No ( ) 3-Uncertaln ( ) 63 (If not mentioned above, ask) Have you heard or read about "socialized medicine"? 1-Yes ( ) 2-No ( ) 3-Uncertaln ( ) 64 (If "yes") What do you feel are its good points? _____________________________ 65. 6b What are Its bad points? __________________________ 67. 68 COMMUNITY AND PUBLIC HEALTH Now I would like to ask you a question or two about Public Health Service. Have you (or any members of your family) been personally examined or advised by a public health Inurse or officer within the past year? 1-Yes. self only ( ) 2-Yes, other members ( ) 3-Self and others ( ) d-l!o ( ) 5-Uncertain ( ) Do you feel that this community has any major health problem? 1-Yes ( ) 2-No ( ) 3-Uncertaln ( ) 69 70 (If "yes" t# 70) What is It? ______________________________________ _ _ _ In some places a Comi.iittee or Have you heard l-Yts representatives of different organizations have gotten together In council to ftovelop plans for improving health in Ithe community. of anything like that? ( ) 2-No ( ) 3-Uncertain ( ) (If "yes" tp 7 1 ) Do you think representatives of the organizations in this community ought to organize some kind of a health committee or council? 1-Yes ( ) 2-No ( ) 3-Uncertaln ( ) 4-0ne in community now ( ) 71j_ 72_ * Have you heard or read of the Michigan State Medical Society? 1-Yes ( ) 2-No ( ; 3-Uncertaln { ) (If "yes") In general do you feel that it works mostly for Doctors' in­ terests, the interest of people in general, or for both? 1-Doctors' interests ( ) 2-Interest of people ( ) 3-Interests of both ( ) 4-Neither ( ) 5-Uncertain ( ) In general how do you feel about what it does? 1-Like ( ) 2-Dlslike ( ) 3-Unaertain ( ) Code for No. 16 (Symptoms page) Give number of informant (Prom symptoms page line 18 ) PART III CONTROL ITEMS Schedule No. Code for No. 16 (Symptoms page) Give number of informant (Prom symptoms page line 18) HEALTH CONTROL ITEM Do you have a certain doctor to whom you (and members of your family) go for most of your ills? 1-Yes ( ) 2-No, go to more than one ( ) 3-Nc, have no doctor ( ) 4-Uncertain ( ) (If "1" or "2" to 6) Is he an M.D. (are they M.D.'s)? 1-Yes ( ) 2-None M D.'s ( ) 3-0ne M D. ( ) 4-Uncertaln ( ) 5-Other________________ _______________________ (If "none M D.'s" to 7) What kind of doctor (s) is he (are they)? (If "yes" to 6) In what town is his office located? 9-10 How far is it to his office? (Check code in miles) 1-1 to 5 ( ) 2-6 to 10 ( ) 3-11 to 15 ( ) 4-16 to 20 ( 5-21 to 25 ( ) 6-26 to 30 ( ) 7-0ver 30 ( ) ) OTHER CONTROL ITEMS (If single person answering for self only, go to 14) Who is the main earner of your family? 1-Informant ( ) 2-Other person (specify relation to informant)__________( Are you (is he) (is she) employed? 1-Yes ( ) 2-No ( ) (If "no") Why aren't you (isn't she) (isn't he) employed right now?_________ What kind of work did you (did he) (did she) do when you were working (when he was working) (when she was working) (when he was-ITving)? (If "yes") What kind of work do you (does she) do? Job________________________________________________________________ Industry_____ ("What sort of place work at?") ("What do they make or do there?") Are you (13 the family's breadwinner) a member of any union? 1-Yes ( ) 2-No ( ) (If "yes") Is that CIO, A P of L, or Independent? 1-CI0 { ) 2-AFL ( ) 3-Independent ( ) 4-Uncertain ( ) J.DU (8) Do you remember the name of the last school you went to?_ What was the last grade or year you completed in school? -Some high 1-No schooling 5-Completed high 2-1-1* years grammar 7-Some college 3-5-7 years grammar ) 8-Completed college 4-Completed grammar i 20 (If married fem*le head responding) What was the last grade or year your husbtand completed in school? 5-Some high 1-No schooling 2-1-4 years grammar 6-Completed high 7-Some college 5-5-7 years grammar 4-Completed grammar 8-Completed college (If Open Country) About how many miles Is It to the nearest town having a doctor? 0-Town or city ( ) 1-1 to 5 ( ) 2-6 to 10 ( ) 3-11 --2 -- 25 - '( ) 6-26 to 30 ( ) 7-over 4-16 to 20 ( ) 5 1 to Do you live on a farm? 0-Town or city ( 1-Yes ( ) 21 22 to 15 ( 30 ( ) 23. 2-No ( (If "yes") Did you (the head) work 100 or more days off the farm during the past year? 1-Yes ( ) 2-No( ) 3-Uncertaln ( ) 24_ Do you or your family rent or own the place where you live? 1-R«_nt ( ) 2-Own ( ) 3-0ther (specify)__________________ 25. Is there a telephone in your home (place where you live)? 1-Yes ( ) 2-No ( ) 26_ (If "yes") Is It listed either In your name or your family's name? 1-Yes ( ) 2-No ( ) 27. Dc you happen to have a car (in your family)? 1-Yes ( ) 2-No ( ) 28_ Do you have running water In the place where you live? 1-Yes ( 7 2-Ns ( ) 29. Do you have an inside toilet la the place where you live? 1-Yes ( ) 2-No ( ) 20. Do you read a daily newspaper? 1-Yes ( ) 2-N0 ( 31. ) 52. Do you have a radio? 1-Yes ( ) 2-No ( ) If "yes": What one radio station dc you listen to most?________ 33-?4_ i, Havw you ever heard a 5-minute health news radio program called "TfeU Me Doctor t 3-Uncertain ( ) 1-Yes ( ) 2-No ( ) '.............. 35. a service of the (If yes) Do you happen to know * is r that program Michigan State Medical Society, or not? 1-Yes ( ) 2-N ( ) 3-Uncertaln ( ) ?6. About how many times a week do you listen to It? 37. O 1-3 or m o r e ( ) 2-0nce or twice ( ) 3 -LvSS than once a W e e k ( ) 4-Practlcally never ( ) About how often do y^u go to church or religious services? 1-Oncc a week or oftener ( ) 2-1 to 3 times a month ( 3-Occasionally ( ) 4-Ncver ( ) 38_ What denomination do you consider yourself? ______________ _ 39. Did you (or any member of your family) serve In any of the United States Armed Forces during World War II? 1-Yes. self only ( ) 2-Yes, other members ( ) 3-3elf and othcrs( ) 4-IIQ ( ) 5-Uncertain ( ) 40 (9) Dc y .u remember for certain whether or not you voted In the 1944 Presidential election? ~ 1-Yes, voted ( ) 2-1J.,, didn't vote ( ) 3-No, toe young to vote ( ) 4-Uncertaln ( ) In general, which of the political parties do you favor in the Presidential election this fall? 1-Republlcan ( ) 2-Democratic ( ) 3 -Other (specify) ________________ ( ) ^-Uncertain ( ) 4l_ 42 Would you look at this card and tell me the letter opposite the figure that comes IS It ti , B, C, or what 0 43 (1) B-$l,ooo up to $2,000 (2) C-$2. ooo up to $3,000 (3) B-$3 ,000 up to (*> E-$4,000 up to $5,000 (5) F-$5. 000 up to $7,500 (6) 0-$7. 500 up to $10,000 o o o A-Und er $1,000 <6- (C) (7) H-$1C1,000 or more Code for economic level 1-A ( ) 2-B ( Leave blank (code for population of community) ) 3-C ( ) 4-D ( ) 44_ 45_ M