C Wilson, KM Rhodes, RA Payne Did the UK Quality and Outcomes Framework improve rates of recognition and treatment of cardiovascular risk factors in people with severe mental illness? C Wilson, KM Rhodes, RA Payne
Cardiovascular disease Severe Mental Illness e.g. schizophrenia, bipolar affective disorder Premature morbidity and mortality Life expectancy 12 years less Cardiovascular disease
Quality & Outcomes Framework Introduced 2004 Primary care “payment for performance” Incentivises annual review 2004 to 2011, unspecified 2011 to 2014, cardiovascular specific
Quality & Outcomes Framework Reduces inequalities in CVD risk factor testing Unknown Impact on recognition of CVD risk factors? Management of CVD risk factors?
? Have QOF indicators been associated with improvements in identification and management of CV risk factors in people with SMI?
Methods CPRD analysis UK general practice data 7% of population
Study design Retrospective open cohort study 1995 to 2014 Age ≥35 years Cases with lifetime diagnosis of SMI Random, unmatched controls (5:1)
Outcomes Diagnosis Treatment Diabetes Hypertension Obesity Elevated cholesterol Treatment Lipid-modifying medications Anti-diabetic medications
Interrupted time series analysis Segmented logistic regression Adjustment for age, gender, practice Time periods 1995-2003, pre-QOF 2004-2011, introduction of QOF 2011-2014, change to CV-specific indicators
Interrupted time series analysis Intervention 1 ↓ ↑ Intervention 2 Time
Results
Population characteristics SMI cases Controls Number 67,239 359,951 Follow-up 5.3 years 7.4 years Male 46.6% 49.7% Age (yrs) 1995/6 63.0 56.0 2004/5 57.6 55.6 2013/4 56.8 56.7 Proportion died 20.9% 10.8%
Recording elevated cholesterol 84%↑ Compared to non-SMI p<0.001 OR 1.21 (1.10-1.33) 37%↑ Compared to non-SMI p<0.001 [CLICK] Following the introduction of QOF in 2004, the [CLICK] odds of an SMI patient having elevated cholesterol test results were 1.2-times higher compared to before QOF. [CLICK] This was 37% higher than the change observed in the non-SMI group. The immediate effects appeared to be sustained post-2004. The cardiovascular-specific indicator in 2011 resulted in an increase in the odds of detection of high cholesterol which was 84% greater in the SMI group, but this effect was not sustained post-2011.
Case detection of risk factors Elevated cholesterol Diabetes 37%↑ 21%↑ 84%↑ Obesity Hypertension 21%↑ 19%↑ Similar to cholesterol, the introduction of the QOF incentives in 2004 was associated with increases in the recording of obesity, diabetes and hypertension in the SMI group compared to the non-SMI group, with these differences maintained. Following the change to the 2011 incentives, the odds of recording obesity increased in the SMI group compared to the non-SMI group, although this effect was not sustained post-2011. This was similar to what was observed for cholesterol. However the 2011 incentive did not appear to affect detection of diabetes or hypertension. 39%↑
Treatment of risk factors Anti-diabetic Rx Lipid-lowering Rx No association between QOF and changes in prescribing
Discussion
Improved identification of cardiovascular risk factors Under-recording of CV risk factors in SMI is known Impact of cardiovascular-specific indicators uncertain Possible “catch-up” effect?
No impact on treatment of cardiovascular risk factors At odds with national guidance Under-treatment of SMI population recognised e.g. stroke, arthritis Possible reasons Patient– e.g. cognitive impairment, adherence Physician – e.g. stigma, clinical complexity Service – e.g. care fragmentation, lack of resources
Limitations Can’t distinguish better case-finding from changes in incidence Age differences and changes in coding practice Effect of other unknown interventions (e.g. guidelines) Limited post-2011 data
Summary Incentives for GPs improve detection but not treatment of cardiovascular risk factors in SMI patients Effect of specifically incentivising treatment? Broader role of incentives in reducing inequalities and improving care for SMI patients?