Comments on the Rand Report 351 New Curves in the Course: A Comment on Polich, Armor and Braiker, *The Course of Alcoholism' Robin Room Social Research Group, School of Public Health, University of California, Berkeley, California 94720, U.S.A. The Background of the Report The rise of U.S. federal funding for alcoholism treatment in the early 1970s and the establishment of the National Institute on Alcohol Abuse and Alcoholism in 1971 coincided with the explosion of enthusiasm in the U.S. for comprehensive computerized monitoring systems for human and health services as an instrument of programme management and planning. NIAAA was thus putting its best foot forward as a progressive and accountable agency when it contracted in 1971 for the development and institution of a system which would eventually require fed- eral treatment grantees to fill out client information forms at intake and six months later for every case coming in for treatment [10, 11, pp. 120-122, for orig- inal contract reports, see 12, 13]. The monitoring system instituted then con- tinues today as 'NAPIS', The National Alcoholism Program Information System, see Patterson [14], NIAAA [15, pp. 53-63, 16, p. 57, 17, 18]. The NAPIS system no doubt has yielded a wide variety of information useful for programme management and planning both at the level of the individual programme and at the federal level, although little system-wide NAPIS data has made its way into the scholarly literature. But from the start the aim had been more ambitious than tracking the infiow of clients and output of treatment resources: the system was also intended to measure and show the effects of treat- ment. I believe the impulse here was partly scientific, and partly justificatory, to use the Levines' terms [19]. There was a scientific interest in knowing what kind of treatment worked how well for whom. But there was also a justificatory inter- est in demonstrating how effective and cost-efficient federally-funded treatment was. Thus from the first there was a strong interest in measuring clients' job status and income, to document their restoration as productive members of the economy. A major argument for societal support of alcoholism treatment in the U.S. has always been that it saves money in the long run - whether for the economy as a whole (government-funded treatment), for a private corporation (employee assistance programmes), or for health insurance providers (health insurance coverage of alcoholism). It is an argument that is frequently falsifiable, but is politically irresistible [20]. From the point of view of measuring effects of treatment, the monitoring sys- tem was seen to have two major defects: its follow-up period of six months was too short (a substantial proportion of clients were, in fact, still in treatment), and six-month follow-up interviews were completed with only about one-quarter of