Cryptosporidiosis Outbreak in Huntsville - Case Study

Part I. Outbreak Detection

Centers for Disease Control and Prevention (2010). Cryptosporidiosis in Georgia.  A pdf copy of the document for has been posted on the course shell under Evaluations.  You can also access the document using the direct links below (you may need to cut/paste the link directly into your browser).  This document will provide useful resource information to help with this assignment.

After completing this case study, you should be able to:

1. Discuss epidemiologic clues indicative of a waterborne disease outbreak;
2. Describe a boil-water advisory and discuss problems that might be encountered in issuing such an advisory;
3. Help plan a community survey to determine the prevalence and distribution of a health problem;
4. Interpret a dose-response analysis for an exposure and development of a disease;
5. List activities that should be included in the evaluation of a public water system associated with an outbreak;
6. Define turbidity, total coliform count, and E. coli count, and discuss how each are used to indicate drinking water quality;
7. Discuss the typical steps used in the treatment of surface water at a municipal water treatment plant; and
8. Describe the clinical features, epidemiology, and control of cryptosporidiosis.

On January 21, a physician notified the Muskoka Health Unit (MHU) of a dramatic increase in acute gastroenteritis among students at a college in Huntsville, Ontario. Huntsville is a small town located in central Ontario [population estimate: 18,000]. The physician reported examining hundreds of students with a gastrointestinal illness at the college’s infirmary during the previous week. The majority of students reported having abdominal pain and watery diarrhea of several days duration.

Typically, approximately 100 students were examined at the infirmary each day, the majority of whom presented with upper-respiratory infections or injuries. On January 20 alone, over 200 students were examined at the infirmary for gastroenteritis.

Question 1: Do you think these cases of gastrointestinal illness represent an outbreak at the college? Why or why not? (3 marks)

The physician reported that stool specimens had been collected from selected ill students and had been submitted to the hospital laboratory for testing. Because the physician had an interest in parasitology, he had specifically requested that the stools also be examined for parasites.

The physician reported that cultures for bacterial pathogens had been negative. On microscopic examination, four specimens were determined to be positive for Cryptosporidium. The earliest known onset of illness among the Cryptosporidium-positive students was January 11.

Question 2: How is cryptosporidiosis transmitted? On what sources of infection should public health officials focus? (4 marks)

Question 3: Characteristics among patients can provide clues that indicate one mode of transmission over others. Briefly describe the outbreak characteristics you would expect to see based on the following types of transmission. (10 marks) 

• Person-to-person transmission
• Transmission by a food
• Transmission by public drinking water
• Transmission by contaminated private well water
• Transmission by recreational water

Staff from the college contacted the MHU about the outbreak. Because the majority of the ill students lived on campus and participated in the college meal plan, concerns were voiced regarding possible foodborne transmission of the infection at the college. Within a few hours of the initial report, an environmental public health specialist from MHU was dispatched to the college to interview food service staff and to inspect the kitchen at the main cafeteria for food safety problems.

Question 1: Do you think these cases of gastrointestinal illness represent an outbreak at the college? Why or why not? (3 marks)

Meanwhile, staff from MHU undertook steps to determine whether the cryptosporidiosis problem extended beyond the college.

Question 4: List eight existing sources of information that might help determine if others in the community have cryptosporidiosis. (8 marks – note there are at least 11 potential sources of information)

Question 5: Is Cryptosporidium a reportable disease in Ontario? (2 marks)  

A review of the patient intake log from the local hospital emergency department provided evidence of an increase in patients presenting with gastrointestinal illness starting the week of January 11. However, the majority of these patients were not students from the college. Calls to selected public schools and large businesses in Huntsville revealed widespread absenteeism. In addition, data from local pharmacies indicated an increase in sales of antidiarrheal medications throughout the region.

On the basis of these findings, staff from MHU concluded that the outbreak was not confined to the college but involved the entire community. Reports from three communities bordering Huntsville indicated that those areas were not severely affected.

To characterize the cryptosporidiosis cases and to seek clues about the source of the Huntsville outbreak, investigators from MHU examined medical records from patients with acute gastroenteritis examined at the local hospital’s emergency department. During December 14 through January 25, a total of 98 patients were examined at the emergency department for acute gastroenteritis. The number of visits increased from 8−12 each week through the first week of January to 22−28 each week during the last 2 weeks of January (Figure #1 – attached).

Among the 65 patients with acute gastroenteritis examined at the emergency department since January 11, when the first Cryptosporidium-positive student had onset of illness, the following symptoms were reported: diarrhea (defined as three or more loose stools/day) (87%), stomach pain (80%), nausea (67%), vomiting (33%), fever (30%), and muscle aches (20%). Patients often reported that their diarrhea was watery. Approximately half of patients reported that their symptoms had been present for more than a week at the time they presented to the emergency department.

Patient’s ages ranged from younger than a year to 76 years (mean: 34 years); 63% of patients were female. The majority of the patients lived within Huntsville’s town limits as opposed to other parts of the region. Cases did not appear to cluster by neighborhood of residence, child care center, or school but were widely distributed around the city. Information on specific exposures was unavailable.

Question 2: How is cryptosporidiosis transmitted? On what sources of infection should public health officials focus? (4 marks)

Stool specimens, available from 25 patients, were forwarded to the Public Health Laboratories. Cryptosporidium was identified in 11 (44%) of the 25 specimens. One stool from a child aged 2 years was also positive for Giardia.

Question 6: Interpret the descriptive epidemiology of the outbreak including the epidemic curve (Figure #1 – attached). Were symptoms among patients consistent with cryptosporidiosis? Was clustering of cases apparent by selected demographic characteristics? What was the course of the outbreak and did it appear to be over? (4 marks)

Cases began to increase the second full week of January, peaking the week of January 18. Because reporting for the week of January 25 was incomplete, whether cases were declining is difficult to determine. Given the incubation period of cryptosporidiosis (i.e., average of 7 days), exposures of interest probably occurred as early as the first week of January. Because of the wide geographic distribution of cases within Huntsville’s town limits and the occurrence among all age groups, water was suspected as a possible source o the outbreak.

Investigators reviewed public water system records and collected water specimens from multiple points in the water distribution system.

Routine Total Coliform and E. coli counts from water samples collected from the public water system on January 5 were negative. In addition, samples of water collected from the college in Huntsville on January 22 and the public water system on January 23 were also negative for coliforms and E. coli.

Question 7: What do total coliform counts indicate? What do E. coli counts indicate? What does negative total coliform and E. coli results in treated water mean? (6 marks)

Investigators contacted the directors of the four nursing homes in the Huntsville area. The three nursing homes connected to the public water supply reported substantial numbers of residents with acute gastroenteritis. The nursing home that used a well for their water supply reported no residents with acute gastroenteritis.

On the basis of this information, the MHU issued a boil-water advisory on January 25.

Question 8: What is a boil-water advisory? List four key points that you would include in a BWA. (6 marks)

Studies report that residents often continue to consume untreated water after issuance of a boilwater advisory, resulting in additional cases of the waterborne disease, because they did not hear about the advisory or understand the severity of the situation, the procedures, or the duration of the advisory.

Question 3: Characteristics among patients can provide clues that indicate one mode of transmission over others. Briefly describe the outbreak characteristics you would expect to see based on the following types of transmission. (10 marks)

Question 9: When boil-water advisories are issued, what could water supply operators, local governments, and public health officials do to ensure that all residents are informed of the health risks and consequences of noncompliance with the BWA? (4 marks)

During a BWA members of the public often have concerns and regularly ask questions of the local public health officials.

Question 10: Formulate a response to the following frequently asked questions during a BWA: (8 marks – 2 marks per response)

1. Can I use my dishwasher during a BWA?
2. How long do I have boil water to kill cryptosporidium?
3. Can I use activated charcoal filters (e.g. Brita) to treat my water during a BWA?
4. Can my pet drink the water during a BWA?

Question 11: What studies or investigations might you undertake to confirm the hypothesis that the public water supply was the source of Huntsville’s cryptosporidiosis outbreak? (3 marks)

On January 25, a case-control study was undertaken among patients presenting at the local hospital emergency department. Case-patients were persons who had presented to the emergency department with abdominal pain or diarrhea (defined as three or more loose stools in a 24-hour period) since January 1. Control-subjects were patients presenting to the emergency room on the same dates as the case-patients but for non-gastrointestinal illnesses. 

Twenty (80%) of 25 case-patients and 3 (33%) of 9 control subjects had been exposed to the public water supply at home or work (odds ratio: [OR] 8; P value = 0.03). The case-control study, however, was plagued with problems. Multiple patients initially selected as control subjects, but who were later excluded, reported having been ill with abdominal pain and diarrhea even though they had been examined at the emergency department for non-gastrointestinal illnesses.

Investigators questioned the validity of the results. Therefore, two simultaneous lines of investigation were undertaken to determine if the public water supply was the source of the outbreak: a community survey and an evaluation of the water treatment plant in Huntsville, Ontario.

Staff from the MHU met with epidemiologists from Public Health Ontario to plan the community survey.

Question 12: List four steps you would undertake to conduct a community survey. (4 marks)

The goals of the community survey were to determine the number of persons affected by cryptosporidiosis in Huntsville and to examine the association between public water consumption, other possible risk factors, and gastrointestinal illness.

The goal in selecting a sample is to represent the original target population as accurately as possible. The sample should be similar to the target population with regard to all relevant characteristics, except that it includes fewer persons. Multiple approaches are available for selecting a sample including the following:

• Simple random sampling − Simple random sampling gives every member of the population an equal chance of being selected for the sample. A listing of the population is obtained and members are randomly selected from the list.
• Systematic sampling − Systematic sampling is similar to simple random sampling in that it also gives every member of the population an equal chance of being selected. However, rather than randomly selecting participants from a listing of the population, a starting point on the list is randomly selected, and participants are selected on the basis of a sampling interval (e.g., every tenth person).
• Stratified random sampling − In stratified random sampling, the target population is divided into non-overlapping subsets (i.e., strata) on the basis of at least one characteristic (e.g., sex). A simple random or systematic sample is then selected within each stratum. Each member of a particular stratum has an equal chance of being selected; however, the probability of selection might differ between members in different strata.
• Cluster sampling −In cluster sampling, the population to be studied is divided into natural, geographically distinct groups or clusters (e.g., schools, villages, or camps). A sample of clusters is then selected by using simple random sampling, systematic sampling, or stratified random sampling. After the clusters are selected, all units within the selected clusters are included in the sample. No units from non-selected clusters are included in the sample.
• Multi-stage sampling − Multi-stage sampling is a complex form of cluster sampling. In the first stage, clusters (i.e., primary sampling units) are identified and a sample of the clusters is selected by using simple random sampling, systematic sampling, or stratified random sampling. In the second stage, units within the clusters (i.e., secondary sampling units) are randomly selected.

Question 4: List eight existing sources of information that might help determine if others in the community have cryptosporidiosis. (8 marks – note there are at least 11 potential sources of information)

If a population listing is available, simple random sampling and systematic sampling are conceptually easiest to implement. A stratified random sample should be used when the population can be divided into meaningful subsets and estimates for the different subsets are desired. Cluster sampling, including multistage sampling, is good to use when the population of interest is too large, cannot be enumerated, or is distributed widely. For results from a cluster sample to reach the same level of precision as a simple random sample, the sample size must be appreciably greater, which increases the needed resources. Cluster sampling also requires more complex statistical analysis to account for the mode of sampling.

For this outbreak and setting in which a listing of all potential survey subjects is available, simple random sampling (e.g., random-digit-dialing) or systematic sampling (e.g., selecting households from the telephone directory or a county census) are reasonable approaches. The investigator might want to stratify the population according to those who live within the town limits of Huntsville and those who live in region outside of the town.

To select the survey sample investigators systematically selected 400 listings (i.e., households) from the Huntsville telephone directory. A larger proportion of Huntsville telephone numbers were selected (i.e., were oversampled) to ensure that the sample included an adequate number of persons who had been exposed to the public water supply (i.e., the suspected source of the outbreak).

A PHO epidemiologist drafted a questionnaire for the survey. The questionnaire was piloted with staff from the MHU who were not involved in the investigation.

Options for collecting survey information include face-to-face interviews, telephone interviews, or self-administered questionnaires. The methods used to collect the necessary information will be based primarily on the time available to collect the information (i.e., the need for timely information to support necessary public health actions), likely costs and resources, characteristics of the target population, and the sensitivity of the information collected.

Question 13: Identify one advantage and one disadvantage for each collection method listed: (6 marks)

1. Face to Face Interview
2. Telephone interview
3. Self-administered Questionnaire

Information for the survey was to be collected through telephone interviews. By using the questionnaire developed by the MHU epidemiologist, one adult in each selected household (referred to as the respondent) was to be asked his or her age and sex, place of employment (or school), food and restaurant exposure, home water source, amount of tap water consumed, consumption of ice, and exposures to children in child care centers and to farm animals. The respondent also was to be asked about the age, sex, and place of employment or school of all household members and whether the household member had been ill with abdominal pain or diarrhea (defined as three or more loose stools within a 24-hour period) since January 1.

Question 5: Is Cryptosporidium a reportable disease in Ontario? (2 marks)

On the afternoon of January 30, a total of 12 staff from MHU were trained to administer the survey questionnaire by the telephone. Starting that evening, they telephoned each household on the list.

Question 14: What activities or efforts might help to improve the survey response rate? (3 marks)

By February 5, adult respondents were interviewed at 304 (76%) of the 400 telephone numbers. Fifty-six of the listings were disconnected numbers; 31 had no answer after three calls or had no adult available to complete the interview; and nine adult respondents refused to participate or did not complete the interview. Information was collected from the 304 adult respondents and 507 additional household members for a total of 811 household members.

Investigators calculated the overall attack rate among all household members and attack rates by residence and exposure to the public water supply. Investigators set a P value of 0.05 as the cutoff for statistical significance. The source of home water was based on information provided by the adult respondent from each household. Public works engineers determined the water supply for worksites and schools. Persons whose home, school, or worksite was supplied with public water were considered to have been exposed to the public water supply.

Of the 811 household members interviewed, 363 had been ill with abdominal pain or diarrhea (defined as three or more loose stools within a 24-hour period) since January 1. After adjusting for oversampling of households from Huntsville, the overall attack rate for the Region was 40%. Attack rates varied by residence and exposure to the public water supply.

Among the 489 household members exposed to the public water supply, the attack rate was 67% for females and 55% for males (relative risk: 1.2; 95% confidence interval 1.1−1.4; P value=0.01). Attack rates varied by age group, ranging from 52% to 72%. The highest attack rate was among persons aged 20−29 years, but these differences were not statistically significant.

Residents of the Huntsville were twice as likely as nonresidents to become ill. Persons who were exposed to the public water supply at home, school, or work were 3 times as likely to become ill as those who were not. Both findings were unlikely (i.e., less than 5 chances in 100) to have occurred because of chance alone.

The increased risk for infection among persons residing in Huntsville might reflect the fact that they were more likely to be exposed to the public water supply through their homes than people living in the region but outside of Huntsville town limits.

Among the 304 adult respondents to the survey, 182 were exposed to the public water supply and provided information on the average amount of water they consumed each day (Table #2 – attached).

Question 15: Graph and interpret the association between water consumption and the occurrence of illness. (6 marks)

Among adult respondents exposed to the public water supply, the higher attack rate among females remained significant even after controlling for age and water consumption. No significant association was identified between illness and any other exposure studied (e.g., exposure to specific foods, child care centers, farm animals, or other ill persons).

Question 16: On the basis of Table #1 (attached), 20% of persons not exposed to the public water supply became ill. If the public water supply was the source of the outbreak, what explanations exist for these persons becoming ill? (2 marks)

On January 25, an evaluation of the Huntsville water system was initiated by Public Health and Ministry of the Environment inspectors.

Question 17: What activities would you include in the evaluation of a public water treatment plant? With whom would you talk? What records or data sources would you review? (8 marks)

Question 18: List and draw the typical steps in the treatment of surface water at a public water treatment plant. (6 marks)

The inspectors reviewed plant water-quality data (e.g., records of pH, temperature, turbidity, and total coliform counts) of untreated and treated water. They measured the turbidity of the water at the intake and the outflow from the facility and collected samples of raw and treated water to test for coliforms and Cryptosporidium.

Question 19: What does turbidity indicate? Why would the investigators be interested in turbidity of both untreated and treated water? (4 marks)

Cryptosporidium oocysts were identified in samples of treated water taken from the treatment plant and from four dead-end water mains (i.e., mains that supply water from only one direction such that water stops circulating and becomes stagnant), including one at the college first reporting the outbreak.

Question 20: How is it possible that treated water from the public water system was negative for coliforms yet contained Cryptosporidium oocysts? (2 marks)

Question 21: What water treatment measures are effective against Cryptosporidium? (2 marks)