593

Medical Education

A Survey on the Prevalence of Alcoholism Among the Faculty and House Staff of an Academic Teaching Hospital
BARY J. SIEGEL, MD, and FAITH T. FITZGERALD, MD, Sacramento, California

We studied the extent of alcoholism among faculty and house staff of an urban, university-based teaching hospitaL Of 569 questionnaires sent, 282 (50%) were returned and 271 of these were complete enough to be interpretable. Of those responding, 12 (4%) were classified as alcoholic and 26 (10%) as possibly alcoholic. There was no statistically significant difference in the prevalence of alcoholism in physicians from the different medical specialties or in regard to gender. Nevertheless, with 14% of the respondents to our questionnaire being classified as either alcoholics or possible alcoholics, it appears that this is a pervasive problem in our profession thatdeserves further study.
(Siegel BJ, Fitzgerald FT: A survey on the prevalence of alcoholism among the faculty and house staff of an academic teaching
hospital. West J Med 1988 May; 148:593-595)

In the last decade a great deal has been written about the problem of alcoholism among physicians. Very little is known about the actual prevalence of this problem, however. A 1986 review of the literature on alcoholism and drug abuse among physicians suggests that the true prevalence of alcoholism in this group is unknown.1 To date there have been only two published reports of surveys of physicians where standard measures of alcoholism were used. Both of these surveys, however, were administered only to subgroups of physicians that were not necessarily representative of physicians as a whole, and the prevalence of alcoholism reported in these studies differed markedly.23 Accurately assessing the magnitude of alcoholism in medical professionals, especially physicians, is clearly a matter of considerable importance. The definition, identification, and treatment of impaired physicians is of immediate concern to patients, regulatory boards, and society as a whole. To obtain more information on the prevalence of alcoholism among physicians, we surveyed a broad group of physicians.

Results
House staff and faculty were sent 569 questionnaires. Of these, 282 (50%) were returned, with 271 complete enough to be interpretable and scored in the manner described in the methods section. These constituted the final sample upon which we based our results. A total of 12 (4 %) of the respondents scored 3 or more points and were classified as alcoholic. An additional 26 (10%) respondents scored 2 points and were classified as possibly alcoholic. The remaining 233 respondents scored 1 or fewer points and were classified as nonalcoholic. Of the 271 respondents, 147 (54 %) identified themselves as men. Of these, 6 (4%) were alcoholic and 16 (11 %) were possibly alcoholic. A total of 43 (16%) of the respondents identified themselves as women; one (2 %) was classified as alcoholic and three (7%) were classified as possibly alcoholic. Of the 81 subjects who declined to state their sex, there were 5 (6%) alcoholics and 7 (9%) possible alcoholics (Table 1). There was no statistically significant difference in the prevalence of alcoholism among these three groups (X2 with 4 degrees of freedom (df = 1.8, P=.77). We also examined the influence of the respondents' medical specialty on the prevalence of alcoholism. The only specialties in which no one was identified as being either alcoholic or possibly alcoholic were physical medicine and rehabilitation and therapeutic radiology. The highest prevalence of alcoholism was found in the specialties of family practice and obstetrics and gynecology, where 14% of the respondents were identified as being alcoholic. No one identifying themselves as a member of the specialties of anesthesiology, physical medicine and rehabilitation, therapeutic radiology, or psychiatry was classified as an alcoholic. The highest prevalence of those classified as possibly alcoholic was found in the group containing pathology, radiology, and nuclear medicine, in which 36% of the respondents fell. Among those respondents who declined to state their spe-

Materials and Methods A slightly modified version of the Short Michigan Alcoholism Screening Test was sent to all faculty and house staff of an urban, university-based teaching hospital (Figure 1). The only notable modification of the questionnaire was the addition of a second component to question six, where respondents were asked whether their attendance at a meeting of Alcoholics Anonymous was part of a class assignment or a clinical rotation. Limited demographic information was also
requested. Each alcoholism-indicating response was assigned a one-point value. Respondents scoring zero to one point were classified as nonalcoholic. Those with two points were considered to be possibly alcoholic. Those scoring three or more points were classified as alcoholic. This system is similar to that used by Selzer and colleagues.4 The data were analyzed using the Catmod procedure in the statistical analysis system and a mainframe computer.5

From the Division of General Medicine, University of California Davis Medical Center, Sacramento. Reprint requests to Bary J. Siegel, MD, Department of Internal Medicine, Division of General Medicine, Primary Care Center, Rm 3120, 2221 Stockton Blvd, Sacramento, CA 95817.

594ALCOHOLISM

594

ALCOHOLISM IN PHYSICIANS

TABLE 1.-Effects of Gender on Prevalence of Alcoholism
Gender
Nonalcoholic, No. (96) Alcoholic, No (96)

Possibly

Alcoholic, No. (96)

Total, No. (96)

Men ............ 125 (85) 16 (11) 39 (91) 3 (7) Women ......... Decline to state .... 69 (85) 7 ( 9% Total ......... 233 (86) 26 (10)

6 1 5 12

(4)
(2) (6)

(4)

147 (54) 43 (16) 81 (30) 271

cialty, only 5 % were classified as alcoholic and the rest were classified as nonalcoholic (Figure 2). Despite the extreme variation in the prevalence of alcoholism among various departments, the null hypothesis that the mean of all of the departments is equal cannot be rejected (X2 with 18 df= 27.3, P= .07). Discussion Previous estimates of the prevalence of alcoholism among physicians have varied widely. Murray suggested that it is much higher than that in the general population,6 while other investigators have suggested with equal conviction that it is probably no greater than that of the general population.7'8 Our knowledge of the true extent of the problem is hampered by the application of an often imprecise definition of alcoholism, the surveillance of only highly selected subpopulations, such as hospitalized patients or those physicians in trouble with medical disciplinary or licensing boards, and, perhaps, the failure of physicians to make this diagnosis in their peers.19 Thus far, there have been only two reports in the literature wherein alcoholism-indicating surveys have been used to determine the prevalence of alcoholism in physicians. The first study was conducted on a group of physicians, most of whom were family practitioners, attending a continuing medical education conference. The Self-Administered Alcoholism Screening Test was completed by 399 physicians. By this instrument, the authors concluded that this group had 2 % alcoholics and 5 % possible alcoholics.2 These results are quite different from those published in the only other study in
1. Do you feel you are a normal drinker? (By normal, we mean you drink less than or as much as most people.) 2. Does your spouse, a parent, or other near relative ever worry or complain about your drinking? 3. Do you ever feel guilty about your drinking? 4. Do friends or relatives think you are a normal drinker? 5. Are you able to stop drinking when you want to? 6. Have you ever attended a meeting of Alcoholics Anonymous? If you answered yes, was this part of a class assignment or clinical

which an alcoholism-indicating survey was used. In this study, the Michigan Alcoholism Screening Test was sent to all of the house staff of the Presbyterian Hospital in New York. Of the 417 house staff surveyed, 188 responded. Of these, 13% had scores that indicated either suggested or presumed alcoholism.3 Based on these studies, the prevalence of dysfunctional drinking in physicians would appear to be anywhere from 7 % to 13%. There are many possible explanations for this disparity. In the Lewy survey, questionnaires were mailed, and the participants may have felt that their anonymity was more secure than that of physicians at a continuing education conference. Thus, they may have answered their questionnaires with greater candor, resulting in a higher number of identified alcoholics. Alternatively, the participants in the survey of house staff may have been much younger than those who attended the continuing education conference. If young physicians drink more heavily than older ones, this could also account for the difference in the prevalence of dysfunctional drinking in these two studies. Also, those who attend continuing educational conferences are a self-selected group. Perhaps the conference where the survey took place in the study by Niven and associates was such that it selected for nonalcoholic physicians, and the prevalence of alcoholism, therefore, was found to be lower than that in a random group ofhouse staff. To obtain more precise information on the prevalence of alcoholism in physicians, we decided to apply a structured, validated, alcoholism-indicating questionnaire to a broadbased group of working physicians in multiple disciplines over a wide age range. In our study of the house staff and faculty of a universitybased teaching hospital, 4% of the respondents were identified as being alcoholic and 10% as being possible alcoholics. These findings stand in marked contrast to the prevalence of alcoholism reported in the study of physicians attending a continuing education conference. The prevalence of dysfunctional drinking in our group is also higher than that reported in the house staff at Presbyterian Hospital.
80 Nonolcoholic

70*

Possibly Alcoholic Alcoholic

60
w

50-

:

4030 -

rotation?
7. Has drinking ever created problems between you and your spouse, a parent, or other near relative? 8. Have you ever gotten into trouble at work because of drinking? 9. Have you ever neglected your obligations, your family, or your work for two or more days in a row because you were drinking? 10. Have you ever gone to anyone for help about your drinking? 11. Have you ever been in a hospital because of drinking? 12. Have you ever been arrested for drunken driving, driving while intoxicated, or driving underthe influence of alcoholic beverages? 13. Have you ever been arrested, even for a few hours, because of other

20 10

drunken behavior?
Figure 1.-Modified
with modifications

Short Michigan Alcoholism
in bold-faced

Screening Test,

Dl Ob Pe PM Ps Su 0 Figure 2.-The graphs show the number of respondents by medical specialty. An anesthesiology; FP = family practice; IM = internal medicine and related specialties; Dl = pathology, radiology, and nuclear medicine; Ob obstetrics and gynecology; Pe = pediatrics; PM = physical medicine and rehabilitation and therapeutic radiology; Ps psychiatry; Su surgery and its subspecialties; 0 = declines to
An FP

Lt
IM
= = =

LLL LLkt

=

placed

type.

state

THl

THE WESTERN JOURNAL OF MEDICINE

o

MAY 1988

0

148

o

5

595
9

There was no statistically significant difference in our cohort in the incidence of alcoholism by sex of the respondent, but almost 30% of our respondents declined to identify their gender, perhaps out of fear of being identified. Thus, we can say little from our data about the true relationship between gender and alcoholism in physicians. Medical specialty also bore no statistically significant relationship to alcoholism in our study. However, a pronounced variation was noted in the prevalence of alcoholism in different specialties and may have achieved statistical significance had the cohort been larger. It was striking to note the high number of alcoholics in the specialties of family practice and obstetrics and gynecology compared to none in anesthesiology, physical medicine and rehabilitation, therapeutic radiology, or psychiatry. Since slightly more than half of those surveyed did not return their questionnaires, it would be hazardous to draw any firm conclusions from this study-we have no knowledge of the prevalence of alcoholism among those who did not return the questionnaire. Yet, with 14% of those who did respond being classified as either alcoholics or possible alcoholics, it appears that this may be a pervasive problem that clearly demands further study. If further studies do substantiate this high rate of alcoholism, what can be done about it? How can physicians, hospitals, and state licensing boards concurrently assure the best therapy for those physicians affected, as well as protection for them and for their patients? Rehabilitation can be more successful in alcoholic physicians than in nonphysician alcoholics. In one report, 76% of alcoholic physicians were abstaining one year after discharge from treatment, as op-

posed to 61 % of patients in the general population."0 National, state, and hospital boards have constructed detection and therapy programs for impaired physicians that have been effective.I1-13 Reluctance to identify and "stigmatize" professional colleagues with a diagnosis of alcoholism is a major impediment to care.9 We must continue to design programs aimed at educating physicians and medical students in the realities of alcoholism and its treatment, the prevalence of this disease, and the availability of Alcoholics Anonymous and other rehabilitation programs.
REFERENCES

1. Brewster JM: Prevalence of alcohol and other drug problems among physicians. JAMA 1986; 255:1913-1920 2. Niven RG, Hurt RD, Morse RM, et al: Alcoholism in physicians. Mayo Clin Proc 1984; 59:12-16 3. Lewy R: Alcoholism in house staff physicians: An occupational hazard. J Occup Med 1986; 28:79-81 4. Selzer ML, Vinokur A, van Rooijen L: A self-administered Short Michigan Alcoholism Screening Test (SMAST). J Stud Alcohol 1975; 36:117-126 5. SAS User's Guide: Statistics. Cary, NC, SAS Institute, 1985 6. Murray RM: Alcoholism amongst male doctors in Scotland. Lancet 1976; 2:729-731 7. Bissell L, Jones RW: The alcoholic physician: A survey. Am J Psychiatry 1976; 133:1142-1 146 8. Vaillant GE, Brighton JR, McArthur C: Physicians' use of mood-altering drugs. N Engl J Med 1970; 282:365-370 9. Talbott GD, Benson EB: Impaired physicians: The dilemma of identification. Postgrad Med 1980; 68:56-64 10. Kliner DJ, Spicer J, Barnett P: Treatment outcomes of alcoholic physicians. J Stud Alcohol 1980; 41:1217-1220 1 1. Spickard A, Billings FT Jr: Sounding board: Alcoholism in a medicalschool faculty. N Engl J Med 1981; 306:1646-1648 12. Talbott GD, Gallegos KV, Wilson PO, et al: The Medical Association of Georgia's impaired physicians program-Review ofthe first 1,000 physicians: Analysis of specialty. JAMA 1987; 257:2927-2930 13. Shore JH: The Oregon experience with impaired physicians on probation-An eight-year follow-up. JAMA 1987; 257:2931-2934