I blogged about Malcolm Gladwell's coverage of programs like this, just recently. To borrow from his latest column, he provided "story" reasoning, whereas this Annals article provides a "technical account" explanation.
The long and short of it, from the intro and discussion:
California law also provides judges the option of offering such individuals an opportunity to complete an alcoholism treatment program in lieu of custody. Before the implementation of Serial Inebriate Program, local treatment programs were unwilling to accept these clients because of their recidivist behavior, and jails rarely housed them longer than 72 hours. In 1999, the San Diego Police Department recruited a treatment provider to collaborate in the development of a novel pilot program tailored to this population. The San Diego Police Department also secured the support of the city attorney to develop new booking and sentencing procedures. Importantly, the public defender lent its critical support to program development after concluding clients would be afforded valuable new support and care. Volunteers of America staff were asked to define the criteria that should constitute a "chronic inebriate" (and therefore Serial Inebriate Program client), and the superior court endorsed a trial program...
...This study documents the extraordinary consumption of EMS, ED, and inpatient resources by one city’s population of homeless chronic alcoholics. Assuming an average observation period of 4.5 hours (Dunford, unpublished data) these patients consumed nearly 15,000 hours of ED staff time at 2 of San Diego’s major regional hospitals, which equates to a 34% chance that a Serial Inebriate Program client was occupying an ED bed at one of these facilities at any moment during the 4-year study. These data provide evidence that a relatively small number of individuals can have a large impact on a community’s safety net.
This study also demonstrates that a community-supported alcohol treatment strategy that incorporates legal consequence can reduce the consumption of emergency health care resources. There was a 50% decline in the use of ED, inpatient, and EMS resources for the 156 individuals who accepted a 6-month outpatient treatment program in lieu of custody. Conversely, there was no change in resource consumption by the 112 individuals who elected not to enter treatment. The Serial Inebriate Program’s success derived primarily from its impact on the most recidivist individuals. Those accepting treatment were typically older men who had been transported by EMS and treated in ED twice as often as nonacceptors.
Many more nuggets in the article, with citations to many more articles on homelessness and ED frequent fliers (yet another medicine / airline analogy). And, of course, profound limitations that leave many questions of costs and long-term effectiveness unanswered (EM research, sigh).
Moreover, San Diego's program is not as revolutionary as Nevada's or Denver's, in that it relies on courts and legal threats more than largesse and social work. The authors expressed surprise that they were able to persuade even 156 of San Diego's chronic alcoholics to sign onto 6 months of treatment (even with the threat of jail) -- Gladwell reported no problems recruiting recidivists in his article.
Either way, it seems that programs like San Diego's or Nevada's are going to prove their worth. The most hardcore chronic inebriates are just too expensive to be treated and streeted every few days; special programs to handle them are bound to be cost-effective.