The viewpoint expressed by Lee and Kavanaugh¹ is of great interest, and may even underestimate the potential importance of data concerning socioeconomic status and race in clinical trials. In one study low formal education level was associated significantly with higher joint counts, erythrocyte sedimentation rate, and patient questionnaire responses.² In another study, formal education was a more important identifier of poor physical function and high pain scores than age or duration of disease in patients with rheumatoid arthritis, scleroderma, systemic lupus erythematosus, fibromyalgia, and osteoarthritis.³ In the B-HAT study, education level was as prognostic of outcome over 3 years after a myocardial infarction, whether patients were randomised to a β blocker, propranolol, or placebo.⁴

Many physicians continue to believe that the primary reason for associations of low socioeconomic status and poor health is limited access to medical care.⁵ One explanation is that the classical “biomedical model”⁶ suggests that disease outcome and health in general is determined largely, if not entirely, by health professionals, with minimal contribution from patients. That certainly applies in acute care hospitals, the setting of most medical education and research. In chronic diseases, however, all experienced clinicians recognise that patient actions, attitudes, and behaviour contribute importantly to outcomes—the same physician and treatment may lead to very different results in different patients over long periods.

We have suggested that low socioeconomic status serves primarily as a marker to identify patient determined variables in health status rather than limited access to care.^7,8 The prevalence of most chronic diseases is higher in people of low socioeconomic status in most Western countries.⁷ Disparities have widened in recent years in the United Kingdom,⁹ the Netherlands,¹⁰ and the United States,¹¹ despite extensive programmes to reduce them. These and other data appear explicable only to a small extent on the basis of limited access to medical services.

Finally, the higher prevalence of most diseases in people of low socioeconomic status may be particularly relevant in rheumatic diseases, in which comorbidities are associated with poor outcomes such as premature mortality in rheumatoid arthritis.^12,13 We hope that the report of Lee and Kavanaugh will lead to more interest in socioeconomic status and race as important variables in clinical trials.