Environmental Health Specialists' Self-Reported Foodborne

FEATURES

REFERENCESEnvironmental Health specialists’ self-reported Foodborne illness outbreak investigation practices

FIGURE Carol A. Selman Laura R. Green, Ph.D.

however, been criticized (Bryan, 2002; abstract To collect qualitative data on the investigation practices Food and Drug Administration, 2001; TABLE of environmental health specialists with respect to food- Jones et al., 2004). Some have contended that investigations often yield inadequate borne illness outbreaks, the authors convened six focus groups of randomly selected information concerning the etiology and specialists working in public health agencies in eight states. Participants discussed contributing factors of foodborne illness outbreaks. Contributing factors were their investigation activities, methods used to identify contributing factors, success identified in only 57 percent of outbreaks in identifying contributing factors, and the difficulties they faced when conducting reported to the Centers for Disease Con trol and Prevention (CDC) (Olsen, MacK investigations. Findings revealed substantial variability in the type of activities in inon, Goulding, Bean, & Slutsker, 2000). SwhichPEC participantsIAL RengagedEPOR during Tinvestigations, and the amount and nature of the This inadequacy may result, at least in part, from specialists’ ineffective outbreak collaboration between epidemiologists and environmental health specialists during investigation practices and from barriers investigations. Many participants indicated that during investigations they often did encountered during investigations (Bry an, 2002; Ehiri & Morris, 1994; Jones et not identify contributing factors associated with an outbreak. Participants also iden al., 2004). tified several difficulties associated with outbreak investigations, including difficul The Environmental Health Specialists Network (EHS-Net) wished to gain a bet ties associated with restaurant employees, restaurant customers, and environmental ter understanding of specialists’ foodborne health organizations. illness outbreak investigation practices and consequently spearheaded the study report ed here. EHS-Net is a collaborative project introduction tivities because they can identify the patho focused on food and water safety research For the years 1993 to 1997, an annual av gens that caused an outbreak and the cir and includes epidemiologists and environ erage of 550 foodborne illness outbreaks cumstances that led to the introduction or mental health specialists from CDC, the was reported in the United States (Olsen, proliferation of those pathogens (i.e., con Food and Drug Administration (FDA), the MacKinon, Goulding, Bean, & Slutsker, tributing factors such as improper holding U.S. Department of Agriculture (USDA), 2000). More than 40 percent of these out temperatures and worker bare-hand contact the U.S. Environmental Protection Agency breaks were attributed to food service estab with food). Knowledge of the pathogens (U.S. EPA), and nine state public health lishments. Environmental health specialists and contributing factors associated with agencies. Currently, EHS-Net includes Cali along with epidemiologists and laboratori outbreaks can increase understanding of fornia, Connecticut, Georgia, Iowa, Minne ans in federal, state, and local public health the causes of, and prevention strategies for, sota, New York, Oregon, Rhode Island, and agencies are typically involved in investiga outbreaks. (Jones et al., 2004). Tennessee; before 2005, Colorado was an tions of foodborne illness outbreaks. These The quality and effectiveness of food- EHS-Net state and Iowa and Rhode Island investigations are important food safety ac borne illness outbreak investigations have, were not.

16 Volume 70 • Number 6 Reprinted with permission from NEHA FEATURES

REFERENCES

FIGURE

method TABLE 1 In 2004, EHS-Net convened six focus groups of specialists who worked in state or Characteristics of Participants (n = 42) local public health departments in EHS-Net states. The groups met through telephone Characteristic Percentage of conference calls, a cost-effective method of Participants collecting data from participants who are Gender geographically dispersed. This method has Male 50 been found to generate as much information SPECFemale IAL REPORT 50 as the use of face-to-face focus groups (Sil Age verman, 2003). To enroll participants, study recruiters 18–24 2 telephoned specialists randomly selected 25–44 52 from public lists to request their partici 45–54 38 pation in the study. Eligible participants 55 or older 7 had spent at least 30 percent of their time Race conducting restaurant inspections and had White 86 worked in their current positions for at least six months. To minimize the possibil Black 10 ity that participants might know one an Other 5 other, only two participants from any state Hispanic or Spanish origin 7 and only one participant from any agency Education were scheduled in a group together. Each Some postsecondary education (e.g., associate’s degree) 2 group session lasted approximately two 4-year college degree 74 hours, and participants received a $60 in centive for their participation. Postgraduate degree 24 During the focus group meetings, partici Employer pants discussed various topics associated with City/township 17 their restaurant food safety activities. This County/district 57 paper covers only the discussions of partici State 19 pants’ investigation practices for foodborne Other 7 illness outbreaks, the methods they used to identify contributing factors, their success in Certifications* identifying contributing factors, and the dif NEHA Registered Sanitarian or Environmental Health Specialist 31 ficulties they faced when conducting investi NEHA Certified Food Safety Professional 7 gations. The meetings were audio taped, the Registered in state 57 transcripts of the audiotapes were reviewed, FDA standardization 36 and common themes among responses were identified. Participants also completed a short Years working in environmental health survey on their personal characteristics (e.g., 1–5 32 education and certification). 6–10 19 The study protocol was reviewed and ap 11–15 15 proved by CDC’s Institutional Review Board. 16–20 17 >20 17 results Years working in food programs Participant Characteristics 1–5 34 Each focus group had five to eight partici 6–10 22 pants, for a total of 42 participants. Of the 11–15 19 participants, 50 percent were male, 52 per 16–20 10 cent were between 25 and 44 years of age, >20 15 86 percent were white, and 98 percent had a four-year college or a postgraduate degree Position (Table 1). Over half worked for a county or Sanitarian/specialist 83 district and a third were National Environ Manager or senior sanitarian/specialist 17 mental Health Association (NEHA) regis

* These figures total more than 100 percent because participants could answer yes to all questions. tered sanitarians or environmental health specialists. Sixty-eight percent had worked in

Reprinted with permission from NEHA January/February 2008 • Journal of Environmental Health 17 FEATURES

REFERENCES

FIGURE

the environmental health profession for more than five years, and 66 percent had worked TABLE 2 in food programs for more than five years. Seventeen percent identified themselves as Outbreak Investigation Difficulties Identified by Participants managers or senior sanitarians/environmen tal health specialists. Difficulty Category Specific Difficulty Restaurant employees Lack of cooperation Foodborne Illness Outbreak Investigation Practices Restaurant customers Contact constraints About half the participants had been involved SPECIAL REPORLackT of cooperation in foodborne illness outbreak investigations. Lack of knowledge Those who had not participated in an out break investigation said either that their juris Organization Lack of epidemiologic assistance or a team approach diction had not had any outbreaks since their Lack of training and experience in outbreak investigation employment there; that outbreak investiga Lack of support from environmental health management tion was not a priority in their jurisdiction, so Lack of cooperation between agencies they did not conduct outbreak identification Lack of staff or investigation activities; or that environ mental health personnel in their jurisdiction Other Lack of physician cooperation conducted only routine restaurant inspec Notification delay tions in outbreak investigations because epi demiology personnel were responsible for all other investigation activities. Participants who had engaged in outbreak Some participants indicated that during an either that they did not have access to epide investigations described several investigation outbreak investigation they tried to identify miology personnel or that little or no collabo activities, including conducting restaurant as many customers of the outbreak restaurant ration occurred between epidemiologists and inspections and investigations, identifying as possible by culling pertinent information specialists (e.g., “When an outbreak occurs, customers in outbreak restaurants, develop from restaurant credit card receipts and inter … they just send us out to do the inspection ing and administering food history question viewing known customers of the restaurant. and don’t keep us informed….”). naires, collaborating with epidemiologists, Most said they played some role in developing Several participants said they conducted conducting epidemiologic analyses, taking questionnaires to be administered to restau epidemiologic data analyses, such as iden stool or food samples, and collaborating rant customers and sometimes workers. These tifying foods that ill people had consumed with public health nurses. Most indicated questionnaires were designed to help identify in common. Some said they tried to get ill that during an investigation they conducted ill people, symptom characteristics, and foods people to provide stool samples for pathogen inspections in the restaurant associated with linked to the outbreak (i.e., the vehicle). identification, and a few took food samples the outbreak. Several participants said that Many participants said they collaborated for the same reason. Some participants indi they conducted “routine inspections” during with epidemiology personnel during out cated that during an outbreak investigation outbreak investigations. Many, however, said break investigations, although the nature they worked with public health nurses, who outbreak inspections differed from routine of this collaboration varied across jurisdic were primarily involved in conducting inter inspections. Some said outbreak inspections tions. Some participants said they worked views with ill people. focused on the suspected vehicle and the food closely and collaboratively with epidemi handling practices associated with that vehi ologists throughout the investigation (e.g., Identifying Contributing Factors cle; some said outbreak inspections focused “Our communication is very good, and we When asked to describe their methods for iden on the food handling practices most strongly have a very strong team approach”). In some tifying contributing factors associated with out associated with foodborne illness in general jurisdictions, epidemiologists and special breaks, participants tended to focus on identify (often referred to as risk factors), and some ists had very defined investigation roles. For ing pathogens rather than contributing factors. said outbreak inspections focused on talking example, in one jurisdiction, epidemiolo Often participants specifically discussed con to employees to identify those who might be gists interviewed restaurant customers, while tributing factor identification only after prob ill. Similarly, many outbreak inspections fo specialists conducted the restaurant inves ing by the moderator. Participants’ comments cused less on noncritical violations (regula tigation. In other cases, although special suggested that contributing factor identifica tion violations not considered to be strongly ists conducted most investigation activities, tion was often dependent upon identification related to foodborne illness), such as whether they involved epidemiology personnel if the of vehicle, pathogen, or both. Participants pri the floors were clean. Some participants used epidemiologists’ expertise was needed (e.g., marily discussed three sources of information hazard analysis critical control point (HAC to determine pathogen incubation periods). used to determine the pathogen, vehicle, and CP) system terms, such as environmental Participants’ comments also indicated varia contributing factors—illness characteristics, investigations, food flows, and food prepara tion in which group led investigations—in epidemiologic analyses, and restaurant inves tion reviews, and a few participants referred some jurisdictions, epidemiology personnel tigations. Most said illness characteristics were specifically to the HACCP concept. led, and in others, specialists led. Some said often their first clue in an investigation. These

18 Volume 70 • Number 6 Reprinted with permission from NEHA characteristics provided information about the Investigation Difficulties management did not understand the impor possible pathogen involved, which led to a Participants identified 11 factors that nega tance of investigations and therefore did not focus on foods associated with that pathogen tively affected outbreak investigations. We support them. One participant said that state (e.g., “Let’s say … the incubation duration was grouped these factors into four categories: agencies did not notify local agencies of out classic for Bacillus cereus; we would investigate restaurant employees, restaurant custom breaks in a timely manner, and another said the rice [with a focus on] cooling”). Some said ers, organizational, and other (see Table 2). that they needed larger staff to adequately epidemiologic analyses into foods ill people had Participants reported that restaurant em conduct investigations. consumed in common were helpful in focusing ployees did not always cooperate with out Participants also discussed two additional the investigation on suspected foods and the break investigations. Examples included difficulties—lack of cooperation from physi pathogens and practices associated with those employees refusing to answer questions cians and delay in learning of outbreaks. Some foods. Some participants indicated that the res truthfully or at all and managers coach participants indicated that physicians some taurant investigation sometimes provided clues ing employees on “correct” answers to times created difficulties during investigations to the pathogen and contributing factors. For food preparation questions. According to because, when asked to support the investiga example, an interview with a worker might re participants, employees sometimes do not tion by testing patients with appropriate ill veal information about unsafe cooking practic cooperate because they are afraid of nega ness symptoms, they did not always comply. es. Most often, however, information on illness tive consequences for themselves or their Several participants said that there was often characteristics, vehicles, and pathogens was employers, they doubt their food made a delay (anywhere from several days to several used to focus the restaurant investigation (e.g., anyone sick, or they are focused on at weeks) between the start of an outbreak and “If we’re looking at a Norwalk, we’re going to tempting to identify the individuals who their notification of the outbreak, and that this be emphasizing … handwashing; if it’s Salmo complained about their restaurant. Partici delay made it difficult to obtain accurate food nella we’ll probably be looking at their storage, pants indicated, however, that in general, histories and stool samples from ill people, preparation, and service of eggs or chicken”). restaurant employees were very coopera accurate information concerning the prepara tive in outbreak situations. tion of the suspected vehicle, and samples of Effectiveness of Investigations Participants identified four difficulties the suspected vehicle. According to partici When asked how effective their outbreak inves associated with restaurant customers— pants, notification delays occurred because ill tigations were in identifying contributing fac contacting customers, lack of cooperation, people wait to notify health departments of tors to outbreaks, participants again tended to obtaining food histories, and lack of knowl their illness and because departments or agen focus on identifying pathogens rather than con edge. According to participants, contacting cies aware of outbreaks do not always notify tributing factors. Participants said that in many customers was difficult because such at environmental health programs or specialists cases they were able to make an educated guess tempts were typically made during the day, in a timely fashion. about the pathogen involved in the outbreak, when many people were not home. Exam based on the vehicle and the incubation period; ples of customer lack of cooperation cited Discussion however, laboratory confirmation of the patho by participants included reluctance to dis Our findings reveal substantial variability gen was seldom obtained from stool or food cuss illness symptoms and to provide stool in the investigation of the activities of envi samples (e.g., “We might know from the incu samples. In addition, participants said peo ronmental health specialists with respect to bation period, but if we don’t have lab confirma ple cannot easily remember what they ate foodborne illness outbreaks. Of particular tion, we can’t say for sure.”). A few participants, several days before their illness, so obtain interest is the finding that while some spe however, said they were more successful—one ing accurate food histories is difficult. Sev cialists reported conducting routine inspec participant said that pathogens found in food eral participants indicated that the public’s tions during outbreak investigations, others and stool samples were matched in 95 percent lack of knowledge about foodborne illness reported activities such as collecting stool of outbreaks in their jurisdiction. Several par delayed outbreak investigations because and food samples, identifying ill workers, ticipants also indicated that identification of the investigators had to educate complainants and identifying unsafe food handling prac vehicle, and therefore the pathogen, was easier about foodborne illness before they could tices associated with suspected vehicles and with large outbreaks and outbreaks in situa collect the information they needed. For foodborne illness. Routine inspections are tions where there were limited menu choices, example, complainants frequently attribute not likely to identify the pathogens that cause such as catered events. their illness to the last meal eaten and thus outbreaks or the factors that contribute to When prompted to discuss contributing are resistant to discussing other meals. outbreaks, because inspections are typically factors rather than pathogens, participants Participants identified five organizational used only to determine regulatory compli said identifying contributing factors was dif difficulties associated with outbreak inves ance with food safety codes or laws. On the ficult and they were not often able to do so. tigations. A few participants said that dur other hand, activities with a different focus, According to participants, employee turn ing investigations, epidemiology and envi such as stool and food sample collection and over, uncooperative restaurant employees, ronmental health personnel did not work identification of ill workers and unsafe food and the time lapse between preparation of the together or that epidemiology personnel handling practices, are more likely to lead to vehicle and the investigation made it difficult did not provide assistance to environmen the identification of outbreak pathogens and to identify the preparation practices associ tal health. Some said they did not have ad contributing factors, thereby furthering our ated with the suspected vehicle and to deter equate training or experience in outbreak understanding of the causes of outbreaks. mine whether ill workers had been involved investigation. In one jurisdiction, outbreak Participants identified several investigation in food preparation. investigations were not conducted because difficulties associated with restaurant employ-

Reprinted with permission from NEHA January/February 2008 • Journal of Environmental Health 19 ees, restaurant customers, and physicians. To Participants also identified lack of train the Electronic Foodborne Outbreak Reporting address the difficulties posed by uncooperative ing and lack of management support in System (EFORS), a Web-based system that al employees, investigators should try to conduct outbreak investigation as difficulties. Once lows state health departments to report data their restaurant investigation as early as possi the role of environmental health is defined, from foodborne illness outbreaks electronically ble in the process, because anecdotal evidence as suggested above, environmental health (CDC, 2000; CDC, 2005). This system includes suggests that employees and managers are more programs will require resources, training, information on contributing factors, and CDC likely to be cooperative at this point. Difficul and management support to fulfill their out is currently working with state health depart ties associated with restaurant customers, such break investigation roles. ments to increase the reporting of contributing as their lack of knowledge about foodborne ill Many participants said they did not often factors in this system. Fourth, CDC and NEHA ness and reluctance to cooperate with investi identify contributing factors during investiga have developed Epi-Ready Team Training, a gations, may be alleviated by foodborne illness tions. Three factors may influence this lack of nationwide initiative that provides training on education programs (Fein, Lin, & Levy, 1995; identification. First, identifying contributing foodborne illness outbreak investigations to Green et al., 2005). Similarly, difficulties associ factors did not seem to be a priority—many state and local environmental health programs ated with physicians may also be alleviated by participants said they focused on identify (NEHA, 2006). education of physicians about foodborne ill ing pathogens rather than contributing fac Qualitative focus group studies are useful ness diagnosis and investigation. The develop tors. Second, although some participants saw for identifying issues of concern for specific ment of CDC’s physician primer on foodborne pathogen identification as a prerequisite to populations, as our study has done. Data illness diagnosis is an important step in this identification of contributing factors, they also from such studies should not, however, be educational process (CDC, 2004). said that pathogen identification did not hap interpreted in a quantitative sense or general Several findings suggest that collabora pen very often. Third, participants indicated ized to a larger population. Thus, although tion and communication between epide that barriers such as uncooperative restaurant we can be fairly certain of the existence of miology and environmental health pro employees and delay in outbreak notification outbreak investigation practices and diffi grams is an important issue for outbreak impede the identification of contributing fac culties identified by the participants in our investigation. Participants reported a lack tors. Contributing factor identification is criti study, we are unable to determine the ex of epidemiologic assistance and variability cal to improving food safety because it leads tent of these practices and difficulties in the in the amount and nature of the collabora to the development of effective interventions. specialist population. Future research will tion between epidemiology personnel and Earlier suggestions––improving collaboration be needed to make this determination. Our environmental health personnel during between environmental health and epidemiol discussion of this study’s findings should be investigations. In addition, participants ogy, defining the role of environmental health reviewed with these facts in mind. identified the delay between the time when in investigations, and providing training and an outbreak occurs and when environmen management support––should increase the Acknowledgments and disclaimer: The authors tal health personnel learn of the outbreak ability of environmental health to identify wish to thank Sheryl Cates and Katherine Kosa as an investigation difficulty. This delay is contributing factors during investigations. (Food and Agricultural Policy Research, RTI likely to be due at least partially to a lack of CDC is participating in several ongoing ef International) for their assistance with study communication and collaboration among forts to improve investigation and reporting of design and participant recruitment, and the agencies and programs. Both epidemiol foodborne illness outbreaks. First, CDC’s EHS- EHS-Net Working Group (National Center for ogy and environmental health bring valu Net is working to improve identification and re Environmental Health, CDC) for their guid able knowledge and skills to the outbreak porting of contributing factors among the EHS- ance with respect to study questions. The find investigation process. Improvement in col Net states. From this work, training resources ings and conclusions presented in this paper laboration between these two programs and model data collection instruments will be are those of the authors and do not necessarily requires developing and defining the role developed and shared with other environmen represent the views of the Centers for Disease that environmental health can and should tal health programs. Second, CDC has funded Control and Prevention. play in outbreak investigations. A clear the Enteric Disease Investigation Timeline understanding of the role of environmen Study (EDITS), a project focused on collecting Corresponding author: Carol Selman, Nation tal health and its importance in outbreak and analyzing data on foodborne illness com al Center for Environmental Health: RTI In investigations is needed to provide an ad plaint and outbreak investigation timelines, and ternational, Centers for Disease Control and equate basis for managers to ensure col developing recommendations for improvement Prevention, 4770 Buford Highway, MS F28, laboration between the two programs. (Hedberg, 2005). Third, CDC has implemented Atlanta, GA 30341. E-mail: [email protected].

NEHA Sustaining Members can post their URLs on NEHA’s Web site for free. To take advantage of this benefit, please e - mail your organization’s Web site Great address (URL) to [email protected]. Benef it We’ll do the rest! Reciprocal links are appreciated. To access the links on NEHA’s f or Sus t aining Member s Web site, simply visit us at neha.org and click on “Links.”

20 Volume 70 • Number 6 Reprinted with permission from NEHA FEATURES

REFERENCES

Bryan, F. (2002). Where we are in retail food safety, how we got to eaten outside the home as sources of gastrointestinal illness. Jour where we are, and how do we get there? Journal of Environmental nal of Food Protection, 68, 2184-2189. Health, 65, 29-36. Hedberg, C. (November, 2005). Public health and response coordi Centers for Disease Control and Prevention. (2000). Surveillance for nation: The state of the art. In Proceedings of the Institute of Food foodborne-disease outbreaks—United States, 1993–1997. Morbid Technologists’ First Annual Food Protection and Defense Research ity and Mortality Weekly Report, 49, SS-1. Conference. Retrieved June 1, 2006, from http://www.ift.org/ FCentersIGUR for DiseaseE Control and Prevention. (2004). Diagnosis and fooddefense/29-Hedberg.pdf. management of foodborne illnesses: A primer for physicians and Jones, T., Imhoff, B., Samuel, M., Mshar, P., McCombs, K., Hawkins, other health care professionals. Morbidity and Mortality Weekly M., Deneen, V., Cambridge, M., & Olsen, S. (for the Emerging In Report, 53, RR-4. fections Program FoodNet Working Group). (2004). Limitations Centers for Disease Control and Prevention. (2005). Preliminary to successful investigation and reporting of foodborne outbreaks: FoodNet data on the incidence of infection with pathogens trans An analysis of foodborne disease outbreaks in FoodNet catch mitted commonly through food—10 Sites, United States, 2004. ment areas, 1998–99. Clinical Infectious Diseases Supplement, 38, Morbidity and Mortality Weekly Report, 54, 352-356. S297-S302. Ehiri, J., & Morris, G. (1994). Food safety control strategies: A criti National Environmental Health Association. (2006). Epi-Ready team TABcal reviewLE of traditional approaches. International Journal of Envi training program. Retrieved June 1, 2006, from http://www.neha. ronmental Health Research, 4, 254-263. org/research/food_safety.html#team_training. Fein, S., Lin, T., & Levy, A. (1995). Foodborne illness: Perceptions, Olsen, S., MacKinon, L., Goulding, J., Bean, N., & Slutsker, L. 2000. experience, and preventive behaviors in the United States. Journal Surveillance for foodborne disease outbreaks—United States, of Food Protection, 58, 1405-1411. 1993–1997. Morbidity and Mortality Weekly Report, 49, 1-51. Food and Drug Administration. (2001). FDA’s recommended national Silverman, G. (2003). Introduction to telephone focus groups (Report retail food regulatory program standards. Retrieved June 1, 2006, prepared for Market Navigation, Inc.). Retrieved June 1, 2006, from http://www.cfsan.fda.gov/~dms/ret-intr.html. from http://www.mnav.com/phonefoc.htm. Green, L., Selman, C., Scallan, E., Jones, T., Marcus, R., & EHS-Net SPECPopulationIAL Survey RWorkingEPOR Group.T (2005). Beliefs about meals