Veq: WHO English only Distr Limited PROBLEM-BASED TRAINING EXERCISES FOR ENVIRONMENTAL EPIDEMIOLOGY INSTRUCTOR'S GUIDE 2nd edition EEX > aes SZ —— Office of Global and Integrated Environmental Health World Health Organization Geneva 1998aCOLI PROBLEM-BASED TRAINING EXERCISES FOR ENVIRONMENTAL EPIDEMIOLOGY 2nd Edition Instructor’s Guide Editor Steven Markowitz, MD This document provides a teaching resource for instructors of environmental epidemiology: The main purpose of these teaching cases is {0 promote the! understanding and application of epidemiology in the prevention of environmental disease and the promotion of health. These teaching exercises emphasize examples from developing countries and may be useful for use with students in environmental health, epidemiology, and international health. ‘The environmental epidemiology cases included in this set are based on original studies that have been published during the last: decade. ‘The present collection emphasizes entry level epidemiological concepts and is intended to supplement ‘classroom lectures and standard epidemiology textbooks, including the WHO texts, Basic Epidemiology and Environmental Epidemiology. Geneva, 1998(© World Health Organization 1908 “This document isnot issued othe general pubic and all ights are reserved by the World Heath Organization. The document may not be reviewed, abstracted, quoted, reproduced or translated, in pat or in whole, without the prior wren permission of WHO. No partofthis document may be soredin a rtveval system or transmitted in any form or by any means - electronic, mechenicel or other without the prior written permission of WHO. “The views expressed in documents by named authors are solely the responstilty of those authors Reproduced from originals provided.TABLE OF CONTENTS FOREWORD LIST OF CONTRIBUTORS INTRODUCTION: HOW TO SELECT AND USE CASES Case 1 An outbreak of acute neurologic disease in a factory Case 2 Severe infection due to bacterial water contamination Case 3 Pesticide poisoning: an outbreak among antimalarial workers Case 4 Chronie obstructive pulmonary disease Case 5 Lead poisoning among household members Case 6 Epidemic asthma Case 7 Environmental causes of Wilms’ tumor Case 8 Bladder cancer in a chemical factory Case 9 Parathion poisoning Sierra Leone Case 10 Atmospheric fog in Greater London Case 11 Chronic arsenic toxicity Case 12 Water availability and trachoma Case 13 DDT and breast milk ii iii iv 19 27 34 41 52 58 69 86 101 107 14FOREWORD This teaching resource was originally prepared in 1992 and revised in 1997 for the Programme for Promotion of Environmental Health of the World Health Organisation in order to strengthen education, training, and research in the field of environmental health. This document aims to help address the need for practical teaching tools for adult learners with content that is relevant to problems in environmental epidemiology in developing countries. A broad definition of environment has been utilised in keeping with the recognition by WHO that teaching materials should reflect the effects that arise in the general population in the outdoor and indoor environment through air, water, food, consumer products, or soil These problem-based cases were contributed by a number of experts in environmental epidemiology and are adapted from published studies in the past decade. In some cases, slight modifications have been made in the originally reported data in order to make them suitable for teaching purposes. The original collection of exercises published in 1992 was prepared Dr. Tord Kjellstrom, Division of Environmental Health, WHO, Dr. Nancy Hicks of the USA, and Mr. Guido Torelli, Division of Environmental Health, WHO. The revised edition published in 1997 was edited by Dr. Steven Markowitz of the Mount Sinai School of Medicine, New York, USA. A draft was reviewed by the participants in the Consultation on Environmental Epidemiology Training, Geneva, 2-4 June 1991 and again by a panel of experts in environmental epidemiology in 1997. tee This training material is a contribution by WHO to the International Programme on Chemical Safety. Financial assistance of the Government of Sweden in carrying out this project is gratefully acknowledged. iLIST OF CONTRIBUTORS Dr. Ruth ETZEL, Chief, Division of Environmental Hazards and Health Effects, Centers for Disease Control, Atlanta, Georgia, USA Dr. Nancy V. HICKS, Epidemiologist, USA Dr. Tord KIELLSTROM, Director, Global and Integrated Environmental Health, World Health Organisation, Geneva, Switzerland Dr. Steven MARKOWITZ, Associate Professor, Department of Community Medicine, Mount Sinai School of Medicine, New York City, USA Prof. Anthony J. McMICHAEL, Professor, London School of Public Health and Tropical Hygiene, London, England Dr. Linda ROSENSTOCK, Director, National Institute for Occupational Safety and Health, Washington D.C., USA iiiINTRODUCTION: HOW TO SELECT AND USE CASES. Epidemiology is universally recognised as an essential tool in efforts to improve public health. This notion is embodied in the WHO Declaration, Health for All by the Year 2000. In the World Health Assembly resolution in May 1988, specific improvements in the application of epidemiology in attaining health for all were recommended. Member states were urged to make greater use of epidemiological data, concepts, and methods in developing, monitoring, and evaluating health-for-all strategies. This recommendation recognised the need to improve the training of health and related professionals in epidemiological concepts and skills in most developing countries. This collection of cases in environmental epidemiology was developed to provide teachers of epidemiology with practical learning tools of special relevance to environmental health challenges in developing countries. We believe that beginning students of epidemiology and environmental health in any country may find these cases useful. The cases span a broad variety of exposures, diseases, and geographic regions. They address diverse study designs, epidemiologic concepts, and learning objectives. These cases are based on real studies that have been published in the peer-reviewed English language journals during the past decade. These problem-based cases are intended to be used in conjunction with lectures and textbooks in epidemiology. They are intended to be used by students in small groups as a means of applying and integrating newly acquired concepts in epidemiology. While students need to have a general understanding of health in order to participate in case discussion, we have attempted to eliminate the need for a specialised background in medicine or environmental health by providing important medical and environmental facts in the cases when needed. Why do we encourage small group learning? Advances in the theory and application of adult educational methods show that students learn well in small groups that are given tasks (i.c.-cases) that require thinking and interaction. Students thereby help each other “discover” the meaning of previously introduced concepts by together struggling through their application in a set of specific circumstances. Instructors guide the discussions. These small groups are best served if instructors resist usurping meeting time by giving mini-lectures or otherwise readily providing answers that cut short student discussion, A simple and useful technique is to reflect questions addressed to the instructor back to other students in the class, thereby transferring the role of “expert,” at least for the moment, Each case begins with a set of leaming objectives, These should be useful to the instructors in selecting which cases to use, especially in relation to timing the use of cases in relation to lectures. They should also be useful to students by informing them of what they should expect to leam by the end of the case. wvCases are divided into parts. Students are given one part at a time for discussion. Subject to availability of time, instructors may wish to allow students to pace themselves by allowing them to express their readiness to begin on the next part of the case. It is best if one or more students acts as the group recorder if needed. It is useful to encourage that a student write answers on a blackboard or poster paper, if available. In order to guide the selection of cases, we provide case profiles in the following tables. In Table 1, the cases are indexed by exposure, route of exposure, and occupational vs. environmental setting. Table 2 provides a profile of the cases by learning objectives. While these summaries cannot replace perusal of the cases, they may expedite choice of 1 or more cases that best meet the needs of the course. These cases are best used in groups of 6 to 10 students in the presence of 1 instructor or discussion guide. If a larger number of students are present without sufficient discussion guides to allow a 10:1 ratio, then it recommended to divide the large student group into smaller groups of 6 to 10 students. ‘The small groups can record their answers to the case questions, select a reporter, and then report back their answers in a large group session after the small groups have completed each part of the case or after completion of the case. The instructor should bear in mind that the successive parts of the cases often provide information that answer questions from previous parts of the cases. Hence, waiting until small groups complete the case prior to the large group report-back session may subvert some of the challenges posed by the individual parts of the case. In the Instructor's Guide version of these cases, each exercise is accompanied by possible answers to the questions posed. These answers can be used a checklist by the instructors to ensure that all important issues are raised by the students. These answers are intended as a suggested floor for discussion, not a fixed ceiling. Students should be urged to consult with epidemiology textbooks to seek fuller explanations of issues raised by the cases. Having such a textbook available during case discussions might be useful as well. Case users are urged to give us feedback on the strengths and weaknesses of the cases, any errors inadvertently included, or suggestions for additional cases. We would like over time to continue to add cases that address new learning objectives or involve exposures and diseases that are not included in the current collection. We especially welcome members of the WHO Global Environmental Epidemiology Network (GEENET) to submit exercises from their own experience that may be reviewed for possible inclusion in the next editionTable 1. Structured Index of Cases By Topic Topie Case Number Agents Infectious Agents Botulism 1 Salmonella 2 Trachoma 12, Pesticides Organophosphates 39 Chlorinated Hydrocarbons 13 Non-specified # 8 Metals : Lead 6 Arsenic, B ‘Asbestos 8 Air Pollution Ambient 10 Indoor 4 Organic Chemicals Soy Protein 6 Ortho-toluidine (aromatic amines) hit Roules Of Exposure Food/ingestion 1,5,9,13, Water 2,11,12 Skin : 3,8 Air 4,6,7,8,10 Occupational vs. Environmental Occupational 3,78 Para-occupational (direct connections to specific 5,6,13 occupational/industrial activity) Environmental 1,2,4,8,9,10,11,12 viTABLE 2. ORGANISATION OF CASES BY LEARNING OBJECTIVES 5 CASE NUMBERS TT273 74] 5] 6] 7/8 | 9] a) 2B |Objectives/Topics 1 Ientify study design |X [X [|X |X xX 2, Strengths and xt xy, x x ‘weakness of study design 3. Measures of disease [ X |X ~ x x x|x occurrence. E Measures of effect. [|X | X x ¥/ XTX] ¥ 5. Acute disease ¥ ¥ ¥ [x] x outbreak investigation i 6. p-Value, confidence * x X/X]X[xX intervals 7. Concept of causality ¥ x | 8. Confounding and x [x effect modification | | 9. Role of exposure x x | Tx assessment | 10. Epidemiology in Tx ¥ t health policy Key to Case = Neurologic Epidemic 8 = Wilms’ Tumour Numbers 2= Bacterial Water Contamination 3 = Pesticide Poisoning/Malaria 4 = Chronic Arsenic Pulmonary Disease 5 = Lead Poisoning/Batteries 6 =Epidemic Asthma 7 = Bladder Cancer 9 = Pesticides in Sierra Leone 10 = Smog in London 11 = Chronic Arsenic Toxicity 12 = Water/Trachoma 13 = DDT and Breast Milk viiCASE I - ACUTE NEUROLOGIC EPIDEMIC IN A FACTORY Nancy Hicks, PhD and Steven Markowitz, MD First prepared April, 1991 (Hicks); revised January 1997 (Markowitz) Learning objectives 1. Identify and apply steps in investigation of outbreak of acute disease. 2. Calculate and interpret attack rate, case fatality rate, and relative risk. 3. Identify public health measures to take after an outbreak investigation. PART 1 - INTRODUCTION You are a public health officer on a heavily populated Pacific island. You receive a call on October 10 from an occupational physician who describes an outbreak of “unusual illness that began during the last week of September among workers at-a printing factory. “She tells you that a number of workers became ill within a2 to 3 day period. ‘Their symptoms included double vision, drooping eyelids, muscle weakness, and progressive respiratory weakness. Two workers have been hospitalised and have required mechanical ventilation to assist in breathing, Question 1 What questions do you ask the doctor (use the attached Table 1)? 1. Number of people affected 2, Demographic data, including age, sex, race, employment status 3. Potential exposures at the factory, including agents used and job titles and duties 4. Recent changes at the factory Question 2 What would you do first? You would probably call the physicians who are treating the cases in order to get the full account of the illness and the provisional diagnoses.Acute Neurologic Epidemic 1, 1. References: Table 1 Steps in an Outbreak Investigation ‘a, Gather information on index case(s) Determine likelihood that epidemic exists Verify the diagnosis of reported cases. TL ‘Descriptive Phase Establish case definition, including person, time, and place Develop possible etiologic hypotheses, considering potential exposures Define population at risk (include parameters of time and place) Look for additional cases: establish surveillance Measure excess disease occurrence TI Analytic Phase Develop specific hypotheses Choose proper study design Plan, execute and analyse study (inchiding investigation of source with assistance of laboratory) IV. Control Phase Institute control/preventive measures a) Investigating Environmental Disease Outbreaks, A Training Manual. WHO/PEP, 1991 b) Centres for Disease Control. Guidelines for Investigating Clusters of Health Events. MMWR, 39: (RR-11), 1990Acute Neurologic Disorder PART 2 ‘You are told that there are 4'male and 3 female cases, ranging in age from 18 to 66 years. All are Asian and are native to the island. You'are told that two of:the men operate printing presses, filling them with ink and using solvents for occasional cleaning. One additional male worker is a typesetter, who is also exposed to solvents. The fourth affected male employee is a maintenance worker, who repairs the presses and uses solvents to clean the machines on a daily basis. The three ill female workers are proof-reader, an office worker, and the cafeteria cook. None of the women are exposed to chemicals in the plant. The cook was the first to become ill and was one of the hospitalised cases, You are'told that there are 40 employees in total. ‘The plant manager gives you a personne! list that includes names, age, sex, job title, work area, job duties, and. job. duration for each worker. Question3 What are your initial hypotheses about the cause of the illness? Although solvents are used at the printing plant and are known to cause neurological illness, an occupational agent seems unlikely to be responsible for the acute outbreak of disease due to the fact that three workers who are unexposed are affected by the illness, Secondly, the fact that the cook was the first 10 become ill and was severely affected suggests a food-borne toxin might be the cause of the illness. Question 4 | What information do you want next, and whom do you call? You call the physicians who are treating the cases. You still need a full description of the illness and need to develop a case definition.Acute Neurologic Epidemic PART3 Your first calls are to the physicians who are treating the two hospitalised workers. They tell you that the presumptive diagnosis in both cases is botulism. Laboratory confirmation is pending, but, meanwhile, both patients are being treated for botulism. ‘You reftesh your memory about botulism (see Table 2), since this type of poisoning is rare on your island. Question — What is a working case definition that you would use for your investigation? A case definition can include symptoms, physical signs and/or laboratory findings. A trade off between using a narrow versus broad case definition for epidemiological purposes should be discussed, including the impact on the rate of {falsefpositives and rate of false/negative cases. In this epidemic, including laboratory confirmation of botulism would be excessively restrictive, given the fact that only a minority of cases would have positive serum, stools or cultures when the epidemiological investigation was begun, that is, more than ten days after onset of illness in the first cases. A reasonable case definition was that used in the original published report: any person at the factory who had two or more of the symptoms listed on Table 2: ptosis, dysphagia, diplopia, dysarthia, and muscle weakness, Question 6 What information would you want to collect next, and how would you obtain it? There are at least two types of information that you would want to collect, including: @) Identification of all cases in the factory. This should be done in the form of a questionnaire 10 all 40 employees, inquiring about symptoms and food eaten in the factory and at home. 5) You will need information about the preparation and serving of the food at the factory during the past several weeks. This would be obtained by interviewing the cafeteria workers and inspecting the cafeteria, Samples of any foods that might had been used over the past several weeks and remain should be taken. If samples are not available, detailed information about food items served over the past several weeks, including product name, manufacturer, lot number, etc. should be collected.Acute Neurologic Disorder Table2 Seminal Facts about Botulism, Type A ‘Characteristic symptoms: * drooping eyelids (tosis), _- (develop 12 to 36 hours « difficulty swallowing (dysphagia) after exposure) ‘ double vision (diplopia) * difficulty speaking (dysarthia) ‘© muscle weakness, including respiratory muscles Cause: « neurotoxin produced by the bacterium Clostridia " botulism Diagnosis: « characteristic symptoms and physical exam findings , ‘* laboratory confirmation * epidemiologic linkage Laboratory confirmation: ‘* serum and stool levels of toxin: 60% 10 65% are positive if collected within 3 days of ingestion but only: 15% to 20% are positive if collected 4 or more days after ingestion # stool culture of organism: 75% are positive if collected within'3 days of ingestion » but only 35% are positive if collected 4 or more days after ingestion ‘Treatment: : +» supportive therapy, ic. mechani¢al ventilation if needed ‘e administration of anti-toxin | References: 1) Bleck, T. Clostridium Botulinum in Mandell, G, Bennett, J, and Dolin, R (eds) Principles and Practice of Infectious Diseases, 4th edition, Churchill Livingstone, New York, 1995 2) Woodruff B et al. Clinical and Laboratory Comparison of Botulism from Toxin ‘Types A. Band E inthe United States, 1975 - 1988. Jl Inf. Dis, 1992; 166: 1218-1286.Acute Neurologic Epidemic PART 4 For this outbreak, you define-a case of botulism as anyone with two or more of the symptoms on the list in the box, Seminal Facts on Botulism. All7 workers who were initially identified by the plant's occupational physician as recently ill met this definition. Results of the questionnaire were obtained for 39 of the 40 workers at the plant. Six ill employees ate at the factory cafeteria on September 26 and 27, compared to 7 of 32 not-stricken workers who ate at the factory cafeteria on both days. You also leam that the factory cook died from respiratory complications 3 weeks after hospital admissions. Question 7 In the preliminary analysis of results that you will provide to the Health Commissioner, you set up a2 X 2 table, relating exposure and disease. You calculate the attack rate, case-fatality rate, and the relative risk of getting botulism after eating at the factory cafeteria on September 26 and 27. Construct a2 x 2 table using these data Mt Not Ill Ate in Cafeteria 6 7 13 Did not eat in cafeteria ie 25 26 a 32 “In the original data, this value was zero. It has been changed to 1.0 for teaching purposes. Calculate the attack rate Number ill = 6 = 046 = 46% 13 Number who ate in cafeteria Calculate the case fatality rate for the factory workers: deaths = 1 = 0.143 = 14.3% Hofcases 7 the table shown in Question 9, =RR= 6x25) = 150 = 21.4 (95% Confidence Interval = 1.9- 561.9) 1x7 7 (<0.003)Acute Neurologic Disorder PART5 | Meanwhile, additional information was obtained: Following announcements to the medical community and through’ mass.media, 14 additional suspect cases of botulism were reported. Only two of these-cases met the case definition of the study. One was a.68 year old woman in the community (not-a:plant worker) who died at home within 24 hours of becoming ill. The second was her 6 year-old grandson, who was hospitalised and requited mechanical ventilation. Serum specimens from 7 cases and stool samples from 2 cases, all from the plant, were negative for botulism toxin, These were collected 3 weeks after exposure. Laboratory examination of preserved foods remaining in the factory cafeteria showed that 1 of 4 unopened jars of unsalted | peanuts produced by Company B on September 10 was positive for type A botulism toxin. Question 8 How can you explain the occurrence of botulism in the community? Since botulism is rare on this island, it is likely that the community cases were due 10 the same source of contaminated foods as the cases in the factory. Ifso, the food that was contaminated with type A botulinum was present not just in the lant cafeteria but also in the community. Question 9 How do you interpret the results of the serum and stool tests? As evident from Table 2, the fact that the biological samples were taken three weeks afier exposure, does not rule out presence of botulism, since only the minority of samples will be positive if taken after three days following ingestion of the toxin. Question 10 What steps do you take to protect the health of the public? a) Visit the homes of the community cases 10 collect open jars of unsalted peanuts. 4) Visit plant that manufactured implicated peanuts to confiscate additional product, inspect and identify deficient process of production, and shut down ‘factory until deficiency can be corrected. ©) Issue a mass media alert to stores that might carry the unsalted peanut product implicated in outbreak and confiscate potentially affected jars. 4) Issue alert 10 medical community to identify additional cases of botulism and report these to public health authorities. €) Issue product recall and test all batches for type A botulinum toxin.Acute Neurologic Epidemic PART 6 - EPILOGUE Investigation of Company B by the Department of Health revealed that it was a small family- ‘owned business that was not licensed to can products. They did not have the proper steam equipment to process low acid canned foods. The owner reported that some of the glass jars filled with peanuts were noted to have developed bubbles, or other signs of spoilage, and were then discarded. Company B was closed by the Department of Health, A product recall was issued for all jars of peanuts produced by Company B on September 10. 104 jars were obtained. Laboratory testing showed that 34 of the 104 jars tested positive for type ‘A botulinum toxin, compared to 0 to 32 jars randomly selected from other batches produced by Company B. Despite the recall, the 68 year old woman and her. grandson who became ill had purchased Company B peanuts of the September 10 batch at a local grocery store. After they became ill, testing of the peanuts from the partially eaten jar was positive for type A botulinum toxin. Physician awareness of the dangers of botulism on the island improved as a result of this outbreak. Within four months of this outbreak, two additional outbreaks of botulism as a result of ‘home-preserved meats were reported to the Department of Health. ‘The Department of Health issued new regulations requiring that all canneries register with Department of Health and show "proof of good. manufacturing practices.” The number of inspectors and the quality of their training, however, were considered to be insufficient to ensure adequate enforcement, The case was based on a report by Chou LH., Hwang, P.H. & Malison, MD. An Outbreak of uts, International i of Epidemiology 1988; 17, 899-902. This case reflects minor modifications in the original report, which were made for teaching purposes.CASE I - A STUDY OF SEVERE INFECTION FROM BACTERIAL WATER CONTAMINATION Prepared by Linda Rosenstock, MD, MPH; revised by Steven Markowitz, MD February 1990; revised January 1997 Learning Objectives Establishing a case definition; Identifying the study design; Consideration of sampling strategy; Consideration of non-participant bias; Consideration of exposure routes; Recommending changes to prevent recurrence of epidemic, PART 1 - INTRODUCTION Salmonella infections pose a particularly serious risk if they become blood-borne in the host (septicaemia), carrying a high mortality rate. In your role as a health officer, you have been asked to assist in an investigation of salmonella septicaemiia.in Kenyan children. You will be working with.a team of physicians who have been caring for some of the children with this disease at the Kenyatia National Hospital (KNH). ‘The term salmonellosis refers to infection caused by bacteria of the genus Salmonella, of which there are 3 primary species:“1) Salmonella typhi - the cause of enteric "typhoid fever", 2) ‘Salmonella. cholerasuis and 3) Salmonella, enteridis - which includes Salmonella, typhimurium - a subtype of increasing prevalence and antibiotic resistance in Kenya. ‘Salmonella septicaemia results when pathogenic species enter and multiply in the bloodstream. This may or may not be accompanied by the typical findings of gastroenteritis (abdominal. pain, vomiting, diarthea), but is of particular concern because infections may spread and mortality and morbidity (e.g. need for amputation) are high. Infections result from ingestion of food or water contaminated from human or animal sources. Animal sources include poultry, rodents, cattle, pigs, snakes, etc... Salmonellae are easily killed during cooking, The main goal of the study is to lear more about salmonella septicaemia in Kenya, particularly to understand which children are at greatest risk for getting sick, so that environmental interventions to reduce the risk can be undertaken.Bacterial Water Contamination Question 1 Question 2 Re-state the main problem in the form of hypotheses that should be investigated. Hypotheses: The risk of salmonella septicaemia is increased in association with: 1. Place of residence 2. Source of drinking water 3. Types of food and food preparation 4, Domestic and community sanitation 5. Handling of animals What kind of epidemiological study would you suggest? What are some of the potential problems that might be associated with this type of study? The two main choices are a cohort or case-control study. A cohort study would pose problems because the outcome (a case of septicaemia) is likely to be a relatively uncommon event, so a large number of children would need to be followed to get sufficient cases to study. It would be difficult to do, very costly, and time consuming. A more efficient study is a case-control design. The potential difficulties are still numerous, including: 1. Finding an appropriate control group may be difficult. 2. The determination of individual exposure relies on recall or records - which are often inaccurate, 3. There may be confounding (i.e., outcomes may have been wholly caused by some other factor that is correlated with the exposure - e,g., income) 4. The method of case detection may well not be representative of all cases (e.g. only the sickest cases may be included). 10Bacterial Water Contamination PART 2 - DEFINITION OF A CASE ‘You decide to do a’case-control study. You define a case as follo Study Case Definition: Any child admitted to the paediatric wards in Kenyatta National Hospital (KNH) and the Infectious Disease Hospital between the ages of O-and 12 years, who has a positive blood ‘culture for any salmonella bacteria between July 1 and November 30, 1986. Cases will be identified and enrolled prospectively into the study over the S-month period. You plan t0-obtain information from the parent or guardian about environmental exposures and past medical history of each case. Question 3 What are the advantages and limitations of this case definition? Advantages: © Reference to source and ages of cases as well as calendar time. «Inclusion of positive blood cultures as study criteria allows low level of false positives. A disadvantage is restriction to hospitalised cases since non-hospitalised cases may have different risk factors for salmonella infection. If so, one of the study cases would not be representative of the full spectrum of salmonella infection. List the kinds of information you would like to obtain in your interviews with the parent or guardian, a) Geographical location of home 5) Income c) Nutritional status @) Source of drinking water ¢) Types of food and food preparation methods fl Use of latrines or other method of faecal waste g) Use of domestic animals h) Past medical history ilBacterial Water Contamination PART 3 - CHOICE OF CONTROLS ‘You have a long discussion with your colleagues about the choice of a control group. A number of control groups are considered, including: a) community controls - children of the same age in the community who did not get sick, and b) hospital controls. You finally decide to use hospital controls - a set of children with the same age and sex distribution:as the cases and who were admitted consecutively to the Paediatric Emergency Ward, KNH. 108 children are selected as controls. Question Discuss some of the considerations in choosing the control group. List the advantages and disadvantages of community and hospital controls. Would you include children admitted with the diagnosis of diarrhea but who do not have salmonellosis as controls? Review the primary need for the controls to represent the population from which the cases came -i.e., the need for choice of cases and controls to be independent of exposure status. Controls, particularly hospital controls, may be biased with respect to the exposure factors. In the case of hospital controls, their hospital status (sick from other conditions) may be causally related to the same factors (€.g., poor nutrition, sanitation) that are associated with the disease under study. The admitting diagnosis of diarrhea is an example of a diagnosis that should be excluded from the control group. Community controls, randomly selected from the true source population, are more difficult and more expensive to identify, locate, and enroll. 12Bacterial Water Contamination ‘Part 4- CASE PROFILE: RATES OF OCCURRENCE From a total of 4095 paediatric admissions during the survey time period (five months), 60 cases (Children with salmonella septicaemia by blood culture) were identified. Among the 60 cases, 46 (77%) were infected with Salmonella typhimurium, 7 (12%) with other types of non-typhoidal salmonella, and 7 (12%) with Salmonella tyhpi (“typhoid fever". Males and females are equally affected. The age distribution of the cases by type of Salmonella infection is shown in Figure 1 Figure 1 ‘Age Distribution of Non-Typhoidal Salmonella Cases ‘Age Distribution of Typhoidal Cases es - OF — 20 20 | | | - 7 5 £ os) | Zool | : 1" j g $j 10 | 6] 2 2 2 S 10 | $2 = 577 z 8 z | 1 st 3 1 Question 6 | What can you say about the pattern of the age distribution for those getting non- typhoidal septicaemia? How does this compare to those who get typhoidal septicaemia? Non-typhoidal salmonella predominantly affects infants (less than 1 year of age) and exclusively affected children in early childhood (5 years or less) in this study. By contrast, typhoidal salmonella predominantly affected older children than the non- 1yphoidal salmonella. There were many more non-typhoidal than typhoidal cases. 13Bacterial Water Contamination PART 5 - DESCRIPTIVE EPIDEMIOLOGY You decide to Took more closely at those cases who have community-acquired infections, in other words, who had the infection prior to coming to the hospital. The cases who meet this definition are 70% (32/46) for S. typhimurium and 100% (7/7) for S. typhi infection. You then analyse the place of residence for the children with community-acquired infections, 47% (15/32) of those with community-acquired S. typhimurium infection came from Nyanza Province, 16% (5/32) from the easter margins of Central Province, and 13% (4/32) from Eastern Province. Only 13% (4/32) were from Nairobi (3 from a shanty area adjacent to a sewage plant). 86% (6/7) of S. typhi infections were acquired in Nairobi shanty towns. The map of the distribution of community. acquired infections is shown in Figure 2. Figure 2. Map of Kenya showing home areas of children with community acquired Salmonella sepicaemia ‘+= One case of Salmonella species iQuestion 7 Bacterial Water Contamination What do you think about the geographical distribution of cases of the two types of salmonella infection? Are they different? What hypotheses do you have about the distribution? There appears to be a clustering of cases of the non-typhoidal species in a few regions and typhoidal species in Nairobi. Nyanza Province is a warm, humid region. Only 4/32 community-acquired non-typhoidal septicaemia cases come from Nairobi. Hypotheses to explain the clustering should include: © Income Status © Nutritional Status Environmental Exposures You should report to the students that Nyanza province has a high incidence of malaria and Nairobi has a low malaria incidence. Do you think that coming from Nyanza Province is a risk factor for developing community-acquired salmonella septicaemia? What additional information would you need to determine the magnitude of possible risk Without a comparison, one cannot state yet whether there is a statistical association between Nyanza Province residence and becoming a case. This could be assessed by looking at where control admissions come from. (However, among the controls there may be selective referral, in relation to geographic location, of different categories of illness to KNED. ISBacterial Water Contamination PART 6 - COMPARING CASES AND CONTROLS: PLACE OF RESIDENCE ‘You compare the honie locations of those with community-acquired S. typhimurium septicaemia with the hospital control group. ‘The results are shown in Table 1. (Note these data are not detived directly from the published case report): ; Table 1. Home locations of cases of Salmonella typhimurium and controls Cases (N=32) Controls (N=108) Nyanza Province 134 15 (Other than Nyanza Province 7 93 Question 9 Calculate and interpret the odds ratio and 95% confidence interval associated with living in Nyanza province and being hospitalised with Salmonella typhimurium septicaemia. OR = 5.5 with 95% Cl =2.1- 14.6 Question 10 What can you now say about the factor, residence in Nyanza Province, in terms of risk for becoming a case? Is it a causal factor? What hypotheses do you have about these findings? Nyanza Province is strongly associated with being a case. We wouldn't say “causal” without greater specificity. Living in Nyanza province is a marker for some other risk factor(s). Hypotheses are numerous, including: water supply, food supply, nutrition, ‘socio-economic status, and medical treatment. 16Bacterial Water Contamination PART 7. COMPARING CASES AND CONTROLS: HYGIENE AND SANITATION ‘Additional results are available to compare community-acquired cases and controls in terms of other environmental factors. The main findings are shown in Table 2. ‘Table 2.. Prevalence of select environmental factors between cases and controls : Factor Cases Controls OddsRatio 95% CI P value (QN=32)_-(N=108) Pit Jatrine use 31,97%) 49.45%) 373. 5:1-761.1 <<0:0001 ‘No piped water available 25(77%) 23 (21%) 13.2. 4,738.8 <<0.0001 ‘Domestic animals kept by family - 27.83%) 50 (46%) 63 21-202 <0,0002 Drink milk from family cow 21 (66%) 18 (17%) 96 3.6-25.7 <<0.0001 Question 11 Set up the 2 x 2 tables and calculate the odds ratio for each factor. Then, to ical significance, (i.e., the probability that the observed difference is due to chance) calculate the chi-square value and the p value. ‘What do you think about the role of these risk factors in causing salmonella infection? Are they important? How do they fit in with your main hypothesis? What additional hypotheses could explain these findings? Answers to calculations: See Table 2 Similar considerations apply for these risk factors as with residence location in terms of causality, but they are biologically plausible CAUSAL factors, relating to human and animal faecal contamination of food (milk and others) and water. Students should comment on the relative magnitude of ORs Question 12 What other kinds of information would you like to have available to compare cases and controls? Testing of water and food sources for contamination would provide more direct evidence of the association. 17Bacterial Water Contamination PART 8 - PUBLIC HEALTH RESPONSES ‘The most severe disease occurred in the 53 children with non-typhoidal septicaemia, with an 18% mortality rate (no patients with typhoid died); 19% (10/93) had coexisting malaria parasitemia (compared to none with typhoid). Antibiotic resistance was high, especially for S. typhimurium, and had increased significantly between 1980-and 1986, Four of the cases had sickle cell disease; 42% of children with non-typhoidal bacteria had severe protein-énergy malnutrition (PEM). You confirm that malarial infections (81% of’ cases come from malaria endemic areas), malnutrition and ‘sickle cell ‘disease. are all associated with increased host susceptibility to infections like salmonellosis. But the underlying problem of contamination of food and water sources must be addressed. Question 13 What recommendations would you make to reduce the risk of Salmonella infections? Recommendations for interventions that should be undertaken: 1. Provision of protected water supplies for domestic use with priority for Nyanza Province. The little available data from Central Province indicated extensive faecal contamination of rivers and wells. 2. Boiling of all domestic milk and drinking water - a difficult task for a busy rural family. 3. Improvement in animal husbandry, veterinary, and environmental health services for subsistence farmers. 4. In hospital, early recognition of patients at risk, provision of isolation ‘facilities and appropriate drug supply: Question 14 What role can you as a health officer play in helping to assist in implementing these recommendations? Which are most important? Which are most feasible? Recommendations 1 and 2 and 3 relate more directly to interventions for public health officer concern - the group should discuss their own experiences in these types of endeavours, including issues of financial and human resources, community education, etc. Questions 13 and 14 to be discussed on group experiences. Based in part on a report by Nesbitt, A. & Mirza, N.B. Salmonella Septicaemias in Kenyan children. Journal of tropical paediatrics, 1989, 35: 35-39, 18CASE III - PESTICIDE POISONING: AN OUTBREAK AMONG ANTIMALARIAL WORKERS Prepared by Linda Rosenstock MD, MPH; Revised by Steven Markowitz, MD February 1990; revised January 1997 INSTRUCTOR'S NOTES This problem-based exercise provides the student with practice in: Establishing a case definition; Identifying the study design; Consideration of sampling strategy; Consideration of non-participant bias; Consideration of exposure routes; Recommending changes to prevent recurrence of epidemic. PART 1 - INTRODUCTION You are the medical officer recently appointed to be in charge of a-large malaria control programme, You learn that there has been a suspected increase in the number of pesticide poisonings starting soon after the beginning of the last spray:season. Question 1 How would you proceed to investigate this situation? What more would you like to know before getting started? Participants should raise questions about what the "suspicion" of the epidemic is based on. Students should raise issues of. D 2 3d y Person, time and place (may refer back to Table 1 of Case I: “Steps in an Outbreak Investigation) What types of pesticide poisonings Sources of information, including ones useful in epidemiological surveys: case registries, hospital records, outpatient records, workplace records, and individuals (parents, employers, community residents, health care workers) Whether new work, exposures, or work processes have occurred. 19Pesticide Poisoning PART 2 ‘You lear that the pesticide malathion (an organophosphate) has replaced DDT this spraying season, because: 1) the mosquito had become resistant to DDT, 2) malathion is an effective pesticide, and 3) malathion is thought to be a relatively safe pesticide for human use based on much experience, including field trials in Nigeria and Uganda. ‘You leam that there are about 7700 antimalaria workers, making up 1100 teams of 7’ workers each (5 spraymen, 1 mixer, and 1 supervisor). In addition to the reported increase in illness (which suggested organophosphate poisoning), 5 deaths have occurred - 2 in mixers and 3 in spraymen. It is thought that one of the three brands of malathion was associated with the most severe illness (used by 3 of the 5 who died). ‘It is also reported that the:illness was more common on Friday and Saturday than on Sunday. Question 2 What appears to be the main exposure problem in the episode described? Malathion, an organophosphate pesticide that was believed to be relatively safe, has caused unexpected episodes of poisoning, including deaths, The problem is serious, unexplained, and needs prompt attention. Question 3 How can you plan organisationally to investigate this outbreak? The study population is large. Are locally available resources sufficient to undertake the study? Use this as an opportunity to discuss resources, including outside assistance; in this case the study was undertaken in collaboration with WHO and the Centres for Disease Control and Prevention, USA. Question 4 What case definition of "poisoning" would you suggest (use Table 1)? Review importance of case definition for proceeding with a formal study. The type of definition will vary depending on available data - e.g. questionnaire Surveys will by definition rely on interview responses and not actual laboratory data. There is a trade-off between broad case definitions that will include all cases but also non-cases and narrow case definitions that will include fewer non- cases, but also fewer cases. A case definition should be based on knowledge about symptoms, signs and laboratory findings, but depends on feasibility of data collection The case definition used in this study is given in Part 3. 20Pesticide Poisoning Mild poisoning : headache Moderate poisoning muscle twitching, tremor Severe poisoning see pulmonary edema anxiety, irritability ‘sweating, salivation. blurred vision vomiting, diarrhoea, al chest tightness, wheezing bradycardia (slow heart rate) or tachycardia (fast heart rate) confusion seizures, coma involuntary defecation, urination What do you think about: a) Why there are more symptoms on Friday/Saturday than Sunday? b) Why there appears to be a problem with a pesticide that has apparently been safely used in other antimalaria programs? a) Inthis case study, workers were off work on Sunday - the increased number of cases at week's end reflected the cumulative exposure to the pesticide (and progressive decrease in cholinesterase levels). b) Particular properties of the pesticide itself, or the way it is being used need to be considered as reasons for the outbreak. Students should also be encouraged to maintain scepticism about past reports of chemical safety. The chemical may not be as safe as advertised. 21Pesticide Poisoning PART 3 “The occurrence of cases of poisoning has been confirmed. Cases occur predominantly towards the end of the working week. You decide to study it further with a questionnaire survey. You define a case as follows: 1) occurring in a spray team member; 2) having at least 4 of the following 5 symptoms: blurred vision, dizziness, nausea, vomiting, abdominal pain, ‘You decide to interview a random sample (10%) of all the antimalaria workers to ask them about their past and present symptoms and their exposures at work. Question | What type of epidemiologic study is this survey? This is a cross-sectional study - at one point in time subjects will be investigated and exposures and effects will be assessed. Question 7 What are the advantages and weaknesses of: a) this study design? b) this case definition? ¢) this sampling strategy? Tssue Advantages Disadvantages a) cross-sectional ‘Relatively inexpensive; Selection bias - many affected workers study design easy ‘may no longer be working and therefore unavailable to participate in study Recall bias - rely on memory of symptoms in the course of an epidemic 3) case definition | Easy to elicit (symptoms only); |Subjective only; may miss milder cases cases are likely to be "true" | Symptoms are not specific for pesticide cases because definition is poisonings relatively narrow ©) sampling strategy | Easy to perform; random; ‘May not be representative, may not be not influenced by investigator | large enough to study subgroups 22