Screening for cardiovascular risk factors in adults with serious mental illness: a review of the evidence

We found that rates of screening for cardiovascular risk factors in the population with SMI varied considerably in the studies reviewed. Much of this variation is likely attributable to health system differences, such as whether or not a given system has a standard protocol for screening for cardiovascular risk factors among antipsychotic users. Implementation of standard screening protocols is feasible and an effective tool for improving rates of screening. For example, six recent studies showed significant improvements in rates of screening for obesity in patients with SMI following implementation of standard screening protocols [41-46].

While health system differences may account for some of the variation in screening rates across the studies we reviewed, other important factors may help to explain this variability. Screening rates may change over time due to secular trends in medical practice. Variation in screening rates may also be due to differences in measurement technique (e.g. administrative claims data versus chart review) and the time window used to ascertain the screening rate. To be included in our review, study populations had to include – but were not limited to – persons with schizophrenia and bipolar disorder. Differences in diagnostic distribution within study populations could also influence variation in screening rates; however, the majority of study populations reviewed were comprised primarily of persons with schizophrenia and bipolar disorder using antipsychotic medications and no discernable patterns in screening rates by diagnostic composition of study sample were observed in our results.

Consistent with prior work, the results of our review suggest that rates measured over short time windows, for example in the 30 days following initiation of antipsychotic medication, tend to be lower than rates measured over longer periods [47]. In addition, rates are likely influenced by the frequency of patients’ encounters with the healthcare system. Patients who are frequent users of medical care services are more likely to be screened [48]. In addition, prior work has shown that individuals with more pre-existing risk factors for a condition are more likely to be screened for that condition [48]. Screening rates may be higher among those with pre-existing risk factors because providers prioritize screening among patients at heightened risk. This group may also be sicker and therefore more likely to use health services – increasing their chances of being screened due to increased contact with the healthcare system [48].

Variation in rates of screening for cardiovascular risk factors in the population with SMI is likely caused by multiple factors, including providers’ experience treating persons with SMI, [49] co-located versus geographically separate medical and mental health services, [18] and degree of continuity and coordination among medical care providers [50,51]. Consistent with studies evaluating quality of care conducted in the overall US population, [52,53] screening rates for cardiovascular risk factors among those with SMI tended to be high in the Veterans Health Administration (VHA) [22,30,31,54-56]. This is likely due at least partly to the VHA’s ability, as a national system, to use electronic patient data to trigger standard screening protocols for cardiovascular risk factors [57]. Ongoing initiatives to implement electronic health records in other health systems may provide an important tool to improve quality of medical care for persons with SMI [58].

The ongoing implementation of the Affordable Care Act presents a multitude of opportunities to improve screening for and treatment of cardiovascular risk factors among persons with SMI through initiatives such as Accountable Care Organizations (ACOs), Medicaid Health Homes, and state demonstration projects to integrate care for persons with SMI who are dually eligible for Medicare and Medicaid [50]. All three initiatives focus on coordinating care for populations with complex health needs, such as adults with SMI, to ensure that those with positive screenings receive appropriate follow-up treatment. Additionally, these initiatives rely heavily on health information technology, which can facilitate clinical decision support interventions and improve care coordination and patient outcomes [59-61]. Rigorous evaluation of these efforts is critical in order to understand how these new initiatives can best address the high burden of medical conditions and improve long-term health outcomes in the population with SMI.

Health homes, Accountable Care Organizations, and other integrated care and quality improvement initiatives use performance metrics to measure health system performance and, increasingly, determine rates of reimbursement and shared savings for physicians [62]. Quality measurement can be an important tool to drive quality improvement in the healthcare system. In response to the increased risk of metabolic syndrome associated with use of antipsychotics, the National Committee for Quality Assurance (NCQA), a private organization dedicated to improving the quality of health care in the US, released new quality measures in 2014 relating to diabetes and cardiovascular disease screening and monitoring for adults with SMI who are using antipsychotic medications [63]. More specifically, the NCQA defines three new measures: 1) diabetes screening for people with schizophrenia or bipolar disorder who are using antipsychotic medications; 2) diabetes monitoring for people with diabetes and schizophrenia; and 3) cardiovascular monitoring for people with cardiovascular disease and schizophrenia. Measurement of diabetes and cardiovascular disease screening and monitoring for adults with SMI will help to elucidate opportunities for improvement.

To our knowledge, this is the first study to provide a comprehensive review and compilation of screening rates for cardiovascular risk factors in the US population with SMI. The results of our review have several implications. First, the variation in screening rates suggests that researchers and practitioners interested in improving screening for cardiovascular conditions in this vulnerable population need to take into account unique healthcare setting, measurement, and population characteristics when designing and evaluating screening programs. Second, our results suggest that rates of screening for cardiovascular risk factors are very low in some study populations of persons with SMI, suggesting a need for improved implementation of screening in some settings. Third and perhaps most critically, our review identified considerable variation in the measurement and reporting of screening for cardiovascular risk factors in the population with SMI, making it difficult to compare screening rates across studies and with national rates. Given growing national interest in improving care and reducing costs for the population with SMI, there may be opportunities moving forward to standardize measurement and reporting of cardiovascular risk factor screening rates in this population. For example, 95% of persons with schizophrenia are insured by government programs: Medicaid and the Veteran’s Health Administration [64]. These insurers are already required to report multiple administrative-claims based measures of care quality to the federal government. Requiring these programs to report standardized measures of cardiovascular risk screening – for example the new NCQA measures described above – for beneficiaries with SMI would provide a critical tool for quality improvement and allow researchers and policy makers to monitor national screening rates over time.

These results should be interpreted in the context of several limitations. First, the screening measures abstracted from studies were not always directly comparable with one another or with national measures. The timing of measures varied, with some studies measuring a screening indicator within a one year period and others measuring over a longer period of time. Additionally, the patient characteristics of the included studies varied, which could have implications for receipt of cardiovascular screening tests as patient demographic characteristics have associated with receipt of preventive services in prior research [65], it is possible that our search strategy missed relevant articles, although to minimize this risk we searched the reference list of included articles. In addition, because our study does not include interventions or randomized trials, we did not systematically assess the internal validity of the included studies, as is standard in systematic reviews of clinical trials. Screening rates extracted from the studies in our review may be subject to measurement error, particularly those that estimate screening using administrative claims data. Prior research has shown that claims data underestimates services for some chronic conditions, such as obesity. To aid in interpretation of results, we presented information about data source, measurement specification, and the time window in which the screening rate was measured. Other potential sources of bias to our results include selective reporting within studies and publication bias, but we were unable to identify the extent of these threats to validity based on the content of the studies reviewed.