Five reports in this month's issue focus on people with mental illness involved in the criminal justice system—who they are and how programs to provide appropriate treatment are working. In a study linking multiple databases, Robert Constantine, Ph.D., M.P.H., and colleagues found that 10.1% of all persons jailed in a Florida county in a one-year period had a diagnosis of a serious mental illness. The 3,769 individuals in this cohort had a total of 17,463 arrests over four years, accounting for 567,584 jail days. Findings suggest two discrete subgroups of inmates in the cohort (page 451). John Petrila, J.D., LL.M., and colleagues present expenditure data for the same Florida cohort, including direct costs to jails and prisons and costs to other systems. Aggregate costs were nearly $95 million, with a median per-person cost of $15,134 (page 516). A study in the Philadelphia jail system by Jeffrey Draine, Ph.D., and colleagues found that length of jail stays among inmates with and without serious mental illness did not differ significantly—about half of each group was released within 30 days. Examination of release mechanisms indicated that only 19% of those with mental illness left jail in a predictable way that allowed time for the usual process of release planning for this population (page 458). In a two-year follow-up study Virginia A. Hiday, Ph.D., and Bradley Ray, M.A., found that defendants with mental illness who completed a six-month program offered by a North Carolina mental health court were much less likely to be rearrested than those who left the program early and had a much longer period before rearrest (page 463). A literature review by Steve Ryan, M.D., M.B.A., and coauthors examined accumulating data on U.S. postbooking jail diversion programs for adults with serious mental illness. They found a wide range of approaches to diversion and substantial progress in program dissemination—but also a knowledge gap about factors that predict positive outcomes for participants (page 469). In a Taking Issue commentary, Marvin S. Swartz, M.D., notes that many interventions "have been haphazardly implemented for persons with a history of criminal justice involvement without consideration of the effects of the criminogenic environments that they come from or the developmental pathways that have spawned antisocial behavior" (page 431).

Researchers worldwide are identifying specialized treatments for first-episode psychosis that appear to achieve better outcomes than treatment as usual. A next step is to compare outcomes across these programs, for which a risk adjustment model is needed. Risk adjustment is a statistical process for controlling for group differences when nonequivalent groups are compared on outcomes of interest. In this issue a group of Canadian researchers led by Donald Emile Addington, M.B.B.S., describe the development and performance of a robust model that adjusts for 12 risk factors for hospital admission, which is a negative outcome of first-episode psychosis treatment (page 483). In Maine a communitywide system to identify and treat youths at risk of psychosis has been in operation for a decade. The PIER program (Portland Identification and Early Referral), which was established in 2000, has trained more than 7,200 professionals in schools, primary care, and other systems to identify at-risk youths and refer them to the PIER clinical team for assessment and appropriate care. A public education campaign focused on parents also helps in this effort. William R. McFarlane, M.D., and colleagues report identification and referral rates from this public health initiative designed to prevent the onset of psychosis among at-risk youths (page 512).

• Psychosis has played a negligible role in cases resulting from an Indiana law permitting seizure of a firearm from a person with mental illness who is believed to be dangerous (page 478).

• Data reported in the Economic Grand Rounds column show that for each hour of direct care, collateral contacts, which are at the heart of child wraparound services and which are often not billable, require an average of 20 minutes (page 440).

• The Open Forum presents a framework for research and action in five domains to improve long-term rehabilitation of African Americans with psychiatric disabilities (page 508).

• A 28-state survey of certified peer specialists found varied work settings and modalities along with a core set of roles and activities (page 520).

• The Public-Academic Partnerships column describes tools to help community agencies decide whether to participate in proposed research studies (page 446).