• Alcohol problems;
  • natural recovery;
  • persistence;
  • recurrence

Longitudinal studies of samples affected by alcohol use disorders are relatively rare. These studies suggest that most of those who have substance use problems recover without treatment, and most of those who receive treatment do not recover in the medium term [1, 2]. While we have some understanding of the natural history of alcohol use disorders, and the evidence suggests that natural recovery from these disorders is the norm, the extent of natural recovery and the implications for policy and practice warrant more discussion.

In the current issue, Boschloo and colleagues [3] report the findings of a 2-year follow-up of a selected sample of people who met the criteria for an alcohol use disorder. Boschloo et al. administered the Composite International Diagnostic Interview (CIDI) to assess whether the respondent had ever had an alcohol disorder. At the 2-year follow-up they determined whether the alcohol use disorder was still present (i.e. persistent disorder) or whether it had remitted and/or subsequently recurred. Importantly, treatment could be excluded as a reason for remission, as only 4.8% of the sample reported ever receiving treatment for their disorder. Boschloo et al. found that of those who had ever had an alcohol use disorder, 65.2% had remitted by the baseline interview, while 34.8% met the criteria for an alcohol disorder in the last year (persistent group). Of those who were classified as remitted at baseline, 14.6% had a recurrence of their alcohol disorder in the 2-year follow-up. Male gender and the severity of the disorder were the main predictors of persistence and recurrence.

This study contributes to our knowledge about alcohol use disorders in three areas: (i) the extent to which natural recovery from alcohol use disorders is the norm; (ii) the extent to which treatment programmes make a contribution to overall outcomes; and (iii) whether abstinence or controlled consumption are common outcomes.

Only a small number of studies have documented the natural history of those who manifest alcohol use disorders. George Vaillants' two longitudinal studies are fundamental here. In his first cohort study of disadvantaged youth [4], involving 456 boys recruited from a disadvantaged background, he found that by age 60 years 36% had abused alcohol at some stage of their life-course. He characterized alcoholism as a ‘spectrum’ disorder (no clear distinction between those with and without the disorder), with alcohol use escalating gradually until about the age of 40 years. Levels of alcohol use began to decline ‘naturally’ after that age. In a second longitudinal study involving a follow-up at 8 years of about 100 alcoholics receiving treatment [5], 95% of those who had received treatment had relapsed. Those alcoholics who had received treatment had outcomes which were indistinguishable from those not receiving treatment. Vaillant noted that controlled drinking was possible, but that the more usual outcome for those who had had an alcohol use problem was abstinence.

The 1976 Rand Corporation study remains the most important large-scale study of community treatment for those with an alcohol disorder [6]. It involved a 6- and 18-month follow-up of 11 500 males treated in 45 community-based services. Only a minority of those treated were followed-up successfully. Of those followed-up at 6 months, 70% showed a reduction in use and approximately 10% were abstinent at 6 and 18 months. These estimates of treatment outcome were similar to those derived from a sample who were attending Alcoholics Anonymous (AA) meetings. This study generated the controversial finding that most of those who had ‘recovered’ continued drinking in a controlled manner. The Rand study also monitored an untreated sample, of whom 50% showed a reduction in alcohol use. The Rand study also found that the type of treatment received made little or no difference (e.g. hospitalization, residential care, out-patient care). As it is likely that those most affected by their alcohol use are most likely to be lost to follow-up, the results of the Rand study should be considered optimistic assessments of overall treatment outcomes.

Only a few studies have built on these early foundations comparing natural recovery to recovery following treatment [2, 7]. These have tended to confirm that, in community samples of those who experience problems with their alcohol consumption, the majority of those who recover do so without treatment; and that the majority of those who receive treatment do not maintain their recovery into the medium term. The specific type of treatment received appears to have a modest impact on outcomes. While treatment leads to improved outcomes [8], this level of improvement is not large compared to the proportion who recover naturally. Boschloo et al. also confirm that a proportion of those who have recovered will have a recurrence of their problematic alcohol consumption in the years following the remission. The relatively small contribution made by treatments to overall remission rates raises questions about the conceptualization of alcohol disorders as chronic recurring conditions. Providing services to maintain natural recovery may warrant a trial, as would policies to enhance the process of natural recovery. The case for the increased availability of treatment services would be strengthened if there was better evidence for specific treatment outcomes above the outcomes which are potentially attributable to natural recovery.

Declaration of interests


  • Jake M. Najman

  • Queensland Alcohol and Drug Research and Education Centre, Schools of Population Health and Social Science, The University of Queensland, Brisbane, QLD, Australia. E-mail:


  1. Top of page
  2. References