Estimating the effectiveness and benefits of alcohol treatment programmes for use in economic evaluations

A binary probit model based on random utility theory is employed to obtain the implicit determinants of alcohol treatment effectiveness revealed by decisions made by programme evaluators of behavioural changes by patients. A scale of equivalences for the behavioural variables is constructed which uses reductions in alcohol drinking as the unit of account. Because one of the behavioural variables was the change in income from treatment expressed in dollars, the trade-offs contained in the scale of equivalences can be translated into monetary terms using the human capital approach. The method is applied to a sample of 1689 observations from the National Alcoholism Program Information System (NAPIS) tapes for 1977—81.