Toward a conceptual framework of health and its operational definition: an application in the 1958 British birth cohort

Abstract

Background: Defining and measuring Health presents a challenge, partly due to its conceptual pluralism. To measure Health as an ability to adapt and self-manage, we developed an approach within the theoretical framework of resources and reserves over the life course, recently proposed in the literature. We aimed to (i) use the conceptual framework developed to identify indicators of deteriorating health reserves, (ii) construct an overall health measure from these indicators, (iii) evaluate the association between the overall health measure and subsequent health outcomes and (iv) assess the robustness of our method. Methods: We used data from 7,043 individuals born in 1958 in Great Britain included in the National Child Development Study. An overall health measure was constructed via the sum of three selected indicators of deteriorating health reserves in mid-life: chronic widespread pain (CWP), Clinical Interview Schedule - revised (CIS-r), and allostatic load (AL). A three-category variable was defined: impaired/medium/optimal overall health. We explored criterion validity by modelling the relationships between the overall health measure, or each reserve taken separately at 44–45 years, and self-rated health at 46 years and mortality up to 58 years, corresponding to 14 years of follow up, using Cox and logistic regressions respectively. We performed comparative analyses to assess the robustness of the method. Results: Having an impaired overall health measure was significantly associated with all-cause premature mortality (HRimpaired = 2.74 [1.86; 4.05]) and an increased risk of later fair/poor/very poor self-rated health (ORimpaired = 7.50 [6.29; 8.95]). The overall health measure had a greater effect on the self-rated health estimates than each indicator of deteriorating health reserves considered separately (ORAL medium = 1.82 [1.59; 2.09]; ORAL high = 2.74 [2.37; 3.16]; ORCIS-r = 5.20 [4.45; 6.08]; ORCWP = 2.85 [2.53; 3.21]). CIS-r and allostatic load were also associated with premature mortality contrary to chronic widespread pain (HRAL medium1.82 [1.27; 2.61]; HRAL high = 3.10 [2.19; 4.40]; HRCIS-r = 1.77 [1.22; 2.56]; HRCWP = 1.32 [0.98; 1.76]). The multiple comparative analyses conducted allowed us to assess the robustness of our method within this cohort. Conclusions: We proposed a method for measuring Health in mid-life in line with the concept of Health as the ability to adapt and self-manage and the concept of health reserves. This method may be applied and further developed within the field of social and positive epidemiology.