Variations in patient-reported physical health between cardiac and musculoskeletal diseases: systematic review and meta-analysis of population-based studies.

Population-based assessments of physical health are important to evaluate healthcare resource allocation. Normative data on the level of physical impairments attributable to specific diseases and severity levels within these diseases is critical to interpreting such data. Our objective, by means of a systematic review and meta-analysis, was to test the hypothesis that specific diseases which form cardiovascular and musculoskeletal disease spectra are associated with gradients of physical impairments. We examined a cardiovascular disease spectrum which consisted of hypertension, ischaemic heart disease and heart failure, and a musculoskeletal disease spectrum of lower back pain, osteoarthritis and rheumatoid arthritis. Using Medline, EMBASE and CINAHL databases, articles which had examined these morbidities and used either the SF-12 or SF-36 in general or primary care populations were selected; data was extracted independently by three reviewers. Study characteristics were described and the mean physical component summary scores of the SF-12 or SF-36 was analysed by disease, using random-effects meta-analysis. The association between disease and physical health (mean physical component summary scores) was assessed using multilevel meta-regression analysis, adjusting for age, health setting, country, disease definition and SF-12 or 36 format. From this search, 26 articles were identified, yielding 70 separate estimates of mean physical component summary scores across the morbidities from 14 different countries. For the selected conditions, pooled unadjusted mean physical component summary scores were: 44.4 for hypertension, 38.9 for ischaemic heart disease, 35.9 for heart failure, 39.5 for lower back pain, 36.0 for osteoarthritis and 36.5 for rheumatoid arthritis. The adjusted meta-regression showed mean physical component summary score difference for ischaemic heart disease of -4.6 (95 % confidence interval -6.0 to -3.2) and heart failure -7.5 (-9.1 to -5.9) compared to the hypertension category. For osteoarthritis -4.2 (-5.3 to -3.0) and rheumatoid arthritis -3.9 (-9.5 to 1.6) compared to the lower back pain category. Our findings provide the benchmark norms for the differences in physical health within and between disease spectra. Improved characterisation of the relative impact of individual conditions on physical health will facilitate public health assessments of chronic diseases as well as assessments of interventions using functional patient-reported outcomes.