A gastroenteritis outbreak due to norovirus associated with a Colorado hotel.

Introduction

Recent epidemiologic investigations into outbreaks of norovirus (previously called Norwalk-like virus) on cruise ships and at conference facilities have had difficulty implicating a vehicle of transmission (Centers for Disease Control and Prevention [CDC], 2002; Love, Jiang, Barrett, Farkas, & Kelly, 2002). The virus is spread from one infected person to another by direct contact; aerosol; or contaminated objects, food, or water (Chin, 2000; Eddy, Ingram, & Leever, 2000; Jaykus, 1999). A low infectious dose also facilitates rapid spread of the disease (Kapikian, Estes, & Chanock, 1996). Although the virus does not replicate in food, in water, or on environmental surfaces, it is extremely persistent in these situations (Arness et al., 2000; Eddy et al., 2000; Jaykus, 1997, 1999). Of 348 norovirus outbreaks reported to CDC between 1996 and 2000, food was implicated in 39 percent, person-to-person contact in 12 percent, water in 3 percent; in 18 percent no transmission route was identified (CDC, 2001). Characterizing the environmental practices and the epidemiologic and laboratory findings that occur in these scenarios is important in identifying factors and practices that will aid in understanding the nature of these outbreaks.

This case study provides an example of how characterizing environmental practices and epidemiologic and laboratory findings can enhance norovirus outbreak research. On May 24, 2000, a local hotel notified Tri-County Health Department (TCHD) in Englewood, Colorado, that numerous individuals from the same corporation attending two professional meetings at the hotel had become ill. The attendees complained of nausea, vomiting, and diarrhea. Seventy-nine people from the Denver metro area had attended a one-day professional meeting held on May 22, 2000, and 54 people from out of state attended a second professional meeting that began on the evening of May 22, 2000, and ended on May 25, 2000. Although the Denver group and the out-of-state group worked for the same corporation, they attended separate meetings at the hotel and had little or no interaction with each other. No other meetings or banquets were held at the hotel during this time. The hotel prepared buffet-style meals and provided snacks for both meetings.

Methods

Environmental Investigation

Upon notification of the outbreak, TCHD initiated an environmental investigation of the hotel banquet kitchen to identify and implement control measures that would prevent additional illnesses and to evaluate food-handling practices such as food storage, preparation, sanitizing, handwashing, glove use, and service practices.

TCHD requested assistance from the CDC Environmental Health Services Branch (CDC-EHSB) on May 26, 2000, to investigate the environmental precursors that precipitate foodborne illness in retail food establishments. During informal interviews, each hotel kitchen employee was individually asked questions to determine why actions or behaviors occurred in the preparation of food for the professional meetings. The kinds of information solicited daring the interviews are listed below (actual questions are given in parentheses):

* identification of the food worker's role in the kitchen (What is your job here?);

* identification of the food worker's past work experience and education level (What is your background and where have you worked previously?);

* identification of the food worker's food safety knowledge (Have you had any formal or informal food safety training? From what you know, what can you do to prevent foodborne-illness outbreaks?);

* identification of the working dynamics within the kitchen (What do you like about your job here? What do you dislike?);

* identification of the food worker's role in the preparation of the foods implicated in the outbreak (In connection with the foods prepared for the professional group on May 22 and 23, what was your job, what did you do?);

* identification of the flow of foods, steps involved, ingredients, and so forth (Tell us how you make the food item,);

* determination of the food workers' beliefs about foodborne illness, their experience with foodborne illness, and their opinions about outbreaks (Have you ever heard about foodborne outbreaks in places you have worked or anywhere around town?);

* determination of whether employees work when ill and whether they were ill just before or after the outbreak, as well as identification of family members who could be a source of the organism (Have you been ill in the last few weeks? Has anyone at work or in your family been ill? Do you have children, are they in child care, and have they been ill? What is the policy about sick leave here? What do you do when you are ill?); and

* identification of information indicative of system weaknesses (If ibis outbreak did occur because of something that happened here, why would it have happened? What do you think would have caused it?).

Epidemiologic Investigation

The purpose of the epidemiologic investigation was to identity, those who were ill and to determine the cause of illness. Menus were obtained for foods prepared by the hotel on May 22 and May 23 for the two professional meetings. Questionnaires were developed and administered to the Denver and out-of state groups by face-to-face or telephone interviews. Interviewees were asked to provide demographic information; state whether they were ill before, during, or after attending the meeting; give the time of onset, the symptoms, and the duration of illness; and identify what foods they ate.

Eight hotel employees who complained of illness also were interviewed. Four of the eight ill employees worked as banquet servers, and the other tour worked in capacities unrelated to the food department. In addition, a small sample group of hotel guests who were unrelated to the professional group meetings hut who had eaten at the hotel restaurant were interviewed to determine whether the outbreak had extended beyond participants in the two professional meetings.

Data collected from the interviews were entered into Epi Info version 6.04d (Dean et al., 1995). To analyze the risk associated with eating specific foods for each group, calculation of relative risk, 95 percent confidence limits, and multivariate analyses were performed with SAS[R] Version 8.2 for Windows 2000 (SAS Institute, Inc., 1999). Multivariate analysis was conducted with generalized linear models. Since the response variable is binary, and the outcome occurred frequently, the authors used a binomial distribution with a log link function to calculate relative risks, instead of calculating odds ratios. Possible over-dispersion relative to the binomial model was accounted for (McCullagh & Nelder, 1983). A case was defined as diarrhea, vomiting, or both experienced by a hotel guest or hotel employee on or after May 22, 2000.

Laboratory Investigation

TCHD made stool specimen collection kits available to guests, hotel employees, and food handlers for bacterial and viral assessment. Three stool specimens were collected from three ill out-of-state individuals. The stool specimens were submitted to the laboratory of the Colorado Department of Public Health and Environment and were examined for Salmonella, Shigella, Escherichia coli O157:H7, and Campylobacter. The stool specimens were then transported to CDC's Viral Gastroenteritis Section in Atlanta for detection of noroviruses with reverse transcription polymerase chain reaction (RT-PCR) testing and sequencing (Ando et al., 1995; Eddy et al., 2000; Glass et al., 1990; Jaykus, 1997).

Results

Environmental Investigation

The hotel kitchen had a staff of 13. Employees described the kitchen as being very busy during the week of May 21. Timecards were not available, and the staff gave conflicting accounts of when co-workers had worked and who was assigned to prepare specific food items. Kitchen employees did not report being ill before the event. One employee, however, indicated that his children had been ill with flulike symptoms, but it was unclear when they had been ill or what their symptoms had been. Spanish was the native language for some staff members. The chef, who could speak English and Spanish, acted as interpreter for one kitchen employee during the informal interview.

Inquiries revealed that the kitchen staff had received minimal food safety training when hired. The executive chef was not knowledgeable about basic principles of food safety theory and practice. Only one kitchen employee had received formal food safety training. According to the kitchen manager, the hotel corporation did not require any food safety training. The only required training consisted of a safety-training booklet given to employees that covered blood-borne pathogens, hazard communication, fire safety, back safety, housekeeping, security, kitchen safety, and maintenance.

Several critical violations were observed in the kitchen over the course of the investigation. Colorado Retail Food Establishment Rules and Regulations defines a critical item as more likely than other violations to contribute to food contamination, illness, or environmental degradation (Colorado Department of Public Health and Environment, 1999). First, employees used knives for multiple tasks without cleaning and sanitizing them after each use. These tasks included the preparation of both raw meat and ready-to-eat products. Second, buckets storing wet wiping cloths for sanitizing in-place food equipment did not contain any measurable chlorine or quaternary ammonia residual. Third, employees who engaged in food preparation did not wash their hands between glove changes and used gloves for multiple tasks. Although employees did wear gloves during meal preparation, there was bare hand contact with some produce such as lettuce and tomatoes. Lettuce also was not washed prior to service. No food samples were available for environmental testing.

Epidemiologic Investigation

One hundred and thirteen (85.0 percent) of the 133 corporate attendees were interviewed. Interviews were completed for 61 (77.2 percent) of the 79 attendees in the Denver group and 52 (96.3 percent) of the 54 attendees in the out-of-state group. Twenty corporate attendees either refused to be interviewed or could not be located because of incomplete personal information. Forty (65.6 percent) of the 61 interviewees from the Denver group and 24 (46.2 percent) of the 52 interviewees from the out-of-state group met the case definition of having diarrhea, vomiting, or both on or after May 22, 2000. Other symptoms, such as fever, cramping, and headache, were reported less frequently.

An epidemic curve was constructed for both groups (Figure 1). The median duration of symptoms was 24 hours for both groups, with a range of five to 72 hours for the Denver group and six to 48 hours for the out-of-state group. The mean age was 43 years (SD = 7.8) for the Denver group and 45 years (SD = 8.9) for the out-of-state group. Males and females were similarly affected in the Denver group, but in the out-of-state group, females were disproportionately affected: 18 females (58.1 percent) were affected, compared with four males (23.5 percent). The eight ill hotel employees were not included in the analysis because many ate leftover food from various meals prepared at the hotel, and specific exposures could not be well assessed. Hotel guests who were interviewed but were not related to the professional meetings either exhibited no symptoms or had symptoms that were not consistent with this outbreak; therefore, they were not included in the analysis. No other reports of illness related to the hotel were received through passive surveillance. On the basis of the symptoms and their duration, viral gastroenteritis was suspected.

[FIGURE 1 OMITTED]

Multivariate analysis was unable to implicate a specific vehicle for the Denver group. Individuals in the out-of-state group who ate the garden salad (RR = 3.0, 95 percent confidence interval [CI] = 1.1-7.3) at dinner on May 22 had an increased risk of becoming ill. Consuming water on May 22 (RR = 2.1, 95 percent CI = 1.1-3.9) and May 23 (RR = 2.8, 95 percent CI = 1.2-6.6) also was significantly associated with illness for the out-of-state group. Multivariate analyses for the May 22 dinner did not demonstrate a multiplicative effect from consumption of water and garden salad. In addition, the model was unable to demonstrate whether consumption of the water or the garden salad was the better predictor of illness.

Laboratory Investigation

The presence of norovirus was detected by RT-PCR in all three stool specimens submitted by ill out-of-state individuals. The norovirus strain was determined to belong to genogroup I (GI) by sequence analysis of the amplified RT-PCR product.

Discussion

The laboratory investigation confirmed norovirus as the agent in this outbreak. The statistical analysis conducted in the epidemiologic investigation was unable to implicate a specific vehicle associated with disease transmission in either the Denver or the out-of-state groups. The inability to statistically implicate a vehicle in this outbreak is consistent with other norovirus outbreak investigations (CDC, 2002; Love et al., 2002; Peipins et al., 2002).

Although a vehicle could not be determined, the environmental investigation provided an opportunity to understand potential routes of norovirus transmission. Three identified areas included food-handling practices, food handler perceptions, and hotel policies. One of the documented food-handling practices was that lettuce used to prepare salads was not washed by the kitchen staff prior to service. Statements made by kitchen staff indicated that food handlers believed washing lettuce before service would cause it to wilt, thereby reducing the quality of the product. No other outbreaks or illnesses were reported in association with this brand of lettuce during this period. It was unlikely that the unwashed product was a contributing factor in the outbreak, but the practice of not washing lettuce could support norovirus transmission.

A second food-handling practice involved lack of handwashing; a likely cause of norovirus transmission is lack of handwashing by all asymptomatic food handler. Although the food handlers wore gloves, many reported that they neglected to change their gloves after completing each task and neglected to wash their hands between glove changes. There was also bare-hand contact with produce during salad preparation. Wearing gloves for an extended period of time and using gloves for more than one task can result in increased glove defects and the potential for cross-contamination of foods. The transfer of relatively small populations of Shigella, E. coli O157:H7, and viruses from hands to food presents a high risk of illness if the pathogens are not removed by adequate handwashing (Snyder, 1998). Information obtained during employee interviews suggests that other factors potentially contributing to the outbreak may have included lack of controls such as sneeze guards and supervision by hotel staff of patrons at the buffet line.

Hotel policies that may have enabled norovirus transmission included a lack of food safety knowledge and training among kitchen employees and a lack of corporate emphasis on food safety. There was no reporting policy requiring fond workers to notify their supervisor and cease food-handling activities when they are infected with a communicable disease that might be transmitted through food. Also, lack of adequate supervision resulted in critical violations and the potential for transmission of pathogenic microorganisms. Last, language barriers between employees and management might have hampered communication in areas such as food safety education or management expectations of employees' work performance.

Local health departments play a meaningful role in educating retail food establishments in the prevention of foodborne illnesses. Education alone, however, does not always result in application of the knowledge gained (Clayton, Griffith, Price, & Peters, 2002). Research has indicated that improved food safety behaviors and practices often are sustained only for short periods of time. This phenomenon may result from natural "set points" that cause the system to revert to its original food-handling pattern as a result of underlying circumstances. To create an atmosphere that is conducive to meaningful long term changes, local health departments should first emphasize assessment and understanding of food safety barriers within food establishments in order to discover hazards and develop appropriate educational efforts.

After the barriers and supporting circumstances at the hotel were identified, training was designed to address the needs of the food establishment. Primary emphasis was placed on correcting identified critical violations. Hands-on training was provided to all kitchen staff, who were shown physically how to wash hands, use latex gloves, and prepare sanitizing solutions. TCHD also provided a food safety training class to all kitchen staff. On subsequent visits to the kitchen, the health department required staff to demonstrate their food safety knowledge. Furthermore, recommendations were provided for refining the hotel's food safety policies and training procedures. After the outbreak, TCHD continued to monitor closely the hotel's food-handling practices to reinforce educational messages.

Conclusion

Even though no specific vehicle was identified in this outbreak, the investigation provided a meaningful opportunity to describe the environmental, epidemiologic, and laboratory circumstances that could have supported a norovirus outbreak By describing the practices and antecedents, TCHD was able to assist in strengthening the hotel's food safety program. Case studies such as this one, taken together, can identify practices and policies that may warrant additional emphasis by a food safety program and may provide information concerning the type of research needed to characterize norovirus transmission.

Acknowledgments: The authors would like to acknowledge David Olson, Ph.D., chief of the Biometry Branch, Division of Environmental Hazards and Health Effects, National Center for Environmental Health, CDC, in Atlanta. Georgia, who provided technical assistance in interpreting the multivariate analyses; Richard L. Vogt, M.D., executive director of Tri-County Health Department, who provided a critical review of the manuscript; and the Tri-County Health Department staff who played an integral part in resolving the outbreak.

REFERENCES

Ando, T., Monroe, S.S., Gentsch, J.R., Jin, Q., Lewis, D.C., & Glass, R.I. (1995). Detection and differentiation of antigenically distinct small round-structured viruses (Norwalk-like viruses) by reverse transcription-PCR and southern hybridation. Journal of Clinical Microbiology, 33(1), 64-71.

Arness, M.K., Barker, T.L., Canham, M.L., Ceislak, T.J., Cuthie, J.C., Fankhauser, R.L., Feighner, B.H., Hoedbeeke, E.L., Humphrey, C.D., Jenkins, C.D., Monroe, S.S., Polyak, C.S., Skillman, D.R., & Taylor, D.N. (2000). Norwalk-like viral gastroenteritis outbreak in U.S. Army trainees. Emerging Infectious Diseases, 6(2), 204-207.

Centers for Disease Control and Prevention. (2001). Norwalk-like viruses: Public health consequences and outbreak management. Morbidity and Mortality Weekly Report; 50(RR-9), 1-18.

Centers for Disease Control and Prevention. (2002). Outbreaks of gastroenteritis associated with noroviruses on cruise ships--United States. Morbidity and Mortality Weekly Report, 51(49), 1112-1115.

Chin, J., Ed. (2000). Control of communicable diseases manual (17th ed., Rev.). Washington, DC: American Public Health Association.

Clayton, D., Griffith, C., Price, P, & Peters, A. (2002). Food handlers' beliefs and self-reported practices. International Journal of Environmental Health Research, 12(1), 25-39.

Colorado Department of Public Health and Environment. (1999). Colorado retail food establishment rides and regulations. Denver, CO: Author.

Dean, A.G., Dean, J.A., Coulombier, D., Brendel, K.A., Smith, D.C., Burton, A.H., Dicker, R.C., Sullivan, K., Fagan, R.F., & Arner, T.G. (1994). Epi info, version 6: A word processing, database; and statistical program for public health on IBM-compatible microcomputers. Atlanta, GA: Centers for Disease Control and Prevention.

Eddy, C., Ingram, T., & Leever, J. (2000). The Norwalk virus: A guidance template for local environmental health professionals. Journal of Environmental Health, 63(2), 9-11.

Glass, R.I., Griffin, P.M., Juranek, D.D., LeBaron, C.W., Lew, J.F., Torok, T., Wahlquist, S.P., & Wells, J.G. (1990). Recommendations for collection of laboratory specimens associated with outbreaks of gastroenteritis. Morbidity and Mortality Weekly Report, 39(RR-14), 1-13.

Jaykus, L. (1997). Epidemiology and detection as options for control of viral and parasitic foodborne disease. Emerging Infectious Diseases, 3(4), 529-539.

Jaykus, L. (1999). Foodborne viruses: Emerging agents or emerging techniques? Dairy, Food and Environmental Sanitation, 19(9), 664.

Kapikian, A.Z., Estes, M.K., & Chanock, R.M. (1996). Norwalk group of viruses. In B.N. Fields, DM. Knipe, & P.M. Howley (Eds.), Fields virology (3rd ed., pp. 783-810). Philadelphia: Lippincott-Raven.

Love, S.S., Jiang, X., Barrett, E., Farkas, T., & Kelly, S. (2002). A large hotel outbreak of Norwalk-like virus gastroenteritis among three groups of guests and hotel employees in Virginia. Epidemiology and Injection, 129(1), 127-132.

McCullagh, P, & Nelder, J. (1983). Generalized linear models. London: Chapman and Hall.

Peipins, L.A., Highfill, K.A., Barrett, E., Monti, M.M., Hackler, R., Huang, P., & Jiang, X. (2002). A Norwalk-like virus outbreak on the Appalachian Trail. Journal of Environmental Health, 64(9), 18-23.

SAS Institute, Inc. (1999). SAS/STAT user's guide, version 8. Cary, NC: Author.

Snyder, P.O. (1998). Hand washing for retail food operations--A review. Dairy, Food and Environmental Sanitation, 18(3), 149-162.

Corresponding Author: Laura Dippold, Disease Intervention Coordinator, Division of Epidemiology, Planning, and Communication, Tri-County Health Department, 7000 E. Belleview Ave., Suite 301, Greenwood Village, CO 80111. E-mail: [e-mail].