Review: prescribing rates of drugs for medical conditions are lower in people with mental illness
Question: Are there differences in prescribing for medical conditions in people with and without mental illnesses?
Outcomes: Prescribing rates. The population without mental illness was the comparator, with prescribing rates in people with mental illness being classified as adequate (≥95% of prescribing levels in people without mental illness), inequitable (≥90% and <95%), suboptimal (≥80% and <90%) and inadequate (<80%).
Design: Systematic review with meta-analysis.
Data sources: MEDLINE/PubMed, EMBASE, SCOPUS, Web of Science and the full text databases from Blackwell-Wiley, SCIENCE DIRECT, Ingenta SELECT and Springer-Verlag's LINK were searched from inception to November 2010 for interventional or observational studies. Selected journals and reference lists of identified studies were hand searched.
Study selection and analysis: Two reviewers appraised studies and selected interventional or observational studies assessing prescriptions for medical conditions in people with and without mental illness. Studies in people with dementia or delirium were excluded. Two reviewers extracted data on prescribed medication rates in people with and without mental illness. Participants with mental illness were classified into: severe mental illness (including schizophrenia), affective disorder and other mental illness. Summary meta-analyses were performed where relative risks were converted into OR and then pooled using Statsdirect. Heterogeneity was investigated using the I2 statistic. Fixed effects meta-analyses were performed when heterogeneity was low, and random-effects meta-analyses when heterogeneity was high.
Twenty-three studies met inclusion criteria (1 931 509 participants, 61 drug-level comparisons). Combining all mental illness and all drug types, the presence of mental illness was associated with lower prescription rates (OR 0.78, 95% CI 0.73 to 0.84; p=0.0001). Severe mental illness was associated with less equitable prescribing (36 analyses, >1.5 million participants; OR 0.74, 95% CI 0.63 to 0.86). Prescribing rates for ACE/ARBs, β-blockers, statins and the non-aspirin anticoagulants clopidogrel and ticlopidine were significantly lower in people with severe mental illness compared with those without (see online supplementary webextra table). For this population, there were no significant differences between groups for anticholesterol drugs as a whole (statins and non-statins), or for anticoagulants as a whole (aspirin and non-aspirin). Affective disorder (13 analyses, n=232 882) was associated with a non-significant trend to lower prescribing compared with no mental illness (OR 0.75, 95% CI 0.55 to 1.02, p=0.07) although differences were significant when fixed effects meta-analyses were performed (data not shown). For this population, there were no significant differences between groups for β-blockers and anticholesterol drugs (see online supplementary webextra table). Other mental illness (12 analyses, n=19 637) was associated with lower prescribing compared with no mental illness (OR 0.95, 95% CI 0.92 to 0.98). Prescribing rates for ACE/ARBs were lower in this population compared with people with no mental illness (see online supplementary webextra table).
Prescribing rates for a number of commonly used drugs for medical disorders are lower in people with severe mental illness, especially drugs used for cardiovascular conditions.
There is an increasing body of research showing that severe mental illness (SMI) is associated with higher mortality and physical health morbidity as compared to general population. Cardiovascular events are one of the common causes of excessive natural mortality in individuals with SMI, possibly explained by unhealthy lifestyle choices, prescription of multiple psychotropic medications leading to metabolic side-effects, and inadequate access to healthcare.
This meta-analysis eloquently demonstrates that there is suboptimal prescribing of physical health medications in individuals with mental illness. The majority of the medications reviewed were for cardiovascular indications, which may further reflect the high cardiovascular morbidity in this population.
An inherent problem with large meta-analyses is heterogeneity. Even in this meta-analysis, a number of medical and mental disorders have been lumped together, making inferences difficult at times. For example, figure 2 in the Mitchell et al paper suggests that there was a statistically significant reduction in prescribing of antihypertensives and statins in patients with severe mental illness compared to controls. However, prescribing patterns for other classes of medications appear inconclusive. Mitchell and colleagues have made commendable efforts to address heterogeneity and acknowledge other limitations, including a lack of a priori protocol and inconsistent definition of mental illness in the included studies.
It is difficult to generalise the results of this study to a socialised healthcare system like the UK, or to a mixed healthcare system like Canada, as most of the included studies were conducted in the USA (19 of the 23), where the majority of health services are provided by private insurance. The suboptimal prescribing noted in this analysis may not be entirely attributed to prescriber factors; inadequate healthcare insurance coverage in the USA may lead to reduced patient preference for prescribed treatment. Nevertheless, among health professionals there is a need for increased awareness of the higher prevalence of physical health mortality and morbidity in those with severe mental illness. Family physicians and mental health professionals must work collaboratively for surveillance of cardiovascular risk factors in this vulnerable population.
A systematic review investigating psychotropic polypharmacy and its correlation with cardiovascular morbidity is suggested for future research.