Two research reports in this month's issue focus on ensuring that treatment decisions are driven by the preferences of persons with mental illness—a key feature of a transformed system. In a two-year study of 106 outpatients who had completed a psychiatric advance directive, Debra S. Srebnik, Ph.D., and Joan Russo, Ph.D., examined whether their crisis care was consistent with their expressed preferences. Across 90 crisis events in which directives were accessed, the average rate of care consistent with directive instructions was 67%. Clients who had a surrogate decision maker who accessed the directive were more likely to receive directive-consistent care (page 1157). In a brief report Jared R. Adams, B.S., and colleagues present the results of a pilot study in which they interviewed 30 adults with severe mental illness about their preferences in regard to shared decision making. About three-quarters preferred an autonomous or shared role in treatment decisions, especially in regard to medications; however, only 37% rated their current roles as more than passive (page 1219). In a related commentary Marvin S. Swartz, M.D., and Jeffrey W. Swanson, Ph.D., highlight the importance of these empirical studies, which provide support for a recovery orientation and point to opportunities for improving care (page 1164).

Preventing incarceration of people with mental illness and ensuring that those who spend time in correctional settings receive needed care are critical areas for research. Jonathan D. Prince, Ph.D., and colleagues examined risk factors for incarceration in a cohort of 538 persons over four years after they were discharged from a first hospital admission for psychosis. Forty-seven were incarcerated—20 had multiple incarcerations. The most significant predictors of incarceration were being male or black and having been incarcerated before admission. The authors encourage clinicians to routinely assess incarceration history in an effort to avert clients' future jail time (page 1173). In a survey of 418 incarcerated adult males about their use of mental health services, Robert D. Morgan, Ph.D., and colleagues found that many inmates were reluctant to access services available in their correctional facility. The findings point to specific areas for inmate education about treatment and about the qualifications of professionals who provide it (page 1181).

Since Sigmund Freud's The Future of an Illusion, many psychiatrists have had to counter accusations of having no religious beliefs. In fact, several surveys in the latter half of the 20th century found that psychiatrists were more likely than the rest of the population to be nonreligious. Farr A. Curlin, M.D., and colleagues found that psychiatrists have not changed significantly in this regard. The authors analyzed responses to a recent survey of 1,144 practicing U.S. physicians, including 100 psychiatrists, and found not only that psychiatrists are less religious than other physicians but also that religious physicians are less willing than nonreligious physicians to refer patients to psychiatrists (page 1193). In a Taking Issue commentary on the findings of this study, Greer Sullivan, M.D., M.S.P.H., points out that "significant cultural differences often exist between mental health care providers and the very communities in which we practice." She calls for clinicians to develop "community cultural competence" in order to make treatment more acceptable and accessible (page 1145).

The most common ways of assessing adherence to antipsychotic medications are self-report and physician rating. How do these subjective measures compare with objective measures, such as in-home pill counts, electronic pill bottle caps, and plasma concentrations? Not well at all, according to a study conducted by Dawn I. Velligan, Ph.D., and colleagues. This group used electronic monitoring as an imperfect standard against which to judge other adherence measures in a sample of 52 outpatients with schizophrenia. Thirty-five percent of the patients were not taking a therapeutic dosage. However, neither patients nor physicians were able to identify adherence. The authors suggest that "smart" pill containers may help physicians make more informed decisions about medications and dosages (page 1187).

• The State Mental Health Policy column reports on current priorities of the nation's state mental health agency directors and the forces reshaping these agencies (page 1148).

• Six years ago, on September 11, 2001, Lauren LaPorta, M.D., was in charge of a psychiatric inpatient unit across the river from the World Trade Center. In Personal Accounts, she summons images of that day (page 1154).