In this era of prevention and health promotion, the health-related behaviors of medical professionals have drawn considerable attention and are a common focus of research. Today's physicians not only serve as providers of care for their patients, but also are expected to model the advice they impart.1,2 Addiction to chemical substances is a problem that may afflict any person, and it has long been a concern among physicians. Brewster2 found reference in the literature to concerns of physician addiction to alcohol, cocaine, and morphine as early as 1869. Multiple cross-sectional studies have since tracked the use of tobacco and alcohol among medical students and physicians. Although tobacco use among medical students and physicians has decreased steadily over the past 50 years, the patterns of alcohol intake in this group have remained relatively stable.3 The purpose of this study is to assess the prevalence of tobacco use and the patterns of alcohol consumption among 1996 graduating medical students using a cross-sectional survey.
A questionnaire was developed to assess social and health-related habits and experiences of fourth-year medical students. The questionnaire included sections from the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System Questionnaire. Topics addressed included overall general health status and practices, social and professional relationships, and demographics.
This instrument was placed in the campus mailbox of all graduating medical students at 8 US medical schools (4 public and 4 private), with 2 schools in each of the 4 regions commonly designated by the Association of American Medical Colleges for geographic categorization. A randomization scheme was not used. Questionnaires required less than 10 minutes to complete and were returned individually in unmarked, stamped envelopes provided to students. No incentive was offered for participation in the study and no individual follow-up was possible. Data were collected between April and June of 1996. Statistical analyses were conducted using SPSS (version 6.1, SPSS, Inc).
A total of 1001 questionnaires were placed and 548 were returned (55%), with individual schools' response rates ranging from 44% to 69%. Of the returned questionnaires, 61% were from students at public universities and 39% from students at private institutions. The sample included a greater proportion of males (56%) than females (44%), and more whites (73%) than other racial groups (27%) (Table 1). Respondents' ages ranged from 23 to 47, with a median age of 26 years (mean, 27.5).
Among graduating students, 2% reported currently being smokers, and 13% reported ever having been smokers. Differences in tobacco use as a function of age and gender were not statistically significant, with the exception of a history of previous smoking, which was more common with increasing age.
Frequent alcohol use (3 or more days a week) was reported by 18% of students, and 21% of students reported at least one episode of binge drinking (5 or more drinks in one sitting) in the past 30 days. Male students were significantly more likely than female students to exhibit both of these behaviors (P=.03 and P<.01). Frequent alcohol use was most common in whites and blacks and among older students. Eighteen percent of respondents consumed more than 2 drinks at each drinking session, and male students were more likely to report this drinking behavior than were female students (P=.01). White students were much more likely than any of the other racial groups to report binge drinking, and age groups did not significantly differ regarding binge drinking.
Eighteen percent of women and 11% of men (P=.02) believed that their alcohol intake increased in medical school. Asians and the "25 and younger" age category were the racial and age groups most likely to report an increase in alcohol intake during medical school.
Results from this survey suggest an important decline in the prevalence of tobacco use among medical students. The 2% prevalence of smoking in this sample is the lowest rate reported in the literature among a large sample of medical students and is indicative of a continued acknowledgment by medical professionals of the dangers of tobacco use.4 Although 20% of the students in the oldest age group reported ever having smoked, this was true of only 8% in the youngest age group. Considering that few individuals begin smoking after age 18, this is an important marker for a much decreased use of tobacco among medical students.5
The preventive care beliefs and practices of health professionals are noted by the general public and may influence patient behaviors.1 For this reason, the decline in physician smoking in recent years has important repercussions for the nation's health. The 1979 Surgeon General's report on smoking and health concluded that between 10% and 25% of smokers who are advised to quit by their physician may quit smoking or reduce the amount they smoke.6 Several subsequent studies demonstrated a relationship between physician advice concerning tobacco consumption and the subsequent use of tobacco by patients.7- 9
Although the prevalence of tobacco consumption by physicians has declined over the past half century, the patterns of physician alcohol intake have remained stable or shown a slight increase.10 In general, younger physicians consume alcohol with the same frequency and in the same amounts as their age-related peers.11,12 But as physicians age, their alcohol intake tends to increase, in contrast to the intake of the general population, which tends to decrease with age.10 Estimates of physician impairment due to alcoholism are as high as 10%, comparable to the estimated 12% in the general population.13,14 Seven studies conducted among medical students between 1977 and 1990 found that over this time period, the prevalence of alcohol abuse ranged from 7% to 17%.15- 21 Among this study's sample of 1996 senior medical students, 21% reported an episode of binge drinking within the last 30 days, and 18% averaged more than 2 drinks per drinking session. Binge drinking and heavy alcohol consumption are possible markers for alcohol abuse.
The patterns of alcohol intake for this sample of medical students are similar to those for the age-related general population. Men tended to drink more often and more heavily than their female peers and were significantly more likely to have engaged in an episode of binge drinking within the last 30 days. Interestingly, women were more likely than their male peers to report an increase in alcohol intake while in medical school, which supports the hypothesis of Flaherty and Richman22 that there is a gender convergence in the drinking patterns among medical students that results from the occupational hazards of medical school. These findings may indicate the unique pressures placed on female medical students compared with their male counterparts. Possible examples of these gender-specific pressures include cultural expectations of women regarding child-rearing and domestic responsibilities and the deemphasis of a woman's career compared with that of her male colleagues. Additionally, women in the medical field must compete in a profession still largely dominated by men.23
Results from this study must be viewed in light of the limitations of a self-report survey design. These include recall bias, reporting errors, and nonresponse bias. Though self-reported alcohol consumption has generally been shown to be a valid measurement, it is possible that underreporting of both alcohol and tobacco use may have occurred because of the negative associations with use of these substances.24,25 This study had a 45% nonresponse rate, introducing a nonresponder bias of unknown quantity into the results. Data collection for this study occurred toward the end of the last year of medical school, which may have artificially inflated the results of the alcohol intake questions, especially those related to binge drinking. Finally, this questionnaire was not distributed in a strictly random fashion and its reliability and validity were never completely assessed.
Despite these limitations in the study's comparability with other studies, it provides key information on a large, geographically distributed sample of medical students whose demographics are similar to the enrollment of all US medical students with regard to gender, age, race, and enrollment patterns.26 Based on published studies concerning the validity of substance use questionnaires24,25 and given the educational level and medical background of the study sample, there is a high likelihood that the results are a true representation of actual substance intake.
Mangus RS, Hawkins CE, Miller MJ. Tobacco and Alcohol Use Among 1996 Medical School Graduates. JAMA. 1998;280(13):1192-1195. doi:10.1001/jama.280.13.1192-JMS1007-4-1