[Longitudinal research on treated alcoholism: Systematic review of long-term follow-ups]

To analyze the methodological design of long-term longitudinal studies of alcoholics who have undergone treatment.

Systematic review of prospective longitudinal studies of treated alcoholics with follow-up periods of at least 8 years, and published in English, during the period 1983-2009. The review includes 9 studies which analyzed: follow-up evaluations, information sources, the period of reference in the final assessment, the measurement of alcohol use, the operationalization of drinking patterns, and data analysis.

2434 patients (ranging between 57 and 850 per study) were studied on average for a period of 15.3 (D.T. = 3.2) years (range 8-20 years). Studies differ in the number of intermediate evaluations (0, 1, 3, 4) and the length of the period considered for the final evaluation (the whole period, the last 3 years, the last year, last 6 months, last month or present situation). Drinking patterns tend to be used as the main outcome variable, but they are operationalized in non-equivalent forms and the analysis is performed using classical cross-sectional statistical techniques. Alcoholism evolution is studied by analyzing the evolution of drinking patterns and patients' life situation during the period under study.

The scarce amount of data available on the evolution of treated alcoholism cannot be compared across studies due to considerable methodological differences. There is a need to promote common methodological criteria in relation to follow-up strategies. New statistical methods that permit longitudinal analysis for non-balanced and correlated data should also be incorporated.