Alcohol: No Ordinary Commodity's goal of policy relevance leads naturally and appropriately to a discussion of system-level issues in treatment and early intervention (Chapter 12 of Babor et al. 2003). Defining what constitutes a ‘system’ has proved elusive, and many countries’‘alcohol treatment systems’ are systems only in name (Klingemann et al. 1992). Babor et al. (2003) grapple with two aspects of the system question, namely what principles should guide (a) how we identify, understand and prioritize need for intervention in the population, and (b) how we organize available intervention services. Babor et al. (2003) do an admirable job of proposing a system for defining population need for services, which is one of the toughest challenges in public health planning (Andrews & Henderson 2000). The six conditions for a population approach (Box 12.1) surmount the usual problem of equating diagnostic prevalence with treatment need. As the authors are clearly aware, not all diagnosed cases want treatment, not all would benefit if they received it and many others will remit without it. A further excellent aspect of the authors’‘conditions for a population approach for alcohol-related problems’ is its attention to the economic benefits of treatment. To provide treatment solely on compassionate grounds is laudable, but we must recognize a political reality: there will probably always be a limit on how sympathetic the average citizen will feel towards someone who has a problem with alcohol or other drugs. For this reason, any politically viable population approach must establish the benefits of treatment not just to patients, but also to the broader society who will be asked to devote public resources to it. Chapter 12 could have gone further in proposing a system to organize the interventions it describes so ably. The ‘stepped care’ system seems promising in this regard (Sobell & Sobell 1999). Under this model, alcohol-related mutual help organizations could constitute the first line of defense for help-seekers because they do help many individuals, are easy to access and cause minimal intrusion into the individual's life (Humphreys 2004). Policy makers might be particularly interested to learn that several countries have implemented policies that strengthen self-help groups without co-opting them, and that such efforts probably more than pay for themselves through reductions in health-care costs (Humphreys 2004). Those people with alcohol problems who are uninterested in, or not helped by, mutual help groups would become candidates for a brief intervention in a stepped-care treatment system. The authors note correctly that brief interventions cause mean reductions in alcohol consumption in non-dependent problem drinking samples. However, such mean changes prove neither that every non-dependent problem drinker benefits from brief interventions nor that no alcohol-dependent individuals benefit. To set an a priori decision rule based on diagnosis may be less wise than simply giving brief interventions to everyone and then defining the severe cases as those who do not benefit. Many treatment decisions in other areas of medicine are made in a pragmatic fashion in which low-level interventions are used for both treatment and assessment of whether more intensive services are needed. This seems to me a more promising approach for organizing an alcohol treatment system than is the continued pursuit of a priori rules for assigning patients to one type of psychotherapy or another (the ability of the psychotherapy matching hypothesis to keep walking around, post-Project MATCH, with so many empirical stakes in its heart would put Count Dracula to shame). In a stepped-care treatment system, actual data on how each patient responds or fails to respond to lower intensity intervention are used to assign them to the next level of care (e.g. out-patient, residential, telephone follow-ups). I would close with a final comment about systems for planning and organizing alcohol-related services. I share the authors’ belief that science can and must inform policy; but I am sure they would agree with me that empirical comparison of systems is very difficult. For example, conducting the Fort Bragg evaluation of two alternative systems for organizing mental health care cost nearly US$100 million in today's terms (Bickman et al. 1995). As scientists, we can often only test parts of our proposed systems, leaving us always in need of compelling conceptual models in which to arrange our results into coherent policy proposals. Babor et al. (2003) offer an excellent model for assessing and responding to population need, which along with other work in the field on organizing treatment systems (e.g. Sobell & Sobell 1999) should raise the policy relevance of alcohol research. I thank John Finney for comments on these ideas and the US Department of Veterans Affairs Health Services Research and Development Service and National Institute on Alcohol Abuse and Alcoholism for grant support.
Read full abstract