In this sample of disabled adult Medicaid recipients with heart failure, severe mental illness was not associated with differences in quality of care. In general, study participants had high rates of left ventricular assessment but low rates of prescription filling for medication shown to have benefit in heart failure. This is important as both ACE inhibitor or ARB use and beta-blocker use were associated with improved clinical outcomes.
In a single previous study assessing quality of care among Medicare heart failure patients with and without mental illness, individuals with mental illness had lower rates of left ventricular function evaluation as compared to individuals with no mental illness. However, as with our study, mental illness was not associated with any difference in use of ACE inhibitors or ARBs. The previous study included all mental illness diagnoses, [11
] while our study solely examined individuals with severe mental illness, who have a significantly increased risk of cardiovascular mortality as compared to the general population. [4
] There has been little research on quality of care for heart failure among individuals specifically with SMI, while studies of quality of care for other cardiovascular diseases among individuals with SMI have been inconsistent. [15
] Unlike many previous studies, our cohort consisted of Medicaid recipients with a disability diagnosis. As a result, both SMI and non-SMI individuals may have represented vulnerable populations with similar socioeconomic backgrounds, a factor which may mediate differences in quality of care among individuals with and without SMI. [10
] Furthermore, Medicaid insurance in general has been associated with inferior quality of care in heart failure. [19
] Our findings of no association between SMI status and quality of care may have been driven by similarities in insurance, socioeconomic status, and other risk factors in the exposed and unexposed populations.
While over 80% of individuals in our study had assessment of left ventricular function, utilization of medications shown to improve outcomes in heart failure was strikingly low: 52.2% of the cohort filled a prescription for either an ACE inhibitor or ARB and 45.5% used a beta-blocker. Studies which have looked at medication utilization have found a large range of compliance with evidence based medications. For comparison, use of ACE inhibitor therapy ranged from 36% to 80% and beta-blocker therapy from 23% to 86% among similar studies in the outpatient setting. [22
] In our study, not only was recommended medication usage found to be low, but medication continuity among takers was also suboptimal. We found that, among people who had used an ACE inhibitor or ARB, on average these medications were only being used 56% for the time on study. Among users of beta-blockers, prescriptions were filled about half of the time. Given our findings, policy should focus on both medication prescribing and continuity among Medicaid enrollees.
Individuals with SMI did not have significant differences in clinical outcomes as compared to individuals without SMI. Previous studies have shown an association between worse outcomes in heart failure both among individuals with any mental disorder [11
] and among individuals with depression. [28
] Our study differs from this previous work by exclusively addressing individuals with SMI as opposed to all psychiatric conditions. Furthermore, both SMI and non-SMI groups in our study had high prevalence of comorbidities including diabetes, cancer, and chronic pulmonary disease, placing all individuals at high risk of clinical events. Similarities in comorbidities and socioeconomic status between individuals with and without SMI may have precluded our ability to detect any differences in clinical outcomes between these two groups.
The quality measures in this study did not play a significant role in mediating an effect of SMI on clinical outcomes. However, we found that all three quality measures were associated with improved clinical outcomes. A study by Fonarow and colleagues questioned the clinical effectiveness of current quality measures. [32
] In that study of patients at discharge, most quality indicators, including left ventricular assessment, had no effect on clinical outcomes. In our study, assessment of left ventricular function was associated with adverse outcomes, a finding that was likely related to disease severity. Fonarow and colleagues did find modest clinical improvements with use of ACE inhibitors, ARBs, and beta-blockers. Similarly, both medication interventions in our study were found to be associated with a trend toward improved clinical outcomes. Utilization of ACE inhibitors or ARBs was associated with a 9% reduction in hospitalization rate and a 26% reduction in mortality, a benefit similar to that observed in clinical trials. [33
] Beta-blocker therapy was associated with a 5% reduction in hospitalizations and a nonsignificant trend toward improvement in both readmissions and mortality.
Our study has several limitations. Use of claims data did not allow us to access information related to systolic function. Given the substantial proportion of heart failure patients with normal ejection fraction in the community, [34
] many of our study participants may have had normal left ventricular systolic function and may not have met heart failure guideline indications for ACE inhibitors or beta-blockers. Nonetheless, our study sample included a substantial proportion of individuals with diabetes, hypertension, and coronary artery disease, who also have indications for these medications. Additionally, our results showed an association between medication use and improvement in clinical outcomes in this cohort.
Although our Medicaid data was rich in information, its limitations may have led to other biases. Claims data may not have perfect precision in assessment of quality indicators. Lack of precision should be random and non-differential, which would reduce power to detect a difference between individuals with and without SMI. Some of our findings may have resulted from unmeasured confounders or residual confounding. Potential contraindications to ACE inhibitors, ARBs, or beta-blockers may have explained why some individuals did not receive a prescription. Nonetheless, several previous studies have similarly used Medicaid administrative data both to evaluate the use of recommended therapies for heart failure [26
] and to associate pharmacotherapy for heart failure with hospitalization outcomes [37
] in the general medical population.
From a quality improvement standpoint, we cannot determine if poor performance was related to physician or patient factors. Further research will need to determine the mechanisms for suboptimal care in heart failure among Medicaid participants.
In conclusion, we found that severe mental illness in heart failure was not associated with reduced quality of care or clinical outcomes among a cohort of disabled Medicaid participants. The entire cohort of heart failure patients had low rates of therapy with ACE inhibitors or ARBs and beta-blockers, despite the fact that these medications were associated with improved mortality. Further research should address why Medicaid recipients with heart failure receive low rates of beneficial medications. Quality improvement programs should consider how best to incorporate the disabled Medicaid population to ensure that they receive optimal medical care.