An assessment of morbidity registers for coronary heart disease in primary care
Moher M., Yudkin P., Turner R., Schofield T., Mant D.
Background. Organised care delivered systematically to all patients with established coronary heart disease (CHD) can reduce their risk factors and improve their quality of life. Therefore, identifying all patients with established CHD in a general practice population is an important first step for delivering this effective healthcare. However, there is little information on how registers are compiled, the factors that predict inclusion on the register or the relationship between registration and level of care provided. Aim. To assess the completeness of morbidity registers for CHD in primary care, the factors that predict inclusion on the register, and the relationship between registration and level of care provided. Method. Observational study at baseline of 1979 patients aged 55 to 75 years with established CHD in 18 general practices recruited for a cluster randomised controlled trial. Results. The proportion of CHD patients correctly identified on practice morbidity registers varied from 29.3% to 100%. Four factors were significantly and independently associated with being on a register: a relevant surgery contact since diagnosis (OR = 2.1, 95% CI = 1.6%-2.9%); a relevant repeat prescription since diagnosis (OR = 1.6, 95% CI = 1.1%-2.3%); a diagnosis of myocardial infarction (OR = 1.5, 95% CI = 1.2%-1.9%); and a revascularisation procedure (OR = 1.5%, 95% CI = 1.1%-2.0%). Inclusion on a register was strongly associated with being adequately assessed (i.e. assessed for smoking status, blood pressure, and cholesterol) (OR = 1.8, 95% CI = 1.3%-2.3%) and with treatment with aspirin or a lipid-lowering agent (OR = 1.4 for each agent). Conclusion. A wide variation in registration levels between practices exists. There is evidence that practices using multiple methods of case detection achieve higher levels of registration. The association between registration and better care does not prove causality but an effective call-recall system is impossible without complete registration.