Following each scenario’s presentation, a participant was asked to indicate on a 5-point Likert-type scale (1=very likely; 5=very unlikely) how likely he or she was to seek care from a physician for the illness and the likelihood of using several treatment modalities (eg, antibiotics, antihistamines, pain relievers, vitamin C) for the presented condition.
Participants were also queried about a variety of demographic variables, current and past smoking status, and their belief in the effectiveness of antibiotics against the common cold. Finally, participants were asked to indicate whether they would see a physician if they had a cold (no specific symptoms were provided to define “cold”). The questionnaire was designed for self-administration and required less than 5 minutes to complete.
Analysis
Likert-based responses on the questionnaire were dichotomized by combining “very likely” and “somewhat likely” into 1 group and “neutral,” “somewhat unlikely,” and “very unlikely” responses into another. These responses and the yes or no responses were analyzed using chi-square tests.
The likelihood of seeking care and using antibiotics to treat a cold were averaged across scenarios. We used bivariate analyses to examine the relationship of demographic characteristics to averaged likelihood of seeking care and antibiotic use.
Stepwise multiple regression analyses were employed to examine the effects of participant characteristics on likelihood of seeking care and likelihood of using antibiotics. Because the regression analysis for each individual scenario produced similar results, we only report the analysis employing averaged likelihood of seeking care and averaged likelihood of using antibiotics. Participant demographic characteristics (continuous and discrete) that were included as independent variables in these analyses were sex, race, age, type of health insurance, and year of college. In analyzing likelihood of seeking care we included the additional variable of belief in the effectiveness of antibiotics in treating the common cold as an independent variable. We performed an additional stepwise multiple regression analysis using belief in antibiotic effectiveness as the dependent variable and participant demographic characteristics as the independent variables.
Results
The demographic characteristics of the total sample and each recruitment site are provided in Table 1. In response to a free-format question on the survey, 24% of students in the total sample reported enrollment in a science-related field of study.
Responses to the questionnaire were first examined by campus. Although the campus samples differed significantly in some ways (eg, sex and racial distributions, average age, year in college, type of health insurance), no significant differences existed between campuses in terms of the study’s variables of focus. Therefore, we conducted all analyses on the total sample.
Antibiotic Effectiveness
Forty-one percent of the total sample believed that antibiotics were effective for treating the common cold Table 2. Of those who reported a belief that antibiotics were effective for cold treatment, 24% would see a physician for a cold (10% of the total sample). Of those who did not believe in the effectiveness of antibiotics for treating colds, 12% would still seek a physician’s care for a cold (7% of the total sample).
Symptom Complex Analysis
Analysis of antibiotic use by symptom complex Table 3 revealed that 63% of all students would be “somewhat likely “or “very likely” to use antibiotics in the scenario of 5 days with discolored discharge, compared with 53% in the 3 days with clear discharge with low-grade fever scenario, and 44% in the scenario of 5 days with clear discharge (P <.001 for each comparison). Percentages were higher for those students who believe antibiotics are effective in treating the common cold: 77%, 73%, and 64%, respectively. Even among those students reporting that antibiotics are not effective in treating colds, a high number of students endorsed “somewhat likely” or “very likely” antibiotic use when presented with the 3 scenarios (32% for 5 days with clear discharge; 42% for 3 days with clear discharge and low-grade fever; 55% for 5 days with discolored discharge).
Participants’ beliefs about whether they would seek a physician’s care were also analyzed by symptom complex Table 4. Fifty-six percent of all students surveyed reported being “somewhat likely” or “very likely” to see a physician in the scenario of 5 days with discolored nasal discharge, which was a significantly greater percentage than those who would see a physician for 3 days with clear discharge and a low-grade fever (44%) and those who would see a physician if they experienced 5 days of clear discharge (29%). Percentages of those seeking medical help followed the same pattern among those students who believe antibiotics are effective in treating the common cold and those who do not believe in the effectiveness of antibiotics.