Current issues of ACP Journal Club are published in Annals of Internal Medicine


Mandatory hospital treatment for alcohol abuse reduced rehospitalization but not job loss over 2 years

ACP J Club. 1992 Jan-Feb;116:7. doi:10.7326/ACPJC-1992-116-1-007

Source Citation

Walsh DC, Hingson RW, Merrigan DM, et al. A randomized trial of treatment options for alcohol-abusing workers. N Engl J Med. 1991 Sep 12;325:775-82.



To evaluate the efficacy of 3 different strategies for treating alcoholism in workers.


Randomized clinical trial of 24 months duration.


Work site of a 10 000-worker industrial plant, with hospital- and community-based treatment sites.


As they entered the company employee-assistance program, employees were recruited if they had an alcohol problem and if the company and union were uncertain whether the problem warranted hospitalization. Employees were excluded if they required detoxification because of a blood alcohol level ≥ 0.2 (43 mmol/L or 2.0 g/L), had a history of delirium tremens or grand mal seizures during withdrawal, were seriously ill, needed psychiatric care, or were about to be jailed or fired. 243 of 371 persons were eligible; 227 agreed to participate. 96% of the recruits were men.


Participants were randomly allocated to one of 3 treatment regimens: compulsory 3-week hospitalization followed by a year of attendance at Alcoholics Anonymous (AA) meetings (n = 73); mandatory attendance at AA meetings ≥ 3 times/wk (n = 83); or a choice of treatment such as hospitalization, AA meetings, outpatient psychotherapy, marriage counseling, or no therapy (n = 71). All participants were required to check in weekly with the employee-assistance program.

Main outcome measures

Length of time to being fired and length of time to hospitalization for supplementary treatment for alcohol abuse.

Main results

The groups did not differ for being fired from work. 31 (14%) of all 227 participants were fired: 11% of the hospital group, 13% of the AA group, and 17% of the choice group. Rates for hospitalization for additional alcohol treatment after the initial intervention were 23% of the hospital group, 38% of the choice group, and 63% of the AA group Repeat hospitalization did not deffer between the choice group and the hospitalization-first group with the AA group having higher rates than both groups. For the choice group compared with AA the rate of rehospitalization was lower (38% vs 63%, P = 0.005). {This absolute risk reduction (ARR) of 25% means that 4 patients would need to be given (NNT) choice in treatment for alcohol abuse (rather than AA) to prevent 1 additional hospitalization (95% CI 3 to 11); the relative risk reduction (RRR) was 39%, CI 16% to 57%}*. The data were similar for initial hospitalization compared with AA (23% vs 63%, P = 0.039); {ARR 39%, NNT 3, CI 2 to 4; RRR 63%, CI 43% to 77%)}*.


Mandatory inpatient treatment for alcoholism was more effective in preventing the necessity for later hospitalization than was attendance at AA meetings or the individual's choice of treatments. The type of rehabilitation program had no effect on job loss.

Sources of funding: National Institute on Alcohol Abuse and Alcoholism; Commonwealth Fund; Robert Wood Johnson Foundation.

Address for article reprint: Dr. D.C. Walsh, Department of Health and Social Behavior, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA.

*Numbers calculated from data in article.


The relative benefits of various types of treatment for alcoholism remain controversial. Most studies have shown little or no benefit from more intensive inpatient programs compared with outpatient alternatives. Because of the nature of Alcoholics Anonymous (AA), few randomized trials of AA as an isolated treatment option have been executed. This study compared in-patient treatment to either AA alone or to a choice of treatment for workers in a single plant. The study was methodologically sound with randomized allocation, balanced distribution of potentially confounding factors, and blinded follow-up for the principal outcomes. No difference between groups was found with respect to the likelihood of being fired or other job outcomes, whereas sustained abstinence was significantly more likely and subsequent hospitalization less likely for the hospital group than for the choice or AA group. Cocaine use at baseline was associated with especially poor outcomes in the AA group. However, the data presented for subsequent hospitalization indicate that, in the hospital and choice groups, cocaine users were actually less likely than all participants to be rehospitalized.

This study shows that well-designed trials of employee-assistance programs can be done. A total of 61% of all alcohol-abusing workers were successfully randomized to the 3 treatment options. Because the outcomes were evaluated in an intention-to-treat fashion, the finding of no difference in the job-related outcomes between groups can be generalized to settings that provide close follow-up, with referral for more intensive treatment if needed. The high rate of subsequent hospitalization for subjects assigned to AA alone shows the potential limitations of this strategy as an isolated method of treatment. Although the cost-savings were small for AA compared with initial hospitalization, this study did not address other daily full- or part-time out-patient treatment options, which may be equally effective but less expensive.

John Schorling, MD, MPH
University of VirginiaCharlottesville, Virginia, USA

John Schorling, MD, MPH
University of Virginia
Charlottesville, Virginia, USA