A Multinomial Model for Comorbidity in England of Longstanding CVD, Diabetes, and Obesity

From a public health perspective, previous research on comorbidity tends to have focused on identifying the most prevalent groupings of illnesses that demonstrate comorbidity, particularly among the elderly population, already in receipt of care. In contrast, little attention has been paid to possible socio-economic factors associated with increased rates of comorbidity or to the possibility of wider unrevealed need. Given the known relationship between CVD, diabetes and obesity and the strong socio-economic gradients in risk factors for each of the three diseases as single morbidities, this paper uses the Health Survey for England to examine the demographic and socio-economic determinants of each of the seven disease combinations in the English population. Using a multinomial logistic model, this research finds that gender is a significant predictor for all seven disease combinations. However, gender was not as influential as individual age or socio-economic profile. With regard to ethnicity, the black population presents a high obesity, diabetes and diabetes-related comorbidity risk, whilst the Asian population presents a high risk for diabetes and diabetes-related comorbidity but a low risk for obesity and comorbidity. Across the seven disease combinations, risk was lowest for those individuals with a high income (4 out of 7), in-work (4 out of 7), home owning (3 out of 7) and degree educated (3 out of 7). Finally, smokers have a lower risk rate of obesity (and related) than ex-smokers relative to individuals that never smoked (in all cases controlling for all other factors). The important influence of socioeconomic factors has implications for the spatial demand for services and the policy solutions adopted to tackle the increasing prevalence of comorbidity.
Comment: Working Paper