Influenza-like symptoms in the college dormitory environment: a survey taken during the 1999-2000 influenza season.

Introduction

Influenza viral infections can disrupt the lifestyles of college students. The symptoms, often involving fever, malaise, or cough, may force students to miss classes and social gatherings, interrupting their educational pursuits and those of the university community. The incidence of influenza has been reported to be higher among dormitory residents than among students in non-dormitory settings (Pons, Canter, & Dolin, 1980). To identify aspects of dormitory life that may increase the risk of self-reported influenza-like symptoms (ILS), this study surveyed dormitory residents at the University of Chicago near the conclusion of the 1999-2000 flu season. To the authors' knowledge, it is the first epidemiologic study of ILS conducted among college dormitory students in 20 years, and the first to explore a comprehensive cross-section within a dormitory environment.

Background

Influenza viruses cause a seasonal respiratory illness that can spread rapidly from person to person throughout a community. Cough, fever, and sore throat are symptoms often associated with influenza, with illness typically lasting from one to two weeks in otherwise healthy individuals (Centers for Disease Control and Prevention [CDC], 2001). The mode of transmission is usually exposure to airborne respiratory secretions contaminated with the virus (Hayden & Palese, 2002). Influenza infection can be readily diagnosed and the disease well characterized, and symptoms, including fever, cough, and myalgia, have been found to correlate well with actual disease (Monto, Gravenstein, Elliot, Colopy, & Schweinle, 2000). Influenza also can be transmitted through an environmental-surface intermediate, as when an infected individual sheds the virus onto a surface and an uninfected, susceptible host later contacts the same surface and then self-inoculates the virus into his or her mucous membranes (Barker, Stevens, & Bloomfield, 2001). The most effective control against influenza is prevention of infection (Piedra, 1995). Increased knowledge about influenza transmission within college dormitories may aid in developing methods of preventing infection.

Survey

The authors designed and implemented a survey to study issues related to influenza. The survey questioned students about ILS symptoms and their college dormitory environment, a setting where factors such as roommates, room carpeting, and dining, laundry, and washroom environments could be evaluated. These variables were chosen because they represented sites where surface or air transmission was likely to occur. Gender and year in college also were assessed. To eliminate confounding variables, the survey asked students if they had received an influenza vaccine, an intervention known to reduce ILS risk (Piedra, 1995).

Methods

Prior to distribution of the survey, approval from the Institutional Review Board of the University of Chicago Hospitals was obtained, and the survey was pretested among a sample of students in the dorms surveyed. A questionnaire then was mailed to 1,704 dorm students, on February 26, 2000, and responses were collected in reply boxes at each dormitory through March 3, 2000. The seven-day reply period was near the end of a 10-week quarter that began January 4 and ended March 10. Surveying later in the quarter would have conflicted with final examinations, a factor likely to lower the response rate. Students were reminded to return the surveys through e-mails and flyers on dorm bulletin boards. Data were analyzed with SPSS 9.0 and Microsoft Excel 97. Two-sided p values were reported. Risks were expressed as proportions or percentages and risk ratios.

Case Definition

Influenza-like symptoms were defined as fever ([greater than or equal to]38[degrees]C for [greater than or equal to]48 hours) and at least two of the following symptoms lasting for at least two days: cough, myalgia (described as "aches and pains" on survey), sore throat, and congested nose. This definition was based in part on findings described by Monto and co-authors in 2000. The survey asked students to circle symptoms from a list, and to recall the date of onset of symptoms and the duration of the symptoms to the best of their knowledge.

Dormitories

The five largest dormitories in an 11-dormitory housing system were surveyed. The dorms represented a broad range of housing settings, and cost considerations limited the survey to only a partial sampling. The sizes of the dormitories ranged from 170 to 650 students. Survey distribution, reminder notices, and collection costs were identical for each dorm. In preparation for the survey, an overview of the university housing system showed that the dorms chosen covered a wide range of housing settings (single, double, and triple bedrooms, private and group baths, etc.). Students had not been randomly assigned to rooms.

Data Set Preparation

Altogether, 721 surveys were returned (42.3 percent gross response). Because the authors wanted to evaluate ILS that arose from viral infections incurred while students were residing in the dorms and not on vacation. ILS cases occurring before January 9 were excluded (34 cases). This cutoff date was chosen because students returned from a holiday break on January 4 and influenza has a one-to-five-day incubation period (Hayden & Palese, 2002). Students whose illness met the symptomatic definition of ILS but who could not recall the date the symptoms started also were excluded (15 cases). Nine other surveys were less than 50 percent complete and were therefore excluded.

The data set also excluded 123 respondents who indicated that they had received a flu vaccine for the 1999-2000 influenza season. Influenza vaccine is known to reduce and confound the risk for ILS. Thus, a total of 540 surveys were evaluated (31.7 percent analyzed responses).

Results

Virus

Influenza A/Sydney/05/97 (H3N2) was prevalent in the United States during the 1999-2000 season for the third consecutive year and was considered well matched to the circulating influenza viruses (CDC, 2000). The A/Sydney/05/97 influenza strain was determined to be the predominant strain in the South Chicago area (where the University is located) for the 1999-2000 flu season by the University of Chicago Hospitals Virology Laboratory.

Respondent Comparison to Housing Population

Comparisons between survey data and official housing data showed no significant differences by class level or gender between surveyed and nonsurveyed dorms, respondents and nonrespondents, and cases included or excluded as a result of ILS criteria. The gender and class level comparisons indicated that the survey responses were not biased toward a particular demographic factor and reflected the overall university housing population. No significant differences were found in ILS incidence by dormitory.

The survey data were collected during winter quarter 2000, and housing data from the University Housing Office were available only for spring quarter 2000; no official housing data were available for winter quarter 2000. Minor differences were found between the number of students enrolled during the winter and spring quarter (1,704 winter students versus 1,632 spring students in the dormitories surveyed, a 4.4 percent difference). These differences likely were due to graduations or transfers, and were considered to have little effect on the outcome of the comparisons.

Symptoms

The most frequent symptoms reported in the survey were nasal congestion (reported by 26.7 percent of respondents, or 144 of 540), cough (24.8 percent, or 134 of 540), and sore throat (25.7 percent, or 139 of 540). Table 1 shows all symptoms reported on the surveys. Symptoms that might be related to flu were reported to occur after January 9 by 238 (44.1 percent) of the respondents, including 61 (11.3 percent) whose symptoms met the case definition of ILS (Table 1). Among the students whose symptoms met the definition of ILS, the average duration of symptoms was 4.7 days (95 percent CI: [+ or -]0.85; range: 2-12 days), and an average of 3.7 class sessions were missed (95 percent CI: [+ or -]0.79). Respondents who missed class because of ILS did not differ significantly by gender or year in college.

Dormitory Rooms

As indicated in Table 2, students with [greater than or equal to]50 percent carpeting on the room floor were at a significantly lower risk for ILS (p = .02). Respondents in doubles who indicated that their roommate slept in a different room reported a lower number of ILS cases (2.4 percent, or 1 of 41) than respondents who slept in the same room with their roommate (10.4 percent, or 27 of 259) (p = 0.10, relative risk [RR] = 4.3). In triples, all beds are in the same room, preventing a similar analysis. The number of self-reported ILS cases increased as the number of roommates increased, but the difference was not statistically significant.

Other Settings

The frequency with which the dining hall was attended or laundry was washed had no significant effect on ILS frequency (Table 3). There also were no significant differences in ILS incidence by gender or year in college. The risk for ILS in private washrooms, defined as four or fewer students per washroom, did not appear to be related to the use of coed versus single-gender washrooms. Nor did the risk appear to be related to the use of private versus communal washrooms (Table 4).

Discussion

This cross-sectional survey of the college dormitory environment highlights the factors associated with the risk of self-reported ILS. The results demonstrate the importance of the living environment in ILS transmission in comparison with other factors such as dining setting, laundry habits, washroom settings, and demographics, including age and gender.

Students with [greater than or equal to]50 percent carpeting in their room reported significantly fewer ILS cases than those with less carpeting (p = .02). The reason for the association of carpeting and a lower ILS risk is not readily clear. The ILS risk was not affected by whether students installed their own carpeting or the carpeting was installed by the university. The ability of the survey instrument to detect this significant finding, however, is of primary interest. Relations strong enough to result in positive associations in this study may suggest areas where transmission of ILS in a dorm is likely to happen.

A possible interpretation of the relation between carpeting and reduced ILS risk may come from a study by Bean and co-authors (1982), which reported that Influenza A viruses can survive 24 to 48 hours on hard, nonporous surfaces such as stainless steel or plastic, but only <8 to 12 hours on cloth, paper, and tissues. In addition, the Influenza A virus could be transferred from stainless-steel surfaces to hands for up to 24 hours, but from tissues to hands for only 15 minutes. Bean and co-authors determined that fomites were likely to be the mode of transmission. Many other respiratory viruses are known to behave similarly on these environmental surfaces (Barker et al., 2001). Carpet has a clothlike surface, and a tiled floor is often a hard, nonporous surface, with a potential for prolonged viral survival. This difference could potentially explain the higher number of ILS cases observed among dorm rooms with [less than or equal to]50 percent carpeting (p = .02).

Although the risk of ILS appeared to increase for roommates in doubles who slept in the same room compared with those who slept in different rooms (RR = 4.3), the difference was not statistically significant. This finding does not necessarily distinguish between air and surface transmission. Physical separation between sleeping roommates can both lessen the exposure to air-borne respiratory secretions and decrease contact with room surfaces or fomites harboring the influenza virus. Relations between ILS risk and other dorm settings were not as strong, suggesting that ILS transmission is more likely to occur in the room environment than in other settings with less potential for contact among students.

While trends were observed with respect to room carpeting and sleeping arrangements, no relation was found between ILS risk and the frequency with which dining halls were attended, the frequency with which laundry was done, or a student's washroom setting. The following factors may contribute to these results. Dining halls, washrooms, and laundry facilities are cleaned regularly by dormitory staff, while the student rooms are not. Irregular cleaning in student rooms may lead to a higher risk for ILS. Furthermore, less time is spent at meals or in the washroom than in rooms studying or socializing, and increased time exposure could increase the risk of transmission. Thus, the dormitory room itself may place students at increased risk for ILS.

This study has limitations. The survey does not provide laboratory confirmation of influenza disease. Results from this survey format, however, are likely to be less biased in coverage and more efficient in gathering data with a large population. Also, to strengthen the case definition for ILS, the authors derived their symptom-based definition from previous studies in which influenza was confirmed in the laboratory. Boivin, Hardy, Tellier, & Maziade (2000) showed fever and cough symptoms to be significantly associated with a sensitive laboratory test for influenza, the polymerase chain reaction (PCR) technique used to amplify nucleic acid sequences in vitro. The case definition used in the study reported here is based on their results and those of Monto (2000), and is stricter than that used in previous studies examining influenza on college campuses (Sobal & Loveland, 1982). The authors recognize that ILS cases in the survey population may not have been detected by the survey, but the definitions utilized were designed to yield higher specificity for influenza disease.

A well-matched influenza vaccine and the presence of the A/Sydney/05/97 influenza strain for the third consecutive year may have contributed to the low number of ILS cases found by the survey. Findings that arise from a small sample size, however, may signal strong correlations that warrant further investigation. Finally, factors not considered in this study may have an effect. Interaction in classrooms or social activities outside the dormitory could contribute to the risk of ILS. Because influenza virus vaccines and antivirals are widely available, an understanding of the epidemiology and consequences of disease in this population would be valuable in recommending potential further intervention, such as more widespread vaccination. Further work is needed in larger populations and on different campuses to further understand the epidemiology of ILS and the true effects of dormitory conditions on ILS transmission.

TABLE 1 Incidence of Symptoms Reported

Symptom              ILS           Other
                  %      Number  %      Number

Fever             11.3   61       1.3     7
Sore throat       10.4   56      25.7   139
Fatigue           10.2   55      24.1   130
Headache           9.6   52      22.2   120
Congested nose     9.6   52      26.7   144
Cough              9.4   51      24.8   134
Myalgia            9.1   49      17.6    95
Chills             8.7   47      10.7    58
Loss of appetite   6.9   37      12.4    67
Dizziness          6.5   35      11.7    63
Vomiting           0.93   5       0.93    5
Diarrhea           0.37   2       0.19    1

N = 540 surveys.

TABLE 2 Incidence of ILS According to Dormitory Room Factors (a)

Physical Factors
                                 Yes           No         p    RR
                               %    Number  %     Number

Carpet covers
[greater than or equal to]50%
  of room floor                7.6  25/330  14.4  25/174  .02  1.9
Beds in separate rooms in
  double                       2.4   1/41   10.4  27/259  .10  4.3

Room Size
                               %     Number               p    RR

Single                        10.1   24/238               .17  1.7
Double                        11.7   28/239               --   --
Triple (b)                    17.5    7/40                --   --

(a) Not all proportions come to 61 of 540 because not all students
answered every question.
(b) Only 1/30 quads fit the ILS definition. Quads were excluded from
further analysis.

TABLE 3 Incidence of ILS According to Dining and Laundry Habits (a)

        Dining Attendance             Laundry Habits
Frequency (b)  % ILS  Number  Frequency (c)  % ILS  Number

 0              0.0    0/9     0             11.8    2/17
 1-4           10.9    6/55    1-3           14.4   21/146
 5-8           12.3   10/81    4-6           10.5   26/248
 9-12          11.9   22/185   7-9            6.2    4/65
13-16          13.5   21/155  10-12          10.9    6/55
17-20           3.8    2/53   13-15          16.7    1/6

(a) Not all proportions come to 61/540 because not all respondents
answered all questions.
(b) Average number of meals per week of the 20 meals served by dining
services.
(c) Average number of times a respondent did laundry during the eight
weeks between the start of the quarter and the time of the survey.
(Laundry machines are located in every dorm in the housing system.)

TABLE 4 Incidence of ILS According to Washroom Settings (a)

Setting   % ILS  Number

Private   10.5   16/152
Communal  11.4   44/385
Coed      12.0   31/258
Single    10.4   29/279

(a) Private washrooms are shared by [less than or equal to]4 students
and consist of a sink, shower, and commode. Washrooms shared by >4
students are communal.

Acknowledgements: The authors are grateful to the following University of Chicago faculty and staff: Ted L. Steck, M.D., Chair of Environmental Studies, for his review of the manuscript, and Sherry Gutman, Associate Dean for Student Services, and Katie Nash, Dean of Students in the College, for their assistance with the surveys. The authors also thank the Women's Board and the Office of the Dean of the College for providing funding.

REFERENCES

Barker, J., Stevens, D., & Bloomfield, S. (2001). Spread and prevention of some common viral infections in community facilities and domestic homes. Journal of Applied Microbiology, 91, 7-21.

Bean, B., Moore, B.M., Sterner, B., Peterson, L.R., Gerding, D.N., & Balfour, H.H. (1982). Survival of influenza viruses on environmental surfaces. Journal of Infectious Diseases, 146(1), 47-50.

Boivin, G., Hardy, I., Tellier, G., & Maziade, J. (2000). Predicting influenza infections during epidemics with use of a clinical case definition. Clinical Infectious Diseases, 31(5), 1166-1169.

Centers for Disease Control and Prevention. (2001). Influenza general information. http://www.cdc.gov/ncidod/diseases/flu/fluinfo.htm (25 Oct. 2001).

Centers for Disease Control and Prevention. (2000). Centers for Disease Control update: Influenza activity--United States and worldwide, 1999-2000 season. Morbidity and Mortality Weekly Report, 49(17), 375-381.

Hayden, F.G., & Palese, P. (2002). Influenza virus. In Clinical Virology (2nd ed.). (pp. 891-920). Washington, DC: ASM Press.

Monto, A.S., Gravenstein S., Elliot M., Colopy M., & Schweinle J. (2000). Clinical signs and symptoms predicting influenza infection. Archives of Internal Medicine, 160(21), 3243-3247.

Piedra, P.A. (1995). Influenza virus pneumonia: Pathogenesis, treatment, and prevention. Seminars in Respiratory Infections, 10(4), 216-223.

Pons, V.G., Canter, J., & Dolin, R. (1980). Influenza A/USSR/77 (H1N1) on a university campus. American Journal of Epidemiology, 111(1), 23-30.

Sobal, J., & Loveland, F.C. (1982). Infectious disease in a total institution: A study of the influenza epidemic of 1978 on a college campus. Public Health Reports, 97(1), 66-72.

Wayne M. Tsuang

John C. Bailar, M.D., Ph.D.

Janet A. Englund, M.D.

Corresponding author: Janet Englund, Pediatric Infectious Diseases, Children's Hospital and Regional Medical Center, University of Washington, 4800 Sand Point Way, NE, #8G-1, Seattle, WA 98105. E-mail: [e-mail].