Are Campylobacter cases low risk for public health follow-up?
Campylobacter is a communicable disease reportable to the local Medical Officer of Health under the Ontario Regulations 559/91 and amendments under the Health Protection and Promotion Act. All laboratory-confirmed cases of Campylobacter residing in Toronto, aged between 5 and 64 years and with no evidence of attending a day nursery or living in a long-term care home, are mailed letters and questionnaires to assess how they acquired their illness and whether they worked in a high-risk occupation. Any laboratory-confirmed cases who did not respond within 30 days after a letter and questionnaire were mailed by TPH were identified as potential participants in this study.
A list of cases reported to TPH between June 11 and December 6, 2012 as having not returned the questionnaire within 30 days was generated on a weekly basis. A telephone survey, conducted by a research student, was administered to all cases reached during the study period in order to assess: awareness of the letter sent by TPH; reasons for not responding; and whether the client worked in a high-risk occupation. Cases reached by telephone were informed that participation was voluntary and were asked for consent for participation. For all cases <18 years of age, the parent was interviewed; in the absence of the parent, the guardian was interviewed. Authenticity of the call from the local health authority was supported by the caller's knowing the participant's name, postal code, and family doctor's name. Potential participants were called a maximum of 10 times. If the participant worked in a high-risk occupation, they were asked if they were amenable to having a Public Health Investigator call to discuss the matter further. If the participant indicated they had worked at all while ill, they were directed to the general public health information available on the TPH website or by calling the TPH Inquiry Line.
Data were analyzed (Microsoft Excel 2010) and deemed significant at a level of p<0.05.
This project underwent Toronto Public Health's research review process and was approved by the Ryerson University Research Ethics Board, no. REB 2012-057.
Attributes of participants
Campylobacter cases included in this study were reported to TPH between April 3, 2012 and November 5, 2012. There was a total of 226 cases identified as non-responders and available for follow-up (versus 279 cases who had previously responded by mail); 172 (76.1%) cases were successfully reached by telephone. One hundred and sixty-two questionnaires were completed (71.7%) with 10 cases declining to answer the questionnaire, giving a cooperation rate of 94.2% (162/172).
The 162 participants were compared to the 279 Campylobacter cases who did return the original mailed-out questionnaire during the same period. Demographic characteristics between study participants and respondents to the original letter did not differ. Two worked in a high-risk occupation, as food handlers.
An average of 2.6 call attempts were needed to reach participants, and Mondays between 11 am and noon were the times with most success. When asked about the original questionnaire, 130 (80.2%) recalled receiving it, and of these 39 claimed they had responded and 84 stated they were still planning to respond. There were 29 responses to an open-ended question pertaining to the reason questionnaires were not returned, the most frequent of which was "too busy" (n=8, 27.6%). When prompted with a defined list of reasons for not returning the questionnaire, 79 responded and the most frequent response was that they "forgot" (54.4%), followed by "recovered from illness" (21.5%) (Table 1).
Of the 119 cases who answered the question on occupation at the time of illness, 53 (44.5%) reported being a student, 47 (39.5%) were employed, 16 (13.4%) were unemployed, and 3 (2.4%) reported being employed in a sensitive occupation category; these included a family physician, a food server, and a line cook. Forty-three cases answered the question concerning working while sick: of the 11 (25.6%) who reported that they did work while ill, none were employed in a high-risk occupation.
When asked an open-ended question pertaining to what would have encouraged cases to return the questionnaire, 25 responded, with the most frequent response being "a reminder" (n=14, 56.0%). However, when prompted with a defined list of suggestions, 63 responded and the most frequent response was "more information on importance of returning questionnaire" (19.1%), followed by "a simpler form" (15.9%) (Table 2).
Participation, attributes, occupation
The completion rate (71.7%) and cooperation rate (94.2%) were high in this study, likely due to the topic's concerning the participant's own health. Other similar studies on personal health issues have not produced as high completion/response rates. For example, in a Swedish telephone survey of non-respondents to a mail survey on general health issues, conducted over two months in 2004, the response rate was 49.3%. (2) In a telephone survey follow-up to a mail survey sent to individuals treated for ulcerative colitis in March 2009 in Minnesota (United States), the response rate was 43%. (3)
Annual summaries from Toronto Public Health show an age distribution of Campylobacter cases with peaks in those aged under 5 years and those aged 20 to 29 years. (4) While children under 5 years are not eligible for the LRE program, the current study showed that those ages 10-29 years comprised 49.4% of all participants, compared to only 37.3% in the same age group who had responded to the mail survey (p=0.012). It may be necessary to consider other ways to survey clients, such as an option for online response; this could be particularly effective with younger-aged clients. In a US randomized survey where over 32,000 persons rated their own health, there was a higher response rate from web surveys, but this resulted in upwardly biased measurements of self-related health compared to mail surveys. The authors suggested that a combination of survey methods (mixed modes) can enhance coverage to a broader mixture of respondents. (5) Although not a health study, a study conducted by the U.S. Department of Defense with over 77,000 military personnel responding, compared demographic and health characteristics of Web responders with paper responders. Web responders were more likely to be younger and overweight or obese. Additionally, web responders provide more complete contact information, including their e-mail addresses. Survey research conducted over the Internet may need to be conducted along with traditional mail survey strategies to reflect behavioural or demographic differences. (6) However, there are challenges with online studies, such as the legal implications where adult/parental consent is required.
In a previous unpublished study by TPH (2011) for cases reported from 2006-2009, 1.7% of all (includes Campylobacter enteritis, amebiasis, giardiasis, yersiniosis) LRE cases reported working in sensitive occupations (i.e., health care, food preparation, or in a daycare centre). Of the three participants in this study who identified their occupation as one considered to be sensitive, one was a family physician. No published evidence of Campylobacter cases being transmitted from a health care worker to their charges was found, although it is theoretically possible. The most frequent sources of Campylobacter in our environment are not people but foods of animal origin, water, and animals. (7) Two of the three participants in this study who identified their occupation as sensitive worked in the food industry, one as a server and the other as a line cook. Campylobacter illness frequently arises in situations in which persons ingest undercooked chicken or unpasteurized milk. (8) As well, a food worker in preparing chicken can transfer Campylobacter from their fingers or utensils to ready-to-eat foods, such as produce, which will not undergo a further step to inactivate Campylobacter. (9,10) Additionally, a food worker preparing chicken could contaminate surfaces with chicken juices; another person could then touch this same surface and put their fingers into their mouth, unknowingly ingesting the bacteria. The restaurant server identified in this study would not likely be handling foods that would present a great risk to the public. However, it is possible that the line cook identified in this study could transmit Campylobacter from feces (i.e., via inadequate handwashing after toileting) to food or common surfaces.
In an outbreak in a Kansas (US) school in 1998, 129 Campylobacter cases met the case definition. (11) Epidemiologic investigation identified the gravy made from an instant mix and canned pineapple, both served at common meal, as the likely sources. A food handler reported severe diarrheal cramps and profuse diarrhea necessitating numerous trips to the bathroom during a lunch-serving shift that day. During another Campylobacter outbreak at a boys' summer camp in Connecticut (US) in 1980, 40 campers and staff became ill. (12) Eating salad prepared by an ill staff member was determined to be the cause. Kitchen staff members were observed to be unhygienic in their handwashing practices, i.e., not using soap and wiping their hands on their aprons.
None of the three participants working in sensitive occupations in this study worked while ill; the family physician particularly mentioned having had "no contact with patients for two weeks" during the (physician's) period of illness. Therefore, the likelihood of direct or indirect transmission of Campylobacter from these three individuals was low. Additionally, in the control group of 279 Campylobacter cases, only two (food handlers) stated that they were in high-risk occupations. Based on information from cases who were successfully contacted and who self-reported on their occupation and whether they had worked while ill, it would appear that at least in this study sample there was little risk to the public with regard to ongoing transmission. Since 43 cases did not state their occupation, it is acknowledged that there could be additional cases working in sensitive occupations beyond what was reported. However, the very few reported instances when Campylobacter has been transmitted from infected persons through a workplace exposure support the classification as low-risk.
When presented with an open-ended question regarding the reason for not returning the questionnaire, 27.6% (8/29) participants volunteered that they were "too busy". However, when prompted with a list of reasons for not returning the questionnaire, the majority responded that they "forgot" (54.4%), followed by "recovered from illness no longer considered it relevant" (21.5%), "form too complex" (10.1%), and "not important to return" (8.9%). These findings suggest that a more detailed explanation in a cover letter explaining as to why the response is being solicited by Public Health, even after the subject has recovered from the illness, and a simplification of the form for mail return may increase response rates.
Respondents were asked for ways that would encourage them to return the questionnaire and only one participant said an "incentive" would have worked. However, in a study in 2005-6 in Los Angeles County among Hispanic women diagnosed with breast cancer, patients were mailed a questionnaire and $10 incentive, and non-respondents were contacted by phone. (13) Subjects were called a minimum of five times within two to three weeks of the mailing if no response was received. A 72% response rate was achieved; 98% completed the survey and 2% were interviewed by telephone. The high response rate was attributed to the incentive, extensive telephone follow-up, a native Spanish-speaking interviewer, and a focused questionnaire. In the current study, lack of an incentive was not an impediment to a high cooperation rate. We would suggest rather that participants are more likely to answer questions on health issues that affect them personally. Additionally, the cover letter to the questionnaire could stress the value of their timely response, i.e., in contributing to detecting outbreaks and epidemics and preventing further transmission in the community.
This study showed that Mondays proved to be the best days for speaking with participants. This was somewhat biased and self-fulfilling because halfway through the study, success on that day of the week became obvious and therefore more calls were placed on Mondays. Similarly, the time most people were home was around the lunch hour and calls were made then. This may be due to the large number of students and unemployed participants, who may be available to take calls when at home having lunch.
Our study has several limitations. The generalizability of this information is limited to other large urban centres with similar population diversity and strong presence of public health authorities. Cities with high levels of distrust or a less ubiquitous public health presence may not find the LRE method of much value. Due to time constraints, the study was only conducted over six months and not a full year. The six months, however, did include the typical enteric months (i.e., late spring, summer, early fall) and there is no reason to believe that those who are sick in winter or the colder months would differ in any way significant to this study's findings. It is worth noting that we relied on an honour system to determine if a client was working in a sensitive occupation and while ill. While the same honour system is assumed when clients are telephoned directly (i.e., not eligible for LRE program), we do not know if this reflects the same information a client would have shared if they responded to the mailed-in questionnaire only (i.e., routine LRE). Additionally, participants may have been led in their answers with the prompts used to facilitate use of categorical data.
Overall, the data collected during this study support that the LRE program should continue for Campylobacter cases. While some cases working in high-risk occupations can be missed, the number was small and confirmed that despite the occupational risk and without contact by PH, clients all chose to cease working during their illness. Taken together with suggestions for how to improve the questionnaire and letter, using a LRE system to monitor a widespread mostly low-morbidity gastroenteric illness can be an effective public health strategy.
Acknowledgements: We thank Ryerson student researcher Luckrezia Awuor for her assistance with this project, TPH Epidemiologist Sylvia Ota for her assistance with daily extractions and quality assurance, and Monica Mitchell for facilitating some of the TPH logistics. We also want to acknowledge funding from the Faculty of Community Services at Ryerson University for this project.
Conflict of Interest: None to declare.
(1.) Toronto Public Health. Low-risk enteric letter-based program evaluation, Communicable Disease Surveillance Report. March 2011. Unpublished report.
(2.) Linden-Bostrom M, Persson C. A selective follow-up study on a public health survey. Eur J Public Health 2012 Jan 19. [Epub ahead of print]. Available at: http://eurpub.oxfordjournals.org/content/early/2012/01/19/eurpub.ckr193.fu ll.pdf+html?sid=a358e03e-109f-486f-ba4f-feaa96c42f69 (Accessed January 2, 2013).
(3.) Ziegenfuss Jy, Burmeister KR, Harris A, Holubar SD, Beebe TJ. Telephone follow-up to a mail survey: When to offer an interview compared to a reminder call. BMC Med Res Method 2012;12:32.
(4.) Toronto Public Health. Enteric, food and waterborne diseases. Communicable diseases in Toronto. 2011.
(5.) Shim JM, Shin E, Johnson TP. Self-rated health assessed by Web versus mail modes in a mixed mode survey: The digital divide effect and the genuine survey mode effect. Med Care 2013;51(9):774-81.
(6.) Smith B, Smith TC, Gray GC, Ryan MA; Millennium Cohort Study Team. When epidemiology meets the Internet: Web-based surveys in the Millennium Cohort Study. Am J Epidemiol 2007;166(11):1345-54.
(7.) Taylor EV, Herman KM, Ailes EC, Fitzgerald C, Yoder JS, Mahon BE, Tauxe RV. Common source outbreaks of Campylobacter infection in the USA, 19972008. Epidemiol Infect 2012;14(15):987-96.
(8.) Centers for Disease Control and Prevention (CDC). Campylobacter jejuni infection associated with unpasteurized milk and cheese--Kansas, 2007. MMWR Morb Mortal Wkly Rep 2009;57(51):1377-79.
(9.) Jimenez M, Soler P, Venanzi JD, Cante P, Varela C, Martinez Navarro F. An outbreak of Campylobacter jejuni enteritis in a school of Madrid, Spain. Euro Surveill 2005;10(4):118-21.
(10.) Calciati E, Lafuente S, De Simo M, Balfagon P, Bartolome R, Cayla JA. Campylobacter outbreak in a Barcelona school. Enferm Infecc Microbiol Clin 2012;30(5):243-45.
(11.) Olsen SJ, Hansen GR, Bartlett L, Fitzgerald C, Sonder A, Manjrekar R, et al. An outbreak of Campylobacter jejuni infections associated with food handler contamination: The use of pulsed-field gel electrophoresis. J Infect Dis 2001;183(1):164-67.
(12.) Blaser MJ, Checko P, Bopp C, Bruce A, Hughes JM. Campylobacter enteritis associated with foodborne transmission. Am J Epidemiol 1982;116(6):886-94.
(13.) Hamilton AS, Hofer TP, Hawley ST, Morrell D, Leventhal M, Deapen D, et al. Latinas and breast cancer outcomes: Population-based sampling, ethnic identity, and acculturation assessment. Cancer Epidemiol Biomarkers Prev 2009;18(7):2022-29.
Received: July 23, 2013
Accepted: October 24, 2013
Marilyn B. Lee, ScM,  Effie Gournis, mph, [2,3] Richard J. Meldrum, PhD 
[1.] School of Occupational and Public Health, Ryerson University, Toronto, ON
[2.] Toronto Public Health, Communicable Disease Control, Toronto, ON
[3.] Dalla Lana School of Public Health, University of Toronto. Toronto, ON
Correspondence: Marilyn B. Lee, Professor, School of Occupational and Public Health, Ryerson University, 350 Victoria St., Toronto, ON M5B 2K3, Tel: 416-979-5154, Fax: 416-979-5377, E-mail: firstname.lastname@example.org
Table 1. Defined List of Reasons Why Participants Did Not Return the Mailed Questionnaire Reasons (When Prompted) No. of Participants (%) Forgot to return 43 (54.4) Recovered from illness--no 17 (21.5) longer considered it relevant Form too complex 8 (10.1) Not important to return 7 (8.9) Lost form 2 (2.5) Illness very mild--did not 2 (2.5) consider it important English a second language 0 Did not want to disclose 0 personal information Total 79 (99.9) Table 2. Defined List of Suggestions Which Would Have Encouraged Return of Questionnaire in the Mail Suggestions (When Prompted) No. of Participants (%) More information on importance 12 (19.0) of returning questionnaire A simpler form 10 (15.9) Telephone contact rather than 6 (9.5) by mail Some form of incentive 1 (1.6) Face-to-face contact rather 0 than by mail Availability of electronic or 0 online submission of questionnaire Offer language interpretation 0 services No factor would have encouraged 34 (54.0) me to return the form Total 63 (100.1)
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||QUANTITATIVE RESEARCH|
|Author:||Lee, Marilyn B.; Gournis, Effie; Meldrum, Richard J.|
|Publication:||Canadian Journal of Public Health|
|Date:||Nov 1, 2013|
|Previous Article:||The rising burden of rheumatoid arthritis surpasses rheumatology supply in Ontario.|
|Next Article:||Occupational exposures to antineoplastic drugs and ionizing radiation in Canadian veterinary settings: findings from a national surveillance project.|