COMPARISON OF TRADITIONAL DETECTION METHODS FOR STOOL PATHOGENS WITH MULTIPLEX MOLECULAR DETECTION IN A MID-SIZED CLINICAL LABORATORY By Elizabeth M. Ward      A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of  Clinical Laboratory SciencesMaster of Science  2016           ABSTRACT COMPARISON OF TRADITIONAL DETECTION METHODS FOR STOOL PATHOGENS WITH MULTIPLEX MOLECULAR DETECTION IN A MID-SIZED CLINICAL LABORATORY By Elizabeth M. Ward    Clinical laboratories may choose to determine the benefits or detriments of closed-system multiplex assays for identification of gastrointestinal pathogens as compared to the performance of traditional culture, staining, and antigen detection methods. A 60 residual stool specimen study compared the performance and costs of the current testing algorithm in use at a 208-bed acute care facility in Michigan, and the BioFire FilmArray GI panel. A study of the costs, both to the laboratory and the patient, of four multiplex molecular assays and the traditional testing algorithm was also conducted. Performance of the BioFire FilmArray was not statistically different from the traditional algorithm. The multiplex molecular methods were not more costly to the laboratory or the patient for all possible testing order sets for traditional testing.          Copyright by ELIZABETH M. WARD 2016    iv  ACKNOWLEDGEMENTS  The contributions of Dr. Alan Petkus, Director of Microbiology are acknowledged for the use of laboratory equipment and supplies, staff time, and invaluable advice and expertise. Acknowledgement also is due to the staff of the Microbiology department for their work on the cultures compared in this study. The students of BLD 801 and the faculty of the MSU BLD program deserve thanks for their feedback on the presentation of the first draft of this thesis. Dr. Rudrik at the State of Michigan Department of Health and Human Services Bureau of Laboratories gave some timely advice and shared invaluable knowledge. Finally, acknowledgement is certainly due Dr. Frances Pouch Downes, Dr. Rachel Morris, and Dr. Alan Petkus for their time and energy serving on the thesis committee.                v  TABLE OF CONTENTS    ...vii INTRODUCTION .......................................................................................................................... 1  CLINICAL FEATURES AND DIAGNOSTIC CHALLENGES .................................................. 4  PATHOGENESIS ........................................................................................................................... 6 Salmonella ................................................................................................................................... 6 Shigella ........................................................................................................................................ 7 Campylobacter ............................................................................................................................ 8 Aeromonas and Plesiomonas ...................................................................................................... 8 Yersinia enterocolitica ................................................................................................................ 9 Shiga Toxin Producing Escherichia coli ................................................................................... 10 Vibrio ......................................................................................................................................... 11 Clostridium difficile................................................................................................................... 11 Rotavirus ................................................................................................................................... 13 Norovirus and Sapovirus ........................................................................................................... 14 Adenovirus ................................................................................................................................ 14 Astrovirus .................................................................................................................................. 15 Cryptosporidium........................................................................................................................ 15 Giardia ...................................................................................................................................... 15 Cyclospora cayatensis ............................................................................................................... 16 Entamoeba histolytica ............................................................................................................... 16  EPIDEMIOLOGY ........................................................................................................................ 18  TRADITIONAL DETECTION METHODS AND SUSCEPTIBILITY TESTING .................... 19 Salmonella, Shigella, and Campylobacter ................................................................................ 20 Aeromonas and Plesiomonas .................................................................................................... 21 STEC and Yersinia enterocolitica ............................................................................................. 22 Vibrio species ............................................................................................................................ 23 C. difficile .................................................................................................................................. 24 Rotavirus and other Enteric Viruses.......................................................................................... 24 Cryptosporidium, Giardia and other Enteric Parasites ............................................................. 25  COMMERCIAL SYNDROMIC MULTIPLEX DETECTION METHODS ............................... 27   vi  Film Array ................................................................................................................................. 27 BD Max ..................................................................................................................................... 28 Luminex GPP ............................................................................................................................ 29 Nanosphere Verigene Enteric Pathogens Test .......................................................................... 30  PROJECT OVERVIEW ............................................................................................................... 32 Project Objectives ..................................................................................................................... 33  MATERIALS AND METHODS .................................................................................................. 35  RESULTS AND CONCLUSIONS............................................................................................... 39 Objective One:........................................................................................................................... 39 Objective Two: .......................................................................................................................... 42 Objective Three: ........................................................................................................................ 47  DISCUSSION ............................................................................................................................... 49                 vii  LIST OF TABLES  Table 1: Reported cases of gastroenteric pathogens in Michigan, 2008-201218 Table 2: Comparison of targets detected by multiplex molecular assays  Table 3: Results of 60 specimen comparison study between BioFire FilmArray and methods currently in use at Hospital A Laboratory  Table 4: Results of 60 specimen comparison study between BioFire FilmArray and traditional methods in use at Hospital A  Table 5: Estimated test costs to the laboratory  Table 6: Estimated test costs to the patient5  Table 7: Average times in hours to actionable result for traditional assays run at Hospital A  Table 8: Statistical comparison of BioFire FilmArray and traditional methods47  Table 9: Comparison of published performance characteristics for multiplex molecular assays48       1  INTRODUCTION  Clinical Laboratory Science has, as all technological fields, never been stagnantprogress is rapid, constant, and necessary. Molecular diagnostic techniques have made the rapid, accurate detection of both genetic disease, like the detection of the BRCA-1 and BRCA-2 genes associated with breast cancer1, and infectious diseases2 possible. However, molecular techniques for infectious disease have traditionally been more expensive than traditional culture and microscopy3. In many instances, department directors or physician groups could weigh the benefits of the molecular assayhigher sensitivity, faster throughput, reduced hands-on time for clinical laboratory staffagainst the difference in cost to determine whether the molecular assay is the best choice for their facilities. For instance, currently available molecular assays can detect methicillin resistant Staphylococcus aureus (MRSA) colonization from a nasal swab in around one hour4. Previously, patients whose histories indicated potential MRSA colonization or risk were placed in contact isolation for up to 24 hours upon admission to the hospital until MRSA cultures were confirmed negative. Removing patients from unnecessary isolation within an hour presents a significant cost-savings5 to the health system, although it may incur additional laboratory expense. Because of changes in reimbursement and the inception of value-based purchasing, a new era in which a healthcare facility must consider expenses as a whole organization rather than department-by-department is beginning; hospitals are no longer reimbursed for inpatient acute care services based on the number of services they provide but on the quality of care provided.6 As new molecular assays come to market, clinical laboratory directors and managers must carefully consider the benefits and the costs of molecular technologies not only to their own departments, but to the organization as a whole.    2  There are more than 1.7 billion cases of diarrheal disease worldwide each year; these diseases are the second leading cause of death among children under five years of age7. Although diarrheal disease is neither as common nor as frequently fatal in the United States as it is in economically poor areas lacking sanitation8, it is still a significant concern in patients in high risk categories, such as the very young, very old, and the immunocompromised9.  In the United here are an estimated 211 to 375 million episodes of diarrheal illness each year, with 1.8 million hospitalizations and 3,100 deaths.10 In Michigan, where 208-bed acute care Hospital A is located, diarrheal disease is not a leading cause of death. In 2013, out of a total 92,463 deaths, 356 deaths (0.3%) in Michigan were attributed al infections, compared to the 1,892 deaths (2.0%) attributed to Influenza and pneumonia or the 1,209 deaths (1.3%) caused by septicemia, and no deaths due to salmonellosis or shigellosis were reported.11 Evaluating molecular methods for detection of gastroenteric pathogens will require careful consideration due to low mortality and incidence of diarrheal disease due to infectious causes in the state and community.  Traditional laboratory detection of gastroenteric pathogens has relied heavily on bacterial cultures, ova and parasite (O&P) examination for parasites, and antigen detection by enzyme immunoassay for more elusive pathogens, such as Rotavirus. Molecular methods for the detection of enteric pathogens have been available for many years, but they had been previously inaccessible for small or medium sized laboratories due to their expense, cumbersome equipment, specialized and costly reagents, and/or the need for designated areas appropriate for open-method amplification assays. Additionally, an individual test had to be run for each analyte, so testing for a large panel of targets was not feasible. Recently, however, multiplex, closed-system molecular methods entering the market have made the possibility of switching to a   3  molecular detection or molecular screening method feasible for most laboratories of any size. These methods can detect multiple pathogens simultaneously, often of different taxa, in mere hours rather than the days or even weeks required for a full workup using traditional methods. They can also detect organisms, such as Norovirus, to which most small to mid-sized hospital laboratories had not previous access. Additionally, the base cost of multiplex molecular assay kits are greater than the reagents for traditional tests, but the benefits to patient and the health system may outweigh increased expense.                  4  CLINICAL FEATURES AND DIAGNOSTIC CHALLENGES  Gastroenteritis is defined as an inflammation of the lining of the intestinal tract, which is commonly caused by bacteria, viruses, or parasites. The symptoms of gastroenteritis can include diarrhea, abdominal pain or cramping, vomiting, headache, fever, or chills.12 Because of the broad range of symptoms, the expansive list of etiologic agents for gastroenteritis, and infectious agents that vary widely by region, identifying microbial causes of gastroenteritis and gathering useful national and international surveillance data is a daunting task. This confusion is compounded by the myriad providers ordering stool tests: general and family practice including, of course, gastroenterologists, but also emergency department physicians, and inpatient physicians, among others. These professionals have different diagnostic goals, protocols, and training, and they may take disparate paths to the diagnosis and treatment of their Often, when surveillance data is collected using diagnosis codes, diarrheal symptoms caused by non-infectious sources are included because the code applied to the patient was non-specific (e.g. the ICD-9 code 787.91 could The combined challenges of varying, non-specific symptoms and complicated test menus and algorithms, often encompassing both in-house and reference laboratory testing, can lead to vastly different laboratory order sets for patients with very similar clinical presentations. Where one physician whose patient presents with abdominal cramping and diarrhea might order a single stool culture, a clinician in an emergency department might order stool cultures, ova and parasite, Clostridium difficile testing, Rotavirus testing, and Cryptosporidium and Giardia testing. If the cause was parasitic, the first physician would miss the diagnosis, but if the cause was Norovirus, both physicians would miss   5  it. Buss et al. [infectious gastroenteritis] is often left unidentified for many reasons, including limited laboratory test menus, physician ordering 13                    6   PATHOGENESIS  The panel of enteric pathogens targeted in stool workups is largely dependent on the geographical area in which the patient population is located. The following pathogens are those included on the multiplex molecular panels later discussed. Salmonella  Salmonella is a genus of motile bacteria belonging to the family Enterobacteriaceae. There are two species in the genus, and these are further subspeciated: Salmonella enterica, the more common human pathogen, has six subspecies; Salmonella bongori is the second species. Traditionally, Salmonella enterica has been serotyped using three antigenic surface structures: the somatic (O), flagellar (H), and Vi antigens. Because the serotype of an isolate is most important for public health and epidemiology purposes, rather than treatment, serotyping is often either performed or confirmed at public health laboratories. Many of the serotypes correlate to specific disease states or are associated with specific vehicles of transmission, so their identification can aid in the investigation of a potential outbreak. Recent developments in the science of whole-genome sequencing of Salmonella isolates promise advancement in the epidemiologic genotyping of Salmonella for public health officials and epidemiologists. The disease state associated with Salmonella is classified as one of two types: typhoidal and non-typhoidal. Non-typhoidal salmonellosis is the classic intestinal disorder characterized by diarrhea, fever, and abdominal cramping and is the most common presentation in the United States. These infections, which can affect people of all ages but are most common in infants and young children, can last a week or more and can lead to systemic infections, usually in immunocompromised hosts. Salmonella species are common in the microbiota of many animal   7  species and populations, so foods of animal origin are typically the source of human infections. However, outbreaks have been linked to direct contact with animals, non-animal food products, water, and person-to-person transmission. 14 Shigella  Shigella species are a genus of non-motile bacteria belonging to the family Enterobacteriaceae. Four subgroups of the genus are classically recognized: Shigella dysenteriae (subgroup A), Shigella flexneri (subgroup B), Shigella boydii (subgroup C), and Shigella sonnei (subgroup D). Genetically, Shigellae are very similar to Escherichia species, making their identification in molecular assays more challenging than some of the more distinct genera; Shigella would be regarded as serologically defined anaerogenic biotypes of E. coli15. Unlike many of the other enteric pathogens, Shigella species are strict human pathogens; their only reservoirs are humans and some of the other large primates. Therefore, transmission is more likely to be caused by person-to-person contact or fecally contaminated food and water. Shigellosis is most common in institutions in which hygiene is difficult to control, such as child care centers, and in developing areas in which sanitation is problematic. Shigella sonnei is the most common cause of shigellosis in the United States, where it can cause protracted outbreaks in institutional populations. These outbreaks often have a large number of asymptomatically infected individuals, contributing to the spread of the organism. Shigella can cause bloody and non-bloody diarrhea. Classic shigellosis presents as watery diarrhea with fever and cramping and may progress to dysentery, characterized by scant stools containing blood, mucus, and pus. Complicated cases of shigellosis can result in hemolytic uremic syndrome (HUS) and Reiter chronic arthritis. Increasingly widespread antimicrobial resistance among Shigella strains is a growing concern. 16   8  Campylobacter  The members of the family Campylobateriaceae cause zoonotic infections in humans and are found not only in food animals like poultry, cattle, sheep, and pigs but also in domestic animals commonly treated as pets. Because so many animals act as reservoirs for these organisms, Campylobacter infections are the most common bacterial cause of gastroenteritis and typically are associated with improperly handled or cooked food products, most notably poultry. Milk-borne unpasteurized milk products as a health food in recent years has brought milk-borne Campylobacter back into the limelight. Campylobacter jejuni is the most commonly reported enteric pathogen isolated from the genus, with Campylobacter coli as the second most common. Because the disease states are indistinct, many laboratories do not distinguish between these two species. Campylobacter gastroenteritis presents with mild to severe bloody or non-bloody diarrhea, abdominal cramping, and fever lasting several days to more than one week. Although most cases are self-limiting, relapses have been noted in about 5-10% of untreated patients. Extraintestinal and systemic infections are possible in Campylobacter infections. Campylobacter jejuni infections have been documented as preceding the development of Guillain-Barré syndrome, reactive arthritis, and 17 Aeromonas and Plesiomonas  Aeromonas and Plesiomonas are not currently reportable organisms in the United States and so may be considered of less epidemiologic relevance than the other enteric bacterial pathogens, but they nevertheless cause gastroenteritis and can lead to serious complications in at-risk populations. Aeromonas species are the only documented human pathogens in a family of water-borne bacteria, the Aeromonadaceae. The clinically relevant Aeromonads are found in   9  fresh or sometimes brackish water, and outbreaks are typically associated with food products: meats, dairy, and fresh produce. Aeromonas gastroenteritis usually presents as acute, watery diarrhea but may range to dysentery or chronic illness. Many patients experience abdominal pain, and a lesser number may have nausea, vomiting, and/or fever. Complicated cases can lead to hemolytic uremic syndrome (HUS) or kidney disease. The disease is typically self-limiting, and the most common pathogens are A. caviae, A. hydrophila, and A. veronii. It is important to consider isolates of Aeromonas carefully in stool culture and to examine both the culture details (relative quantity of organism, etc.) and very strong evidence that some aeromonads are gastrointestinal pathogens, there is no convincing evidence, at present, that all fecal isolates of Aeromonas are involved in diarrheal 18 The genus Plesiomonas is a member of the Enterobacteriaceae family, and the only species in the genus, P. shigelloides, is an emerging pathogen causing food and water-borne enteric infections in humans. Plesiomonas diarrheal infections can present as secretory or dysenteric, and symptoms can last for two weeks or more; there also is a subacute, chronic form that can last from two weeks to three months. Plesiomonas, like Aeromonas, is isolated from aquatic reservoirs and has also been recovered from mammals, fish, amphibians, and reptiles.19   Yersinia enterocolitica Yersinia species are classified as Enterobacteriaceae, and there are three human pathogens in the genus. Yersinia enterocolitica is the cause of gastroenteritis. Y. enterocolitica gastroenteritis is usually caused by consumption of contaminated food or water, and has been associated with consumption of unpasteurized milk, several raw meats, and with pre-packaged deli meats. Consumption of raw or undercooked pork products is the main risk factor for   10  gastroenteritis because swine are the most common carriers of the pathogenic serotypes of the species. The specific serotypes are associated with disease severities, and presentation can range from self-limiting gastroenteritis to terminal ileitis. Common symptoms include diarrhea, fever, and abdominal pain, usually resolving within seven days. Septicemia has developed in high risk patients.20  Shiga Toxin Producing Escherichia coli  Serogroup O157 and other serogroups are commonly classified as STEC (Shiga toxin producing E. coli). The STEC group produces one or both of Shiga toxins 1 and 2; several variants have been identified of Shiga toxin 2. Although the production of Shiga toxins is an important factor in the virulence of STEC organisms, there are other virulence mechanisms that amplify the disease potential of the organism, including attachment and effacement. STEC can cause bloody or non-bloody diarrhea ranging from mild to severe. These organisms have been associated as a precursor to Hemolytic Uremic Syndrome (HUS). It is estimated that 4% or more of those infected with STEC develop this potentially life-threatening condition, and E. coli O157 is estimated to cause at least 80% of the HUS cases in the US annually. STEC do not cause fevers; abdominal cramping, however, is a common symptom. STEC is a colonizer in the gastrointestinal systems of both dairy and beef cattle, and ground beef has been implicated in more outbreaks than any other vehicle of transmission. Outbreaks have also been associated with other animal products: raw milk, sausages, roast beef; vegetable products: apple cider, raw vegetables, bean sprouts; and unchlorinated municipal water. The infectious dose of this organism is quite low (<200 colony forming units), so it spreads easily from person to person also, and outbreaks have been recorded in institutional settings (schools, day cares, long term care facilities). Although O157 STEC is the most common in the United States, more than 150   11  non-O157 STECs have been characterized, and in some countries, these comprise more cases than O157. 21 Vibrio  Vibrio is one of six genera belonging to the family Vibrionaceae. Vibrios are typically associated with marine environments, although it is of note that some species, including Vibrio cholerae, require less sodium for growth and can be found in fresh water. There are more than 200 serogroups of Vibrio cholerae, based on the somatic (O) and flagella (H) antigens, but groups O1 and O139 are the etiologic agents of epidemic and pandemic cholera outbreaks. Cholera is less common in developed countries like the United States than it is in underdeveloped countries because it is associated with poor sanitation and use of contaminated surface water for cooking, bathing, and drinking. Vibrio parahaemolyticus is a major cause of foodborne intestinal infections worldwide; it is the leading cause of foodborne intestinal infections in Asia, and it is the Vibrio species most often isolated in the United States. Vibrio gastroenteritis by species other than V. cholerae is characterized by watery diarrhea, nausea, abdominal cramping, vomiting, low-grade fever, and chills. Because Vibrio is a marine organism and is associated with consumption of shellfish, especially during the warmer months, it is less common in the inland areas of the United States than on the coasts.22  Clostridium difficile  Clostridium difficile is a spore-forming, anaerobic gram positive rod responsible for more than 250,000 cases of diarrheal disease per year in the United States. C. difficile is an often-identified cause of hospital-acquired diarrhea and is estimated to have a cost exceeding $1 billion annually. Clostridium difficile infection (CDI) is associated with toxin production, either the enterotoxin TcdA or the cytotoxin TcdB, or both. Since C. difficile has been isolated from 3-5%   12  of the healthy population, 30% of healthy infants, and 20-30% of the sedentary population, it is important to prove the production of one or both toxins to diagnose CDI. The pathogenesis of CDI requires a combination of colonization and toxin-production. Host risk factors for infection are: age, hospitalization, recent bowel surgery, treatment with proton-pump inhibitors, and recent treatment with antibiotics. CDI symptoms can range from mild and self-limiting diarrhea to foul-smelling, slimy, bloody diarrhea to development of pseudomembranous colitis. The antibiotics most commonly associated with development of CDI are clindamycin, fluoroquinolones, and broad spectrum cephalosporins. Symptoms may occur immediately following a course of antibiotics or as much as 4-6 weeks later. It is important to remember that C. difficile is but one agent of antibiotic-associated diarrhea (AAD) and is by no means the only one. Hypervirulent strains of C. difficile have emerged and been associated with outbreaks in the United States that affected younger patients, those with no underlying disease, and even outpatients. Hypervirulent CDI is associated with megacolon, rupture of the large bowel, and often death. CDI is troublesome in hospitals and community settings: relapse rates range from 20-50%; it is a spore-former, and many commonly-used cleansers do not effectively kill the spores; and asymptomatic carriers can act as reservoirs in both hospital and community settings.23  Since molecular detection methods for C. difficile became available in 2009, reporting of CDI has increased, but it is unclear whether actual cases of CDI have increased. Molecular tests detect toxin genes but not active toxin production, so the clinical significance of positive molecular assay in the absence of toxin testing or with negative toxin testing is unclear. A study published in 2015 followed 1416 hospitalized patients tested for C. difficile after 72 hours or more of hospitalization.  All patients were tested for toxin and PCR, and were grouped into three categories: Toxin +/PCR +, Toxin -/PCR +, or Toxin -/PCR -. The results of the study were that   13  -related complications and deaths occurred in patients with positive toxin immunoassay test results. Patients with a positive molecular test result and a negative toxin immunoassay test result had outcomes that were comparable to patients without C. difficile by 24 Rotavirus Rotaviruses belong to the family Reoviridae and are classified into seven groups, A through G, serologically. Human infections are almost exclusively associated with types A, B, and C, with A the most common. Rotavirus gastroenteritis, although not as deadly in developed countries as in underdeveloped countries, represents a large number of gastroenteritis cases and a large healthcare burden worldwide. Children, especially those under two years of age, are affected the most frequently. The initial source of infection in Rotavirus outbreaks is often an older sibling or parent with a subclinical infection, but large outbreaks have been documented in daycare centers and other, similar institutional environments. Rotaviruses are shed from the intestinal tract before the onset of symptoms and even after diarrheal symptoms have ceased, and they can survive on environmental surfaces quite well, increasing their spread and duration of outbreak. Rotavirus, like all viral gastroenteritis, presents with watery diarrhea, vomiting, anorexia, abdominal pain, and fever, although the symptoms are more severe with Rotavirus than with other viral pathogens. Rotavirus is strongly associated with acute, severely dehydrating gastroenteritis. Symptoms typically last for a shorter period than bacterial gastroenteritis, often 2-3 days. Temperate climates typically experience a cycle of Rotavirus infections peaking in the winter to spring seasons. Because serious complications can occur and Rotavirus is a major concern in developing countries, vaccines were developed and have been increasing in popularity. Since their inception, Rotavirus vaccines have reduced the prevalence of the disease   14  so much that Norovirus is now recognized as the leading cause of acute gastroenteritis in children in the United States.25  Norovirus and Sapovirus Norovirus and Sapovirus belong to the family Caliciviridae in which four genera cause gastroenteritis in human hosts. Norovirus is considered the most important cause of non-bacterial gastroenteritis in both developed and developing countries and has been associated with outbreaks from contaminated food, water, and surfaces. The environmental stability and low infectious dose of this virus are key factors in its clinical and public health importance. Large outbreaks of Norovirus have been associated with large institutional settings: hospitals, nursing homes, schools, military settings, and, perhaps most famously, cruise ships. The Norovirus genotype GII.4 has been found to cause up to 80% of all Norovirus outbreaks in several countries. Sapovirus causes gastroenteritis mainly in children and has been reported as a causative agent of outbreaks among children in daycare settings. No vaccines or antivirals are currently available for Norovirus or Sapovirus, although interest in a vaccine for Norovirus is quite high.26 Adenovirus  The Adenoviridae, of which there are two genera, have a large number of serotypes isolated from humans, but most serotypes do not have a clearly demonstrated role in human disease. Two serotypes commonly referred to as the enteric Adenoviruses have been associated with diarrheal disease in infants. Outbreaks of enteric Adenovirus typically occur in children under age four, with symptoms lasting longer than typical viral diarrhea, sometimes 5-12 days. Prolonged or chronic diarrhea can develop in immunocompromised patients. Significant outbreaks have been reported in hospitals and daycare centers.27    15  Astrovirus Astrovirus infections in humans are primarily associated with children, but there have been reported cases in military troops, nursing homes, and immunocompromised patients. Astrovirus causes mild to moderate gastroenteritis with symptoms lasting 1-4 days. There have been documented subclinical infections.28 Cryptosporidium Cryptosporidium species are intracellular parasites that cause diarrheal outbreaks in human hosts, typically associated with contaminated water or food. In immunocompetent hosts, Cryptosporidium gastroenteritis is a short-term illness characterized by watery diarrhea, abdominal cramps, headache, fatigue, low-grade fever, vomiting, malabsorption and weight loss. Diarrhea will often last 1-2 weeks but is self-limiting. In immunocompromised hosts, however, cryptosporidiosis can progress to a systemic and fatal infection. Although children under the age of 5 are at the highest risk for disease, Cryptosporidium can cause protracted, severe gastroenteritis in immunocompromised adults and the elderly. It is more common for adults in developed countries to experience gastroenteritis than those in developing countries, most likely due to a higher rate of exposure, and thus immunity, in developing countries.29  Giardia Giardia duodenalis is the most common cause of intestinal parasitosis in humans worldwide. The literature often refers to this organism as Giardia lamblia, but G. duodenalis is the currently accepted nomenclature for the species causing human disease. This flagellate infects humans through fecal-oral transmission or ingestion of cyst-contaminated food or water. The inoculum required for disease is low10 to 100 cysts. All age groups are at risk of giardiasis, and outbreaks have been found among diverse populations: daycare centers, hikers,   16  and the immunocompromised. Asymptomatic infections are common, and are associated with G. duodenalis Group B, while the more pathogenic Group A is associated with symptomatic infections. Following an incubation of 12-20 days, G. duodenalis will present with sudden onset watery diarrhea accompanied by low-grade fever, chills, nausea, and abdominal pain. Diarrhea caused by G. duodenalis is often characterized as explosive and foul-smelling but lacking the presence of blood or mucus.30  Cyclospora cayatensis  Cyclospora is an intestinal coccidian parasite. Its oocysts can be detected in the feces of infected humans. Following infection with Cyclospora, clinical symptoms will present within 7-8 days and persist for 2-3 weeks, although cases have been documented with symptoms from 1 day to more than 100 days. Cyclospora is endemic in warm climates, but in temperate climates, like the United States, outbreaks are most often associated with consumption of imported fruits and vegetables. The primary symptom of Cyclospora gastroenteritis is watery, non-bloody diarrhea, and it may be accompanied by abdominal cramps, fever, fatigue, nausea, headache, and vomiting. The symptoms, and their duration, are more problematic among immunocompromised patients.31 Entamoeba histolytica  Entamoeba histolytica is one of the most common protozoal human pathogens worldwide. The cyst of E. histolytica is the infectious form and is typically encountered by ingestion of contaminated food or water or by direct fecal-oral transmission, but sexual transmission can also occur. Cysts in the feces are the diagnostic stage. However, two morphologically similar species, E. dispar, which is not capable of producing disease, and E. histolytica, which is, exist and distort the understanding of the true infective rate of E.   17  histolytica.  E. histolytica is endemic in tropical and subtropical climates, but it is of particular concern in the United States among immigrants from warmer climates, the institutionalized, and residents of the warmer, southern states. Symptomatic infections with E. histolytica can occur with or without tissue invasion. Clinical symptoms may present as dysentery, colitis, or, more rarely, ameboma (disseminated abscess). Amebic dysentery is characterized by diarrhea with cramping, lower abdominal pain, low-grade fever, and bloody stool with mucus. Amebic colitis -ulcers of the intestinal tissue accompanied by intermittent diarrhea over a long period of time.32                  18  EPIDEMIOLOGY  Currently, only Salmonella, Shigella, Campylobacter, Cryptosporidium, Giardia, Cyclospora, STEC (including O157), Yersinia enterocolitica, and Vibrio are reportable diseases in Michigan, so reliable data about the prevalence of the other pathogens is difficult to access. For instance, the Centers for Disease Control and Prevention (CDC) estimates that there are more than 5.4 million cases of Norovirus illness in the United States annually,33 but since there currently is no reliable laboratory network for identifying and reporting Norovirus, true incidence data is unavailable. In 2012, the most recent year for which data is publicly available, Michigan experienced the incidence of these reportable gastrointestinal conditions displayed in Table 1. Table 1: Reported cases of gastroenteric pathogens in Michigan, 2008-2012 including both total cases for 2012 and the mean 5 year prevalence for 2008-2012  Total Number of Cases 2012 Mean 5 year prevalence (2008-2012) Campylobacter 1209 1147 Cryptosporidiosis 352 320 Cyclosporiasis 0 3 Giardiasis 547 611 Salmonellosis 946 926 STEC (Shiga-toxin producing E. coli) 285 116  Shigellosis 260 238 Vibriosis (non-cholera) 7 6 Yersinia enteritis 23 22  Data adapted from: Michigan Department of Community Health34  Shiga toxin producing E. coli (STEC) were reported under multiple categories until 2010 when all categories were consolidated.      19  TRADITIONAL DETECTION METHODS AND SUSCEPTIBILITY TESTING  Hospital A is located in Michigan. The most commonly identified bacterial enteric pathogens are Salmonella species, Shigella species, Campylobacter species, Yersinia enterocolitica, and STEC, as well as Aeromonas species, and Plesiomonas shigelloides. Clostridium difficile testing is performed under a separate algorithm from the other bacteria. Viral pathogens of interest in this area are Rotavirus and Norovirus and, to a lesser extent, Adenovirus. Parasitic pathogens of interest are Cryptosporidium and Giardia species, although travel history may justify broader parasitic testing. Other geographic locales may require methods to detect additional pathogens; for instance, on the East Coast of the United States, additional testing to quickly isolate and identify Vibrio species would be of increased clinical relevance. Traditional detection methods for gastroenteritic pathogens vary between laboratories, but most use an algorithm to detect at least a minimum number of pathogens. Because significant variations exist based on size of laboratory, number of billable tests performed per year, regional trends, and personal preferences of laboratory administrators and physician groups, the algorithm in use at Hospital A is described here. This algorithm was developed based on recommendations Clinical Microbiology Procedures Handbook35 and to maintain compliance with accreditation standards from the College of American Pathologists.36 The basic order for stool culture is designed to routinely detect Campylobacter, Salmonella, and Shigella. A blood agar plate (SBA), MacConkey agar (MAC), and Hektoen Eneteric agar (HE) incubated at 35-37°C in room air and selective Campylobacter (CAMPY) media incubated microaerophilically at 42°C, are used to detect these pathogens.   20  Additionally, a Gram Negative (GN) broth is inoculated and incubated at 35-37°C for 4-6 hours before being sub-cultured to another HE plate to increase the sensitivity of the culture. Using the media above, an experienced clinical laboratory scientist should be able to detect and identify Aeromonas, Plesiomonas, and Vibrio, in addition to the pathogens listed.  For cases of bloody diarrhea, or at the discretion of the physician, an expanded stool culture includes additional media as recommended for the isolation of Yersinia enterocolitica (cefsulodin-irgasan-novobiocin (CIN) agar) and E. coli O157:H7 (sorbitol MacConkey (SMAC) agar). At Hospital A, the expanded stool culture also includes Shiga toxin detection by enzyme immunoassay, Card STAT! EHEC, which is performed on a GN broth following 18-24 hours of incubation. A negative stool culture can be reported no sooner than 52 hours after set up, although many small and medium sized laboratories only report culture results during one shift, so results may take longer to report. Isolation and identification of pathogens may take 72 hours or more to complete, and follow-up testing on reportable organisms may take weeks. Salmonella, Shigella, and Campylobacter  Salmonella species can be recovered from both MAC and HE agar on which they will typically present as lactose non-fermenters. HE agar is more selective and contains a hydrogen sulfide (HS) indicator to make detection of Salmonellae easier and to facilitate detection of lactose- fermenting Salmonella. Shigellae are non-lactose fermenters on both MAC and HE. Lactose negative colonies from either MAC or HE are routinely identified at Hospital A using the Vitek 2 instrument system from bioMérieux. If phenotypic identification yields either Salmonella or Shigella, latex agglutination kits are used to confirm. The Wellcolex Color Salmonella Rapid Latex Agglutination kit from Thermo Scientific can identify only the most   21  common human O serogroups, so it should be used with professional judgement. The Wellcolex Color Shigella Rapid Latex Agglutination kit from Thermo Scientific detects common O serogroups as well. Confirmed Salmonellae and Shigellae, and those identified by the Vitek 2 and not confirmed with latex agglutination for which a high index of suspicion exist, are forwarded to the state public health lab for confirmation and further epidemiologic subtyping. Vitek 2 MIC testing is performed on all Salmonella and Shigella isolates at Hospital A and reported following CLSI guidelines.37  Campylobacter has long presented challenges in laboratory isolation and identificationit is particularly fastidious, requiring microaerophilic conditions and 42°C to grow properly. Current isolation methods at Hospital A support the growth of C. jejuni and C. coli but are not conducive to isolation of additional significant pathogens, C. upsaliensis and C. fetus. Campylobacter is currently identified at Hospital A to the genus level only according to the following algorithm: a CAMPY agar plate is incubated microaerophilically for 48 hours. Upon 48 hour examination, wet-looking, oxidase- and catalase-positive colonies are Gram stained; if the Gram stain is suggestive of Campylobacter (a curved, Gram negative rod), the culture is reported as Campylobacter species isolated and reported to the state health department. Routine antimicrobial susceptibility testing is not recommended for Campylobacter isolates, but given the documented resistance to fluoroquinolones among the genus, fluoroquinolone susceptibility testing should be considered by the physician for patients for whom antimicrobial therapy is required.38 Aeromonas and Plesiomonas Aeromonas species grow well on selective enteric media, but many strains are lactose fermenters and are indistinct among the normal flora. Most strains of Aeromonas are beta   22  hemolytic on SBA and can be distinguished from other beta hemolytic organisms by positive oxidase and indole reactions. Organisms matching this description are identified with the Vitek 2 at Hospital A. Routine susceptibility testing is not currently recommended, and therapy is seldom required in gastroenteritis cases. Susceptibility testing may be required for complicated cases or epidemiologic purposes, as fluoroquinolone resistance has been documented and some strains have been observed as carriers of carbapenemases.39 Plesiomonas shigelloides is the only member of the Enterobacteriaceae that is oxidase positive. It produces lactose negative colonies on MAC and HE. Many strains are also beta hemolytic on SBA. Beta-hemolytic, lactose negative colonies that are oxidase positive are routinely identified using the Vitek 2 at Hospital A. Routine susceptibility testing of Plesiomonas in cases of gastroenteritis is not recommended, but in a complicated case it would be warranted as the organism has documented resistance to aminopenicillins and variable resistance to both tetracycline and the aminoglycosides.40 STEC and Yersinia enterocolitica  testing from patients with acute community-acquired diarrhea (regardless of patient age, season of the year, or presence or absence of blood in the stool) be simultaneously cultured for E. coli O157:H7 (O157 STEC) and tested with an assay that detects Shiga toxins to detect non-O157 STEC41 These guidelines were developed because selective testing, such as testing only bloody stools, will fail to detect all STEC infections. Several studies have indicated that most STEC isolates were recovered from patients without apparently bloody stools. While these updated guidelines are under review at Hospital A, the current algorithm dictates that bloody stools (identified by physician order, diagnosis code, or visual inspection) are cultured on SMAC and   23  tested with ImmunoCard STAT! EHEC to detect Shiga toxins 1 and 2, which the assay does differentiate. Non-sorbitol fermenting colonies on SMAC are identified using Vitek 2, and isolates of O157 STEC or stools positive for Shiga toxin 1 or 2 are forwarded to the state public health lab for follow-up testing. Routine antimicrobial susceptibility testing is not currently recommended for O157 isolates because use of antibiotics is associated with a higher rate of HUS than non-treatment.42 Y. enterocolitica is motile at room temperature but non-motile at 37°C; optimal growth occurs at 25°C to 28°C. Although Yersinia will grow on MacConkey agar, its growth is much slower than other Enterobacteriaceae, and it is easily outcompeted by normal flora. Hospital A uses specialized Yersinia agar (containing a combination of cefsulodin, Irgasan, and novobiocin) incubated at 25°C to isolate Yersinia. Y. enterocolitica produces charactecolonies with a red center and a lighter edge under these incubation conditions at 24 to 48 hours. All such colonies are identified using Vitek 2. Routine susceptibility testing for Y. enterocolitica gastroenteritis is not recommended, but it is necessary in complicated cases or those progressing to septicemia.43 Vibrio species  Hospital A does not include specialized media for Vibrio species in routine stool culture. Vibrio species are oxidase and catalase positive, and most strains grow on routine MAC and HE agars as colorless colonies. SBA plates are routinely screened for oxidase positive Gram negative rods, so most Vibrios would be identified on Vitek 2 as part of routine stool workup at Hospital A especially if the physician notes a high level of suspicion for Vibrios due to travel or dietary history. Routine antimicrobial susceptibility testing is not currently recommended for Vibrios in cases of gastroenteritis.44   24  C. difficile  The current algorithm for CDI detection at Hospital A is t enzyme immunoassay that detects glutamate dehydrogenase (GDH), indicating the presence of C. difficile, and toxins TcdA and TcdB, which are not differentiated by the assay. Specimens that give a result of GDH+/toxin  or GDH-/toxin + are considered discrepant and are resolved with loop-mediated isothermal amplification (LAMP), specifically, the Illumigene C. difficile assay by Meridian Bioscience. Clostridium difficile represents a major complication in the consideration of multiplex molecular panels. Community-acquired CDI is becoming more common, so CDI is no longer a concern only among inpatient populations,45 a fact that supports its inclusion on a comprehensive panel. However, colonization is also becoming more prevalent, even among the healthy adult population, in which it is estimated at 3-5%more asymptomatic C. difficile carriers than there are people with CDI in the community and in the hospital.46 Routinely reporting C. difficile results as a component of a gastroenteritis multiplex panel could lead to over-diagnosis of CDI, data which is included in value-based purchasing decisions and can affect revenue and reimbursement for healthcare facilities, and over-treatment of C. difficile in colonized individuals, which could select for antimicrobial resistance mechanisms in GI flora, increase adverse drug events, and increase the likelihood of CDI development following treatment.47 Current treatment options for CDI are vancomycin therapy and fecal transplantation, but susceptibility testing is not performed as C. difficile is not traditionally cultured.48  Rotavirus and other Enteric Viruses Viral testing at Hospital A is limited to Rotavirus. Antigen detection is performed via enzyme immunoassay cartridgeCard   25  Bioscience. Adenovirus and Norovirus testing are performed at Hospital Aupon order by physician. Astrovirus and Sapovirus tests are not routinely performed at Hospital A reference laboratories and, if requested, samples would need to be sent to an additional laboratory for testing. Cryptosporidium, Giardia and other Enteric Parasites  Cryptosporidium and Giardia may be detected in a number of ways: antigen detection by enzyme immunoassay cartridge tests or direct fluorescent stains is common in many laboratories, and these organisms are included in all ova and parasite (O&P) panels. Many small and mid-sized laboratories no longer offer ova and parasite in house because Giardia duodenalis and Cryptosporidium species are the most common causes of parasitic gastroenteritis in the United Stated. A screen for Cryptosporidium and Giardia may be performed either in conjunction with an ova and parasite at a reference laboratory or with the recommendation that a negative should prompt the physician to order a full ova and parasite if parasitic infection is suspected. This is the algorithm in use at Hospital A, and screening is performed with an EIA cartridge assay, the Alere Giardia/Cryptosporidium Quik Chek. Although there are a few therapeutic agents used to treat Cryptospordium infections, none has proven completely effective. Nitazoxanide is currently the only FDA approved treatment, but it is limited to use in immunocompetent patients.49 Treatment for giardiasis is typically metronidazole or nitazoxanide; although the disease is self-limiting in immunocompetent hosts, treatment can help lessen the duration of the disease and prevent person-to-person transmission.50  Cyclospora and E. histolytica are included in the ova and parasite performed at Hospital AHospital A to detect these pathogens. The treatment for Cyclospora cayetanensis is trimethoprim-sulfamethoxazole.   26  Symptoms typically decrease after about 2.5 days of therapy, and a 7 day course is recommended.51 Because of the complicated nature of E.dispar/E.histolytica morphology, molecular testing methods may allow for more accurate diagnosis and more targeted chemotherapy. Treatment for E. histolytica may require a combination of luminal amebicides (such as paromomycin), which target cysts, and tissue amebicides (like metronidazole), which target trophozoites. No resistance has been documented to date, but follow-up stool examination is always recommended to detect treatment failures.52                   27  COMMERCIAL SYNDROMIC MULTIPLEX DETECTION METHODS The use of syndromic approach in medicine typically refers to a review of the pattern of symptoms a patient experiences, the syndrome, followed by testing and treatment targeted toward the likely cause of the syndrome. In the molecular era, companies have interpreted this as a need for paneled tests that can detect the most likely pathogens for a clinical syndrome all at once.53 l syndromes are seldom specific to a single pathogen, so detection strategies that allow multiple agents to be simultaneously considered can have a significant impact on infectious disease management since multiparametric molecular diagnostic tests can provide a 54 A multiplex molecular assay offers opportunities to detect cases of traditionally difficult to culture organisms. However, the multiplex assay would need to include all common pathogens in order to meet these expectations. An excellent example comes in the form of Aeromonas. Most of the currently available molecular assays do not include Aeromonas in their test menus, but this organism has some clinical relevance. A laboratory moving to a molecular algorithm would have to determine whether culture for Aeromonas would still be performed or whether it would no longer routinely be detected at that facility. A recent review article by Gray and Coupland summarized: Diagnosis of syndromic infections represents a new pathway for the diagnosis of infection and newer molecular diagnostic tools, which will streamline workflows in the routine diagnostic setting, must incorporate all relevant pathogens in order to improve patient management [sic.]55 Film Array  The BioFire FilmArray56 is a table-top multiplex molecular assay notable for its compact size, extremely limited hands-on time, and high number of targets per assay. The FilmArray Gastrointestinal (GI) Panel is run on Cary-Blair preserved soft to diarrheal stools. The specimen   28  is stable in Cary-Blair for up to 4 days at room temperature or refrigerated. The assay takes approximately 1-2 minutes of hands-on time to prepare, and results are available in about 1 hour and 10 minutes. The pouches are run individually, one pouch per instrument, so the number of tests per day is determined by the number of units in the laboratory. Each unit could feasibly run 7 tests in an 8 hour shift.  The FilmArray GI panel detects 22 targets. Not only can the FilmArray detect Campylobacter, Salmonella, and Vibrio to the genus level, it can differentiate between five diarrheagenic E. coli groups and detect four parasitic, five viral, and an additional four bacterial targets. (See Table 2) The FilmArray does not currently offer any way to narrow the field of detection. Some pathogens, like Vibrio, are more endemic in certain regions than others. In the Midwest, they are very uncommon, but they are included on each GI panel. Additionally, the GI panel contains several targets that have not been routinely tested for or reported at Hospital A in the past, including the non-STEC diarrheagenic E. coli groups and the viral targets other than Rotavirus. Transitioning to the FilmArray GI would require education initiatives for all physician and mid-level providers, and these groups may or may not want the tests provided. Multiplex molecular assays are assigned CPT codes based on the number of targets they detect. In the case of the FilmArray GI, the CPT code will be for the highest number of targets despite the fact that physicians may find some of the targets useless. BD Max   The BD Max57 is currently FDA approved for only the Enteric Bacterial Panel, which detects Salmonella, Campylobacter (jejuni and coli only), Shigella spp. and Enteroinvasive E. coli, and Shiga toxins 1 and 2, without speciation for any genus. There are additional panels under development58: a parasite panel that will detect Giardia, Cryptosporidium, and E.   29  histolytica, and extended panels that will detect common viruses (Rotavirus, Norovirus, and Adenovirus) and additional bacteria (Aeromonas, Vibrio, and Y. enterocolitica). The Bacterial Panel is performed on soft to diarrheal stools and can be performed on unpreserved or Cary-Blair preserved stools.  Unpreserved or Cary-Blair stools can be stored up to 24 hours at room temperature or 5 days at 2-8°C prior to testing. The test takes about 1-2 minutes of hands-on time, and results are available in about 3 hours total. The BD Max performs runs of 12 samples and has the capacity for two runs at a time, so it can run 24 samples total with the same or different run start times.59 A total of 4 runs, or 48 samples, could be run in an 8 hour shift.     Although the promise of individual panels for bacterial, extended bacterial, parasitic, and viral gastroenteritis is of valueit offers precisely the opposite of the FilmArraythere is also concern in this offering. If clinicians typically order all the panels, then the cost could be the same or greater, even, for the patient. With batches of 12 taking 3 hours to run, 4 tests per patient could greatly affect the throughput time for samples and the technical staff time required. Additionally, until the expanded panels are FDA approved, the bacterial panel offers detection of only 4 potential pathogens and would require additional culture, enzyme immunoassay (EIA), and other testing.  Luminex GPP  The Luminex xTAG Gastrointestinal Pathogen Panel (GPP) simultaneously detects 14 of the most common gastrointestinal pathogens. The panel can be run on unpreserved or Cary-Blair preserved stools from patients with signs and symptoms congruent with infectious colits or gastroenteritis. This assay can be run on the Luminex 100, Luminex 200, or MAGPIX Luminex instrumentation. The batched run can include up to 90 patient specimens and takes about 5 hours to complete.60  Although other organisms, most notably, Aeromonas and Pleisiomonas, are not   30  currently available on the panel, the nature of the Luminex technology has always been such that any laboratory can develop assays for it, so a laboratory-developed test could be built and validated if desired.  , 14 targets is still within the most expansive CPT code, which includes 12-25 targets. The same cost concerns arise with Luminex, then, as with FilmArray. The software used for interpretations can be so physician ordering could still be targeted to certain panels: a bacterial panel, a viral panel, etc. However, this does not decrease the cost to run the assay. Additionally, a 90 patient run is large for a small or mid-sized laboratory, like Hospital A, where 10 stool cultures per day would be considered high volume.  Nanosphere Verigene Enteric Pathogens Test   The Nanosphere Enteric Pathogens test detects 5 bacterial pathogens, Shiga toxins 1 and 2, and 2 viruses. The test uses Cary-Blair preserved stool with a hands-on time of about 5 minutes and a run time of about 2 hours.61 Stool specimens are stable at 2-8°C in Cary-Blair for up to 48 hours after collection. Each Verigene processer can run 1 specimen at a time, so multiple processors would be necessary to manage specimen volume. 62 One processor could run 4 specimens in an 8 hour shift.  The Nanosphere offers a nicely targeted panel that detects the most common bacterial and viral pathogens. However, with such a limited panel, additional cultures and EIAs would be necessary to complete the laboratoreagent costs would be negated, much like the BD MAX.     31  Table 2: Comparison of targets detected by multiplex molecular assays BioFire FilmArray, BD Max, Luminex xTAG, and Nanosphere Verigene. Target BioFire FilmArray BD Max Luminex xTAG GPP Nanosphere Enteric Pathogens Test Aeromonas spp.   Extended Panels (under development)   Campylobacter (jejuni, coli, upsaliensis)    Bacterial Panel (jejuni and coli only)    (C. jejuni, C. coli, C. lari) Clostridium difficile (Toxin A/B)       Plesiomonas shigelloides      Salmonella    Bacterial Panel     Yersinia enterocolitica    Extended Panels (under development)    Vibrio (parahemolyticus, vulnificus, cholerae)    Extended Panels (under development)    Vibrio cholerae       Enteroaggregative E. coli      Enteropathogenic E. coli      Enterotoxigenic E. coli       Shiga-like toxin-producing E. coli (stx-1. Stx-2, E. coli O157)    Bacterial Panel (stx1 and stx2 only)    (stx 1 and 2 only) Shigella/Enteroinvasive E. coli    Bacterial Panel    (Shigella spp. only) Cryptosporidium    Parasite Panel (under development)    Cyclospora cayetensis      Entamoeba histolytica    Parasite Panel (under development)    Giardia lamblia    Parasite Panel (under development)    Adenovirus F40/41    Extended Panels (under development)    Astrovirus      Norovirus GI/GII    Extended Panels (under development)     Rotavirus A    Extended Panels (under development)     Sapovirus (I,II,IV, and V)         32  PROJECT OVERVIEW  For Hospital A, the BioFire Film Array was used for the preliminary study comparing traditional stool pathogen detection to multiplex molecular assay. This assay format was selected because the instruments were already in place and in use for another specimen source panel, the claimed sensitivity and specificity of the assay were published, and the recent change in pricing and billing tiers made multiplex molecular assay a much more realistic choice than it was in 2014 when it was FDA approved. A 60 sample residual stool specimen collection was used to determine the usefulness of a syndromic multiplex molecular assay in the Hospital A setting. The goal of this project was to determine whether there is increased clinical value and cost benefit to be gained in switching to a molecular algorithm at Hospital A. Clinical value, here, is defined as an increase in sensitivity: more pathogenic organisms would be detected. Cost-benefit in this study is defined as either 1) the multiplex assay would be less expensive to run, less expensive to the patient, or 2) it would present such a significant increase in diagnostic capability that it would be worth a greater expense. Hypothesis 1: Traditional microbiology techniques are more sensitive than the BioFire FilmArray in detecting Salmonella species, Shigella species, Campylobacter species, Plesiomonas shigelloides, STEC, Yersinia enterocolitica,   Cryptosporidium, Giardia, and Rotavirus. Hypothesis 2: Syndromic traditional microbiology testing is less costly to the laboratory than multiplex molecular testing. Hypothesis 3: Syndromic traditional microbiology testing is less costly to the patient than multiplex molecular testing.   33  Project Objectives  The project objectives are to examine the performance of BioFire FilmArray GI and the cost of changing from traditional testing algorithms to a multiplex molecular method for Hospital A and similarly sized hospitals with comparable patient demographics. 1. Perform a comparison study between the BioFire Film Array enteric panel and current testing algorithm at Hospital A. a. Study will include 60 residual stool specimens from patients exhibiting diarrheal and/or bloody stools b. Discrepancies in test results for Salmonella species, Shigella species, Campylobacter species, Plesiomonas shigelloides, STEC, Yersinia enterocolitica,   Cryptosporidium, Giardia, and Rotavirus were resolved via reference laboratory testing, either at the Michigan Department of Health and Human Services state public health laboratory or Warde Laboratories, as commercial reference laboratory, when possible, and as appropriate to the organism. Failure to resolve discrepancies was discussed. Because of the complicated literature and conflicting recommendations surrounding molecular testing for C. difficile, the results of the FilmArray were not analyzed as part of this study, although the results were discussed. c. Discrepancies in test results for all other analytes detected by FilmArray were resolved at the discretion of the Director of Microbiology at Hospital A. d. Comparison of physician ordering practices with the actual infectious agent present in the specimen for the 60 specimens included: test orders placed by physicians were compared to any pathogens detected in the specimen   34  2. Complete a cost analysis to define financial and care benefits of each algorithm to determine which best serves the patients of Hospital A. This analysis was based on the following parameters: a. Estimated test cost to laboratory, including base cost per test and hands-on technical staff time, for each of the methods described  b. Test cost to the patient, based on currently available CPT codes, charge to an uninsured patient, and CMS reimbursement rates, for each of the methods described c. Care benefits were measured based on the time it takes to report an actionable result to the physician 3. Complete a comparison of the published sensitivity and specificity for each assay described (utilizing Negative Predictive Value and Positive Predictive Value) and the comparison study data. With only 60 specimens included in the study, it is unlikely that the study will have sufficient sample sizes for each analyte to reliably interpret statistically significant differences in performance.       35  MATERIALS AND METHODS Sixty residual stool specimens submitted to Hospital A for infectious diarrhea testing (stool culture, expanded stool culture, Cryptosporidium, Giardia, or Rotavirus) that included a diagnosis related to diarrhea or infectious diarrhea and were visibly liquid stools upon inspection were run on the BioFire FilmArray GI panel.  The performance of the FilmArray was compared to traditional tests.  If the FilmArray results were positive for an analyte that was not part of the traditional tests ordered, additional testing was performed, if available. Discrepancies were resolved for the following organisms: Salmonella species, Shigella species, Campylobacter species, Plesiomonas shigelloides, STEC, Yersinia enterocolitica,   Cryptosporidium, Giardia, and Rotavirus. Bacterial discrepancies were resolved via confirmation at the state public health laboratory, when possible. Parasitic discrepancies were resolved via ova and parasite from Warde Laboratory, and Rotavirus were resolved at Warde Laboratory via enzyme immunoassay. Where discrepancies could not be resolved, the failure to resolve will be discussed. All specimens for which reportable disease organisms were identified at Hospital A were forwarded to the state public health laboratory for follow-up testing as required by law or when requested.  Among the 60 study specimens, no STEC organisms were detected during the course of the study by either FilmArray or traditional methods. A STEC negative stool specimen was seeded with E. coli O157 to ensure the Film Array would detect a STEC. To prepare the specimen, 15 mL of a known EIA negative liquid stool in Cary-Blair was seeded with 1 mL of 0.5 McFarland suspension of E. coli O157 ATCC 35750.  The cost analysis included an estimated cost per reportable test for traditional and multiplex assays discussed above and the test cost to the patient.  Estimated cost per reportable   36  test analysis included reagent costs only. The costs for equipment and incidentals were not included because equipment may be purchased, rented, or placed under a reagent contract depending on company, hospital purchasing group, and other factors. Two of the four multiplex assays require equipment already in use at Hospital A but may require additional units or upgrades, while the other two would be new additions. To protect proprietary information, the reagent costs listed for the multiplex assays are based on the base price and do not include any hospital purchasing group contract discounts. Labor costs were evaluated using tech-time based on historical study data from Hospital A determining the number of work load units (one WLU is equivalent to one minute) per assay. The multiplex assays were calculated based on the  average salary for a certified Medical Laboratory Scientist: the median hourly wage for Medical and Clinical Laboratory Technologists is $28.57.63 The test cost to the patient was compared using currently available CPT codes and the estimated cash charge to a patient as well as the published CMS reimbursement for each CPT code. The costs to the patient were compared based on currently available CPT codes. Costs were calculated using the average charge to insurance and average CMS reimbursement for each CPT code.64 The cash charge for an uninsured patient was calculated based on Hospital Apractice of offering an automatic 40% discount on laboratory tests for all uninsured patients, so the charge reported is 60% of the average charge to insurance. Because none of the multiplex assays discussed are currently performed on outpatients at Hospital A, the costs are estimates only.   The time to actionable result was determined for each of the tests ordered by the physician for the 60 samples included in the comparison study. Time to actionable result was   37  defined as the number of hours from receipt of the specimen in the lab to the publication of a the nearest quarter hour. Time to actionable result was calculated for the tests ordered on any of the 60 specimens: stool culture and expanded stool culture, Shiga toxin by EIA, Cryptosporidium/Giardia screen by EIA, Rotavirus by EIA, ova and parasite, and C. difficile rapid screen by EIA with reflex to DNA when required. The identification of a bacterial pathogen in culture was considered actionable, because antimicrobial therapy can be prescribed, if necessary, based on typical institutional resistance patterns, i.e. antibiogram. The publication of required sensitivity testing for Salmonella and Shigella species was not required to meet C. difficile, a discrepant result on EIA routinely required confirmation by DNA testing to be considered actionable.  The mean and median total times to actionable result were calculated for each assay as well as the separate times for positive and negative assays. The median was included to control for possible outliers due to extenuating circumstances. All calculations were done in hours, except ova and parasite, which was conducted in days and hours. Finally, the sensitivity and specificity of the performance of the traditional methods and BioFire FilmArray were compared for Salmonella species, Shigella species, Campylobacter species, Plesiomonas shigelloides, STEC, Yersinia enterocolitica, Cryptosporidium, Giardia, and Rotavirus. These organisms are the currently detected organisms using the algorithm in place at Hospital A. Sensitivity and specificity were calculated using the formulas65: Sensitivity=  true positives           x100         true positives + false negatives    38  Specificity=  true negatives           x100         true negatives + false positives  It was recognized that it was unlikely that there would be a large enough sample size for all analytes included to yield reliably interpreted statistical analysis (a significant P value), but the data was analyzed. The mean of the study data for traditional method and BioFire Film Array -test and a probability of 95% (P=0.05) for each of the 9 analytes listed above. In order to compare the study means, the qualitative data was transformed to quantitative data by assigning a numerical value of 1 to positive results and a value of 0 to negative results. For those discrepant results for which a confirmatory assay was available, the confirmation determined whether the original result was considered a true positive or true accurate one. Statistical calculations were performed using Microsoft Excel.  Because the statistical power of the sample size was low, the published sensitivity and specificity of the         39  RESULTS AND CONCLUSIONS Objective One: Two discrepancies among the nine analytes of interest occurred in the 60 specimen comparison study. The FilmArray detected a Campylobacter species that was not detected by traditional culture. Attempts to isolate the organism on both Campylobacter selective media and chocolate agar were unsuccessful. The state public health laboratory does not have any additional means by which to identify Campylobacter, so no additional testing was available to resolve this discrepancy. The FilmArray also detected one Cryptosporidium that was not confirmed by Hospital A ova and parasite. A third possible discrepancy occurred: an organism isolated on one stool culture that was identified by Vitek 2 as Y. enterocolitca group but could not be speciated by biochemical testing. The specimen was negative on the FilmArray assay for this analyte. This result was excluded from the study, due to the clinical details of the case and the low burden of the organism in the presence of abundant normal flora visualized on primary culture plates. Comparison results for Salmonella species, Shigella species, Campylobacter species, Plesiomonas shigelloides, STEC, Yersinia enterocolitica, Cryptosporidium, Giardia, and Rotavirus are summarized in Table 3, and comparison results for the remaining analytes included on the FilmArray GI panel are summarized in Table 4these results were not included in statistical analysis. There was a discrepancy among the analytes not included in the study: a Cyclospora cayatensis was detected by the FilmArray that was not detected on the ova and parasite performed at the reference laboratory.      40  Table 3: Results of 60 specimen comparison study between BioFire FilmArray and methods currently in use at Hospital A Laboratory: Culture, ImmunoCard STAT! EHEC, QuikChek Cryptosporidium/Giardia, and ImmunoCard STAT! Rotavirus, including resolution of discrepant results.   Positives by Traditional Method Negatives by Traditional Method Positives on FilmArray Negatives on FilmArray Discrepancy Resolution Salmonella 3 57 3 57 n/a Shigella 3 57 3 57 n/a Campylobacter 6 54 7 53 Negative Y. enterocolitica 0 59 0 59 n/a STEC 1 0 1 0 n/a P. shigelloides 2 58 2 58 n/a Cryptosporidium 0 60 1 59 Negative Giardia 1 59 1 59 n/a Rotavirus 1 59 1 59 n/a                         41  Table 4: Results of 60 specimen comparison study between BioFire FilmArray and traditional methods in use at Hospital A: send out ova and parasite, Alere Quik Chek Complete C. difficile, Illumigene C. difficile by Meridian Bioscience, culture, send out enzyme immunoassays for Norovirus, Adenovirus, Astrovirus, and Sapovirus for additional analytes. For  A so testing is not routinely performed and no orders specific to those pathogens were placed for the study specimens.    Positives by Traditional Method Negatives by Traditional Method Positives on FilmArray Negatives on FilmArray Discrepancy Resolution Clostridium difficile (Toxin A/B) 5  55  10  50  2 positive; 1 negative; 1 unresolved*  Vibrio (parahemolyticus, vulnificus, cholerae) 0 60 0 60 n/a Vibrio cholera 0 60 0 60 n/a Enteroaggregative E. coli n/a n/a 2 58 n/a Enteropathogenic E. coli n/a n/a 11 49 n/a Enterotoxigenic E. coli n/a n/a 2 58 n/a Cyclospora cayatensis 0 60 1 59 Negative Entamoeba histolytica 0 60 0 60 n/a Adenovirus F40/41 n/a n/a 0 60 n/a Astrovirus n/a n/a 0 60 n/a Norovirus GI/GII n/a n/a 1 59 positive Sapovirus (I,II,IV, and V) n/a n/a 0 60 n/a *Two specimens were positive on the Cepheid GeneXpert C. difficile assay; one specimen was negative on confirmation, but melt curve analysis by BioFire technical support indicated that the melt curve was very low and may have exceeded the limit of detection/sensitivity of other assays; one specimen was unresolved because additional testing was not available.     42  The comparison for accuracy of physician ordering practices included 59 cases, as one specimen was a seeded STEC specimen and could not be included here. This comparison revealed only 2 cases in which a detected pathogen was not covered by the orders placed. One was the case in which the Campylobacter species detected by the FilmArray could not be confirmed by any other means. The second was positive for Norovirus, in addition to Campylobacter and Cryptosporidium. The Campylobacter was detected, the Cryptosporidium was negative on EIA and ova and parasite, and the Norovirus was not ordered. In the remaining ent. The comparison did, however, reveal cases in which the orders were redundant: thirteen cases included orders for both a Cryptosporidium/Giardia screen and an ova and parasite. Additionally, 48 of the 59 orders included an order for C. difficile testing along with other testing. Objective Two:  The estimated test costs to the laboratory are summarized in Table 5, while the estimated costs to a patient are summarized in Table 6. Test cost is a complex issue. It is clear from the comparison that, for instance, a simple, negative stool culture order would be less expensive to the lab than running any of the molecular assays. However, a positive stool culture, for which, as an example, two gram negative identifications were performed because there was a lactose negative colony on MAC and a HS positive colony on HE, and then a sensitivity and Salmonella latex were performed because the identification was Salmonella, would cost the laboratory $49.10. The BD Max panel, at $32.96 is less expensive, at its base price. Of the 57 orders placed for stool culture, 31 (54%) were for an expanded stool culture. Even if the tests are negative, the cost adds up: expanded stool culture plus Cryptosporidium/Giardia Screen is $50.01. Add C. difficile testing, as 48 of the providers did, and the cost increases to $63.35, or, if   43  the DNA confirmation is required, $118.51. Based on the vastly different order sets possible and the different complements of analytes on the multiplex panels, some order sets would be less expensive on multiplex assay and others with traditional assay. Because not all traditional algorithms are less expensive to the laboratory than molecular algorithms, Hypothesis Two must be rejected. Table 5: Estimated test costs to the laboratory based on reagent costs for current in place tests and quoted prices for multiplex molecular assays. Labor cost includes minutes per test based on historical study data from Hospital A and published hands on time from multiplex manufacturers and the median hourly wage for Medical and Clinical Laboratory Technologists: $28.5766. Assay Reagent Cost Labor Cost Labor+ Reagent Cost Stool Culture $8.77 $9.60 $18.37 Expanded Stool Culture with Shiga Toxin (Meridian Bioscience ImmunoCard STAT! EHEC) $23.93 $12.00 $35.93 Gram negative Identification (Vitek 2) $3.95 each $0.96 $4.91 Gram negative Sensitivity (Vitek 2) $6.10 $0.96 $7.06 Salmonella Latex (Wellcolex Color Salmonella Rapid Latex Agglutination Assay by Thermo Scientific) $11.45 $2.40 $13.85 Shigella Latex (Wellcolex Color Shigella Rapid Latex Agglutination Assay by Thermo Scientific) $12.14 $2.40 $14.54 Cryptosporidium/Giardia Screen (Alere QuikChek) $11.68 $2.40 $14.08 Rotavirus (Meridian Bioscience ImmunoCard STAT! Rotavirus) $7.14 $2.40 $9.54 C. difficile Rapid Screen (Alere QuikChek Complete C. difficile) $10.94 $2.40 $13.34 C. difficile by DNA (Meridian Bioscience Illumigene C. difficile) $50.36 $4.80 55.16 BioFire FilmArray GI $155.00 $0.96 $155.96 Nanosphere Verigene $72.25 $2.40 $74.65 BD Max Bacterial: $32.00 Parasite: $32.00 $0.96 each panel $32.96 each panel Luminex Data not available Data not available Data not available   44  Test costs to the patient are equally as complex as those to the laboratory. The base charge for a stool culture is less expensive than a multiplex assay, but the charges increase the more tests that are ordered. An expanded stool culture, Shiga toxin, and Cryptosporidium and Giardia screen would result in a charge to insurance of $317.98 if the culture was negative and at least $400.18 if the culture was positive for Salmonella or Shigella. The cash charge for the same order would be $190.79 for a negative culture and at least $240.11 for a culture positive for Salmonella or Shigella. Add the C. difficile testing the majority of physicians ordered, and the insurance charge for a negative culture goes up to a minimum of $450.39 while the cash charge to a patient goes up to $270.23. If an ova and parasite is considered, or if the C. difficile rapid screen reflexes to a DNA test, or if multiple identifications are performed on the expanded stool culture, the traditional tests could cost more than even the most expensive of the multiplex assays. Because not all traditional algorithms are less expensive to the patient than the multiplex algorithms, hypothesis three must be rejected.              45  Table 6: Estimated test costs to the patient based on average charge to insurance, average CMS reimbursement rates, and the Hospital A cash charge to an uninsured patient (60% of the charge to insurance) Assay CPT Code Average Charge to Insurance Discounted Charge to Uninsured Patients CMS Reimbursement Stool Culture (negative) 87045 $28.47 $17.08 $11.00 Expanded Stool Culture  87046 x2 $122.18 $73.31 $25.72 Shiga Toxin (Meridian Bioscience ImmunoCard STAT! EHEC) 87427 $96.12 $57.67  $16.33 Gram negative Identification (Vitek 2) 87077 $28.47 $17.08 $11.00 Gram negative Sensitivity (Vitek 2) 87186 $53.73 $32.24 $11.78 Salmonella Latex (Wellcolex Color Salmonella Rapid Latex Agglutination Assay by Thermo Scientific) 87147 $28.44 per antiserum $17.06 per antiserum $7.05 per antiserum Shigella Latex (Wellcolex Color Shigella Rapid Latex Agglutination Assay by Thermo Scientific) 87147 $28.44 per antiserum $17.06 per antiserum $7.05 per antiserum Cryptosporidium/Giardia Screen (Alere QuikChek) 87328 $25.41 $15.25  $16.33  87329 $74.27 $44.56 $16.33 Rotavirus (Meridian Bioscience ImmunoCard STAT! Rotavirus) 87425 $191.91 $115.15 $16.33 C. difficile Rapid Screen (Alere QuikChek Complete C. difficile) 87324 $104.14 $62.48 $16.33  87449 $28.27 $16.96 $16.33 C. difficile by DNA (Meridian Bioscience Illumigene C. difficile) 87493 $59.49 $35.69 $47.80 Ova and Parasite (send out) 87209 $90.79 $54.74 $24.49  87177 $43.35 $26.01 $12.12  87207 x2 $125.10 $75.06 $16.32 Norovirus (send out PCR) 87798 $70.71 $42.43 $47.80 Adenovirus (send out PCR) 87798 $70.71 $42.43 $47.80 Astrovirus (send out PCR) 87798 $70.71 $42.43 $47.80 Sapovirus (send out PCR) 87798 $70.71 $42.43 $47.80 BioFire FilmArray GI 87507 $567.75 $340.65 $567.75 Nanosphere Verigene 87506 $290.74 $174.44 $290.74 BD Max 87505 $174.76 $104.86 $174.76  87506 $290.74 $174.44 $290.74 Luminex 87507 $567.75 $340.65 $567.75    46   The average time to actionable result for the 59 applicable specimens compared in this study is displayed in Table 7. The total number of each traditional test ordered is as follows: 27 stool cultures, 30 expanded stool cultures with Shiga toxin, 23 Cryptosporidium/Giardia screens, 6 Rotavirus, 22 ova and parasite, and 48 C. difficile rapid screens. The multiplex molecular assays, if run on only one shift per day, would have a maximum time to actionable result of 24 hours if run 7 days per week. The minimum time to actionable result would be the turn-around-time of the assay: 1 hour and 10 minutes for the FilmArray, 3 hours for the BD Max, 5 hours for the Luminex, and 2 hours for Nanosphere.   Table 7: Average times in hours to actionable result for traditional assays run at Hospital A  Stool Culture and Expanded Stool Culture Shiga Toxin by EIA (Meridian Bioscience ImmunoCard STAT! EHEC) Crypto/ Giardia Screen by EIA (Alere QuikChek Cryptosporidium/ Giardia Rotavirus by EIA (Meridian Bioscience ImmunoCard STAT! Rotavirus) Ova and parasite (send out test) C. difficile Rapid Screen by EIA with DNA confirm-ation if indicated Total Mean 62.2 hours 28.1 hours 17 hours 13.3 hours 8 days, 18.25 hours 4.6 hours Total Median 50.5 hours 24 hours 15 hours 16.2 hours 8 days, 9.75 hours 1.5 hours Mean for Positive 67.6 hours n/a 8.5 hours 16 hours 9 days, 22.75 hours 1.9 hours  Median for Positive 51.8 hours n/a 8.5 hours 16 hours 9 days, 22.75 hours 1.6 hours Mean for Negative 53.7 hours 28.1 hours 17.3 hours 12.8 hours 8 days, 17 hours 4.8 hours Median for Negative 50.5 hours 24 hours 15.8 hours 16.5 hours 8 days, 8 hours 1.5 hours C. difficile Rapid Screen (Alere QuikChek Complete C. difficile); DNA confirmation (Illumigene C. difficile by Meridian Bioscience)   47  Objective Three: The statistical comparison of the traditional methods with the BioFire FilmArray is summarized in Table 8, and the comparison of the published performance characteristics of the multiplex assays is compared in Table 9. The sensitivity of the BioFire FilmArray and Traditional methods were the same in the comparison study, so Hypothesis One must be rejected. Table 8: Statistical comparison of BioFire FilmArray and traditional methods based on a 60 specimen comparison study performed at Hospital A.  Sensitivity and Specificity Traditional Sensitivity and Specificity FilmArray t-test Probability <=0.05? Salmonella Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal Shigella Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal Campylobacter Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 98% 0.32 No Y. enterocolitica Sensitivity: n/a Specificity: 100% Sensitivity: n/a Specificity: 100% Means are equal Means are equal STEC Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal P. shigelloides Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal Cryptosporidium Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 98% 0.32 No Giardia Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal Rotavirus Sensitivity: 100% Specificity: 100% Sensitivity: 100% Specificity: 100% Means are equal Means are equal                 48  Table 9: Comparison of published performance characteristics for multiplex molecular assays  FilmArray67 BDMax68 69 Verigene70 Luminex71  Cary-Blair Cary-Blair/ 10% formalin Unpreserved Cary-Blair No data available Salmonella Sensitivity: 100% Specificity: 99.6% Sens: 85%  Spec: 99.1% Sens: 91.7%  Spec: 98.9% Sens: 86.4%  Spec: 99.4% Sens: 100% Spec: 98.4% Shigella Sensitivity: 95.9% Specificity: 99.9% Sens: 100%  Spec: 99.7% Sens: 100%  Spec: 99.4% Sens: 66.7%  Spec: 98.8% Sens: 100% Spec: 98.5%  Campylobacter Sensitivity: 97.1% Specificity: 98.4% Sens: 96.2% Spec: 98.7% Sens: 100%  Spec: 97.5% Sens: 90.9%  Spec: 98.7% Sens: 100% Spec: 98.2% Y. enterocolitica Sensitivity: 100% Specificity: 100% n/a n/a Sens: 100%  Spec: 100% n/a  STEC/ Shiga toxins Sensitivity: 100% Specificity: 99.7% Sens: 75%  Spec: 99.3% Sens: 100%  Spec: 99% Sens: 100%  Spec: 99.7% Sens:  100% Spec: 98.6% P. shigelloides Sensitivity: 100% Specificity: 99.0% n/a n/a n/a n/a Cryptosporidium Sensitivity: 100% Specificity: 99.6% Sens: 90.3%  Spec: 99.8% Sens: 100%  Spec: 99.5% n/a Sens: 92.3% Spec: 95.5% Giardia Sensitivity: 100% Specificity: 99.5% Sens: 95.5%  Spec: 99.7% Sens: 94.4%  Spec: 100% n/a Sens: 100% Spec: 96.7% Rotavirus Sensitivity: 100% Specificity: 99.2% n/a n/a Sens: 98%  Spec: 100% Sens: 100% Spec: 99.8%            49  DISCUSSION  The limitations of this study include the sample size: the sample was too small to generate enough statistical power for a true comparison of means or computation of valid sensitivity and specificity for the analytes of interest. The gold standard for comparison in any molecular study should be another FDA-approved PCR, so the use of culture, ova and parasite, and enzyme immunoassay in this study may have introduced some bias to the study when determining true positives and negatives. Finally, using prospective samples in such a small study increased the chances of the low positivity that contributed to low statistical power.   There were several limitations to this study. It is difficult to assemble a set of at least 20 positives for each of 20 analytes, even in a retrospective study. The number of positive specimens can be boosted by seeding specimens with multiple analytes, but it is not the true nature of the specimen. Finally, if a laboratory does not currently perform testing for some of the analytes included, it may be difficult and expensive to locate a reference laboratory to confirm discordant results. When analytes are particularly labile, such as Campylobacter species in common transport media72, transporting the specimens to another laboratory for confirmatory testing could contribute to incorrect confirmatory results.   The transition to a multiplex molecular algorithm for any infectious disease testing creates a complicated change in the diagnosis and monitoring of infectious disease for public health laboratories, but the transition to gastrointestinal pathogen detection by molecular assays may create the most profound complications. Traditionally, pathogens that have important epidemiologic features are forwarded to public health laboratories for further testing, either required, universal submission, in the cases of Salmonella and Shigella, or when requested in the investigation of a possible outbreak. If clinical laboratories identify Salmonella and Shigella   50  infections via molecular assay, they will either have to culture the specimens retroactively in order to send an isolate to the public health laboratory or will cease to isolate organisms and simply forward the specimen to the public health laboratory. A shift that requires public health laboratories to isolate organisms themselves, in addition to the serotyping or gene sequencing performed to characterize the organisms, could be prohibitively expensive. However, the clinical laboratory budget will no longer include the isolation of these pathogens by culture, so an additional financial burden would fall to them if the public health laboratories are unable to isolate the organisms themselves.   Regardless of which laboratory takes responsibility for isolating epidemiologically important pathogens for further testing, there is an additional complicating factor: detection of non-viable organisms. The gene sequences detected by molecular assays do not indicate whether the organism from which they were isolated was viable or non-viable at the time of detection. It will be impossible for any laboratory to isolate a pathogen in culture if it was non-viable in the sample. Likewise, because the assay cannot distinguish the viability of an organism, a physician may receive a result indicating the presence of an organism that was not viable and may decide on a course of treatment based on the assumption that the organism was viable. Additionally, disease control efforts often require a negative culture result. For example, food service workers infected with Salmonella must have a negative test before they can return to work. Residual Salmonella DNA in the specimen will generate a positive result and keep employees away from work unnecessarily.  Molecular detection of C. difficile is another diagnostic challenge. In the case of C. difficile, the molecular target is a gene that causes the production of toxin A, toxin B, or both toxins A and B, which are responsible for the pathogenesis of C. difficile disease. Without active   51  toxin production, C. difficile does not cause the disease state. Molecular testing can detect the gene for toxin production even when toxin is not being actively produced and lead to inappropriate treatment with an antimicrobial agent, which may in itself induce toxin production. The indistinct detection of a target genetic sequence can confound more than the diagnosis of C. difficile disease. Some organisms, in low numbers and in the presence of abundant normal microbiota, may not cause disease: Aeromonas, for instance. The microbiome of some people include potential pathogens like C. difficile. Detection of this microbe is incidental to the pathology occurring in the patient. The inability to distinguish activity of a genetic sequence, only its presence, as well as the inability to determine the burden of an organism would represent a shift in the current understanding of gastroenteritis testing for laboratorians and providers.  A multiplex result would provide a great deal more information than clinicians are accustomed to receiving, and the prompt changes to reporting parameters that would require physician education. Reports for the familiar pathogens, like Salmonella, Shigella, and Campylobacter, would essentially remain unchanged. However, there may be pathogens included in the report that are novel to the clinicians. For instance, the BioFire FilmArray includes enteropathogenic E. coli, enteroaggregative E.coli, Norovirus, Sapovirus, Astrovirus, and other targets that clinicians may never have ordered and for which they are unfamiliar with the treatment and public health reporting requirements. Introducing testing for these unfamiliar pathogens will require an extensive education initiative. Laboratories will also need to determine how to build orders and reports in the laboratory information system for the multiplex assays in use. A report with 20 analytes on it, all of them requiring qualifying methodology comments or explanatory comments for results, will be lengthy and complicated. All clinicians will need to be instructed on its reading and interpretation. Additionally, for a large panel like the FilmArray GI   52  panel, all results will need to be reported in order to charge for the full panel. However, if a physician only requires bacterial testing, a lesser CPT code will be charged based on the number of targets covered by the physician orders. If results are hidden to create a smaller, more targeted panel, the profit margin for the test will decrease.   Finally, the issue of pathogens that are not included in the multiplex panel would need to be addressed in the transition to a multiplex assay. The FilmArray GI panel is extensive, but it does not include Aeromonas, notably, among the bacterial pathogens. A laboratory making the transition to this assay would need to consider whether it would still culture to detect Aeromonas or cease to identify it altogether. Continuing to culture for some organisms would decrease the profit margin of the multiplex molecular test. Aeromonas, although a recognized pathogen, is of questionable relevance, but some of the multiplex assays exclude more relevant pathogens, such as Yersinia enterocolitica or Plesiomonas shigelloides. Giardia and Cryptosporidium are the most common parasitic causes of gastroenteritis in the United States, so a negative result for those and high index of suspicion of parasitic infection could prompt or even reflex an order for ova and parasite.   Suggestions for further research include a full comparison or validation with retrospective samples representing a full complement of positivity for all analytes on the panel of interest. A side-by-side comparison of multiple multiplex assays would be of value. A large-scale study of the implications of detecting and reporting non-viable organisms would clarify the importance of that issue in the care of patients. Additionally, a study exploring the financial impact of a shift to multiplex molecular detection of gastrointestinal pathogens on public health laboratories would be relevant.    53             REFERENCES                54  REFERENCES  1 Struewing JP, Hartge P, Wacholder S, et al. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med. 1997; 336: 1401-1408. doi: 10.1056/NEJM199705153362001  2 Tang Y, Procop GW, Persing DH. Molecular diagnostics of infectious diseases. Clin Chem. 1997; 43(11):2021-2038. http://www.clinchem.org/content/43/11/2021.full. Accessed January 24, 2016.  3 Moroni P, Pisoni G, Cremonesi P, Castiglioni B. Staphylococcus. In: Liu D, ed. Molecular Detection of Human Bacterial Pathogens. Boca Raton, FL: CRC Press; 2011: 307-322. https://books.google.com/books? id=nnGhc44bypAC&printsec=frontcover#v=onepage&q&f=false. Accessed January 24, 2016.  4 Redefining active MRSA management: Xpert MRSA. Cepheid website. http://www.cepheid.com/us/search?searchword=mrsa. Updated 2015. Accessed January 24, 2016.  5 Gase KA, Reese SM. Journal club: Impact of a rapid screening for discontinuation of methicillin-resistant Staphylococcus aureus contact precautions. Am J Inf Cont. 2016. Published online. doi: 10.1016/j.ajic.2015.11.011. Accessed January 24, 2016.  6 Understanding value-based purchasing. Lake Superior Quality Innovation Network website. http://www.lsqin.org. Updated October 2015. Accessed October 30, 2015.  7Diarrhoeal disease fact sheet. World Health Organization website.  http://www.who.int/mediacentre/factsheets/ fs330/en/. Updated April 2013. Accessed April 18, 2015. 8 Diarrhoeal disease fact sheet. World Health Organization website. http://www.who.int/mediacentre/factsheets/ fs330/en/. Updated April 2013. Accessed April 18, 2015.  9 Pang X, Hodinka RL. Gastroenteritis viruses. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 1617-1632.  10 Humphries RM, Linscott AJ. Laboratory diagnosis of bacterial gastroenteritis. Clin Microbiol Rev. 2015; 28. doi: 10.1128/CMR.00073-14. 11 Deaths by detailed cause and sex, Michigan residents, 2013. Michigan Department of Health and Human Services website. http://www.mdch.state.mi.us/pha/osr/deaths/DXCause113.asp. Accessed January 24, 2016.     55   12 Gastroenteritis. U.S. National Library of Medicine Medline Plus website. Published July 7, 2014. Updated August 25, 2015. Accessed September 15, 2015.  13 Buss SN, Leber A, Cahpin K, et al. Multicenter evaluation of the BioFire FilmArray gastrointestinal panel for etiologic diagnosis of infectious gastroenteritis. J. Clin. Microbiol. 2015; 15(3): 915-925. doi: 10.1128/JCM.02674-14.  14 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 685-713.  15 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 697.  16 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 685-713.  17 Fitzgerald C, Nachamkin I. Campylobacter and Arcobacter. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 998-1012.  18 Horneman AJ. Aeromonas. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 752-761.  19 Forsythe SJ, Abbott SL, Pitout J. Klebsiella, Enterobacter, Citrobacter,Cronobacter, Serratia, Plesiomonas, and other Enterobacteriaceae. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 714-737.  20 Petersen JM, Gladney LM, Schriefer ME. Yersinia. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 738-751.  21 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 685-713.  22 Tarr CL, Bopp CA, Farmer JJ. Vibrio and Related Organisms. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 1. 11th ed. Washington, D.C.: ASM Press; 2015: 762-772.  23 Stevens DL, Bryant AE, Carroll K. Clostridium. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 1. 11th ed. Washington, D.C.: ASM Press; 2015: 940-966.      56   24 Polage CR, Gyorke CE, Kennedy MA, et. al. Overdiagnosis of Clostridium difficile infection in the molecular test era. JAMA Intern Med. 2015; 175(11): 1792-1801. doi:10.1001/jamainternmed.2015.4114.  25 Pang X, Hodinka RL. Gastroenteritis Viruses. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 1617-1632.   26 Pang X, Hodinka RL. Gastroenteritis Viruses. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 1617-1632.  27 Pang X, Hodinka RL. Gastroenteritis Viruses. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 1617-1632.  28 Pang X, Hodinka RL. Gastroenteritis Viruses. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 1617-1632.  29 Xiao L, Cama V. Cryptosporidium. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology, Vol. 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2435-2447.  30 Novak-Weekley S, Leber AL. Intestinal and Urogenital Amebae, Flagellates, and Ciliates. In: Jorgensen JH, Pfaller MA, eds.  Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2399-2424.  31 Lindsay DS, Weiss LM. Cytoisospora, Cyclospora, and Sarcocystis. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2425-2434.  32 Novak-Weekley S, Leber AL. Intestinal and Urogenital Amebae, Flagellates, and Ciliates. In: Jorgensen JH, Pfaller MA, eds.  Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2399-2424.  33 Estimates of Foodborne Illness in the United States. Centers for Disease Control and Prevention website. http://www.cdc.gov/foodborneburden/2011-foodborne-estimates.html. Updated January 8, 2014. Accessed January 31, 2016.  34 Reportable infectious diseases in Michigan, 2008-2012. Michigan Department of Community Health Bureau of Disease Control, Prevention and Epidemiology web site.   http://www.michigan.gov/documents/mdch/2008-2012_Epi_Profile_Report_457437_7.pdf. Accessed September 27, 2015.  35 Garcia LS, ed. Clinical Microbiology Procedures Manual. 3rd ed.[CD-ROM] Washington, D.C.: ASM Press; 2010.     57   36 Microbiology checklist: CAP accreditation program. College of American Pathologists website. http://www.cap.org. Updated July 28, 2015. Accessed January 30, 2016.  37 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 685-713.  38 Fitzgerald C, Nachamkin I. Campylobacter and Arcobacter. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 998-1012.  39 Horneman AJ. Aeromonas. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 752-761.  40 Forsythe SJ, Abbott SL, Pitout J. Klebsiella, Enterobacter, Citrobacter,Cronobacter, Serratia, Plesiomonas, and other Enterobacteriaceae. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 714-737.  41 Gould LH, Bopp C, Strockbine N, et al. Recommendations for diagnosis of shiga toxin-producing Escherichia coli infections by clinical laboratories. MMWR, 2009, 58(RR-12):1-14. http://europepmc.org/abstract/med/19834454. Accessed January 26, 2016.  42 Strockbine NA, Bopp CA, Fields PJ, Kaper JB, Nataro JP. Escherichia, Shigella, and Salmonella. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 685-713.  43 Petersen JM, Gladney LM, Schriefer ME. Yersinia. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol 1. 11th ed. Washington, D.C.: ASM Press; 2015: 738-751.  44 Tarr CL, Bopp CA, Farmer JJ. Vibrio and Related Organisms. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 1. 11th ed. Washington, D.C.: ASM Press; 2015: 762-772.  45 Stevens DL, Bryant AE, Carroll K. Clostridium. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 1. 11th ed. Washington, D.C.: ASM Press; 2015: 940-966.  46 Dubberke ER, Burnham CD. Diagnosis of Clostridium difficile infection: Treat the patient, not the test. JAMA Intern Med. 2015; 175(11):1801-1802.  47 Dubberke ER, Burnham CD. Diagnosis of Clostridium difficile infection: Treat the patient, not the test. JAMA Intern Med. 2015; 175(11):1801-1802.  48 Stevens DL, Bryant AE, Carroll K. Clostridium. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 1. 11th ed. Washington, D.C.: ASM Press; 2015: 940-966.    58    49 Xiao L, Cama V. Cryptosporidium. In: Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology, Vol. 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2435-2447.  50 Novak-Weekley S, Leber AL. Intestinal and Urogenital Amebae, Flagellates, and Ciliates. In: Jorgensen JH, Pfaller MA, eds.  Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2399-2424.  51 Lindsay DS, Weiss LM. Cytoisospora, Cyclospora, and Sarcocystis. In Jorgensen JH, Pfaller MA, eds. Manual of Clinical Microbiology. Vol. 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2425-2434.  52 Novak-Weekley S, Leber AL. Intestinal and Urogenital Amebae, Flagellates, and Ciliates. In: Jorgensen JH, Pfaller MA, eds.  Manual of Clinical Microbiology. Vol 2. 11th ed. Washington, D.C.: ASM Press; 2015: 2399-2424.  53 Syndromic Approach. Biofire Diagnostics, LLC website. http://www.biofiredx.com/about-us/syndromic-approach/. Accessed January 31, 2016.  54 Gray J, Coupland LJ. The increasing application of multiplex nucleic acid detection tests to the diagnosis of syndromic infections. Epidemiol. Infect. 2014; 142: 1-11. doi: 10.1017/S0950268813002367.  55 Gray J, Coupland LJ. The increasing application of multiplex nucleic acid detection tests to the diagnosis of syndromic infections. Epidemiol. Infect. 2014; 142: 1-11. doi: 10.1017/S0950268813002367.  56 FilmArray. Biofire Diagnostics, LLC web site. http://filmarray.com/the-panels/. Updated 2014. Accessed June 17, 2014. 57 BD MAX Enteric Bacterial Panel [Package Insert]. BD, GeneOhm Sciences Canada, Inc web site. http://moleculardiagnostics.bd.com/product/max/. Updated 2014. Accessed September 20, 2015.  58 BD MAX. BD Diagnostics. Sparks, MD: BD Diagnostics; April 2014.  59 BD MAX Enteric Bacterial Panel [Package Insert]. BD, GeneOhm Sciences Canada, Inc web site. http://moleculardiagnostics.bd.com/product/max/. Updated 2014. Accessed September 20, 2015.  60 Luminex xTAG Gastrointestinal Pathogen Panel (GPP). Luminex Corporation web site.  https://www.luminexcorp.com/clinical/infectious-disease/gastrointestinal-pathogen-panel/resources/ Updated 2014. Accessed September 20, 2015.       59   61 Enteric Pathogens Test. Nanosphere, Inc. web site. http://www.nanosphere.us/products/enteric-pathogens-test. Updated 2014. Accessed September 27, 2015.  62 Verigene Enteric Pathogens Nucleic Acid Test (EP) [Package Insert]. Nanosphere, Inc. web site. http://www.nanosphere.us/sites/default/files/support-docs/ep_package_insert.pdf. Updated October 2014. Accessed September 27, 2015.   63 Occupational Employment and Wages, May 2014. United States Department of Labor Bureau of Labor Statistics website. http://www.bls.gov/oes/current/oes292011.htm. Updated March 25, 2015. Accessed March 24, 2016.   64 2016 CodeMap® Reimbursement Manual for Lab and Pathology. Shaumberg, IL: Wheaton Partners, LLC; 2016.  65 Doucette LJ. Mathematics for the Clinical Laboratory 2nd ed. Maryland Heights, MO: Saunders Elsevier; 2011: 344-383.   66 Occupational Employment and Wages, May 2014. United States Department of Labor Bureau of Labor Statistics website. http://www.bls.gov/oes/current/oes292011.htm. Updated March 25, 2015. Accessed March 24, 2016.  67 BioFire Diagnostics, LLC. FilmArray gastrointestinal (GI) panel: Instruction booklet. BioFire: Salt Lake City, UT; 2015: 38.  68 GeneOhm Sciences. BD MAX enteric bacterial panel package insert. GeneOhm Sciences, Quebec, Canada; 2014: 11-15.  69 GeneOhm Sciences. BD MAX enteric parasite panel package insert. GeneOhm Sciences, Quebec, Canada; 2015: 11-14.  70 Nanosphere. Verigene enteric pathogens nucleic acid test (EP) package insert. Nanosphere, Inc: Northbrook, IL; 2016: 22-26.  71 Luminex Corp. The clinical performance and value of xTAG GPP (gastrointestinal pathogen panel). Luminex Corp: Austin, TX; 2013: 4.  72 Wang WL, Reller LB, Smallwood B, Luechtefeld NW, Blaser MJ. Evaluation of transport media for Campylobacter jejuni in human fecal specimens. J Clin Microbiol. 1983; 18(4): 803-807. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC270910/.