As a clinician, the outcomes that I consider most important among my patients with serious mental illness have changed quite a bit over time. I spend less time these days focusing on the frequency or intensity of their specific symptoms and more time looking at their overall quality of life. In brief, I have prioritized "recovery-oriented outcomes," as summarized by Charles G. Curie when he was the director of the Substance Abuse and Mental Health Services Administration (2001—2006), as follows: Do they have "a place called home, a decent job and a date on Saturday night?"

Yet I do not find myself typically prioritizing questions such as "When was the last time you had a physical exam, went to the dentist, had a flu shot, etc.?" In this issue of the Journal, Druss and colleagues (1) remind me that perhaps I should, and they remind us all of the sobering statistic that people with serious mental illness continue to have a life expectancy of 25 years less than that of the general population. Further, there is evidence that this gap may be broadening rather than narrowing over time (2). While premature death by suicide as well as the sequelae of comorbid drug and alcohol abuse certainly accounts for part of this, it is becoming increasingly clear that much of the early mortality is accounted for by inadequate attention to and care of preventable and treatable general medical disorders, such as diabetes, emphysema, infections, and hypertension and other vascular diseases. If we want to pursue real recovery-oriented outcomes, then meaningful efforts to address this "mortality gap" must be added to our priorities, both clinically and on our research agenda.

To some extent, the broad recognition of the effects of atypical antipsychotics on the risk for metabolic syndrome has mobilized interest and action in this area. It is becoming increasingly standard practice in mental health settings to routinely screen for diabetes and hyperlipidemia, at least among patients taking specific medications. But this is just a beginning.

There are a myriad of reasons why individuals with serious mental illness are less likely than others to receive adequate general medical care, especially preventive services. Druss and colleagues organize these into patient, provider, and system factors and pre­sent a randomized, clinical trial of a model that tries to improve preventive health services by addressing barriers in each of these domains. The core of the model involved having a "medical care manager" within a community mental health setting whose activities were designed around overcoming hypothesized barriers across all three domains. For example, some patients with serious mental illness may simply be unaware of the risks involved in not engaging in basic preventive care, while others may not know where or how to go to get these kinds of services even if they were aware of the need. As such, there is an educational component to the care manager's role involving both information about where and how to receive preventive services in the area, what kind of preventive services are available in the area, as well as the potential benefits of utilizing these services and the risks in not doing so. The care managers utilized a motivational interviewing approach to strengthen client autonomy and evoke personal reasons that a client may wish to take advantage of these services (3). They also helped clients create "action plans" delineating what specific steps are involved. To address provider barriers, the care managers established relationships with a cadre of willing providers and, with client permission, kept them informed of relevant changes in the clients' health status, medications, etc. Perhaps most importantly, they also engaged in some very practical help to address system barriers, such as providing clients with money for public transportation to get to appointments and even taking them to appointments when necessary.

The study was conducted in an urban community mental health center with about 400 seriously mentally ill adults, most of whom were African American and economically disadvantaged. Subjects were rated at 6 and 12 months after intake on a variety of variables that informed the quality of primary care and preventive services they were receiving as well as direct health outcomes. Not surprisingly, there were significant group differences in many of the measures of the quality of primary care, but less so in terms of direct health outcomes. Some of the improvements in the quality measures were fairly dramatic. For example, at intake, only about one-third of patients in both groups had undergone a physical exam by a primary care provider over the previous year. At the 12-month follow-up evaluation, this increased to over 70% for the care management group, while there was no change for the comparison group. Significant improvements in the treatment group were also seen in the use of appropriate screening measures and vaccinations. On an overall measure of appropriate preventive services, the care management group increased almost threefold, from 21.5% to 58.7%, with no change in the comparison group (21.6% to 21.8%).

In terms of the actual health outcome variables, among the domains measured in the Medical Outcomes Study 36-Item Short-Form Health Survey, the only ones that differentiated the treatment and comparison groups were social functioning and the mental component summary score. Domains that more directly reflect physical status, such as bodily pain, physical functioning, or general health, were not statistically different between the groups. Given the time lag expected between good preventive services and directly measurable health outcomes, it would be surprising to see clear effects within the time frame of a 1-year study. It is notable that among those patients who had predetermined risk factors for metabolic syndrome, there was evidence of an improvement in their overall cardiovascular risk, even over the relatively short time of the study.

There is a lot of work ahead in this area. The costs and overall economic balance sheet for this kind of effort need to be elucidated. The outcomes of this type of model need to be compared with other collaborative and colocalization models in which mental health and primary care services are provided under the same roof. Hybrids of these models might be explored (i.e., some direct preventive service provision in the mental health setting along with appropriate care management functions as indicated). And of course the ultimate test will involve direct health outcomes over longer periods of time.

Perhaps it is the overwhelming nature of the problems imposed by having a serious mental illness that have led many of us to ignore or minimize the need to better attend to the more mundane general preventive measures. If tormenting hallucinations can be better controlled by a medicine that may increase appetite, a few extra pounds or slightly higher fasting blood glucose seems a reasonable price to pay. If we can get an unmotivated or paranoid patient to make at least some of their visits to the mental health center, should we really push him or her to go to yet another doctor's appointment for a routine check-up? If we are serious about wanting to close the mortality gap, the answer is yes.