Search

Developmental studies of mental disorders based on epidemiological data are often based on cross-sectional retrospective surveys. Under such designs, observations are right-censored, causing underestimation of lifetime prevalences and correlations, and inducing bias in latent trait models on the observations. In this paper we propose a Partial Likelihood (PL) method to estimate unbiased IRT models of lifetime predisposition to develop a certain outcome. A two-step estimation procedure corrects the IRT likelihood of outcome appearance with a function depending on (a) projected outcome frequencies at the end of the risk period, and (b) outcome censoring status at the time of the observation. Simulation results showed that the PL method yielded good recovery of true frequencies and intercepts. Slopes were best estimated when events were sufficiently correlated. When PL is applied to lifetime mental health disorders (assessed in the ESEMeD project surveys), estimated univariate prevalences were, on average, 1.4 times above raw estimates, and 2.06 higher in the case of bivariate prevalences.

Background: We examined the extent to which disability mediates the observed associations of common mental and physical conditions with perceived health., Methods and Findings: WHO World Mental Health (WMH) Surveys carried out in 22 countries worldwide (n = 51,344 respondents, 72.0% response rate). We assessed nine common mental conditions with the WHO Composite International Diagnostic Interview (CIDI), and ten chronic physical with a checklist. A visual analog scale (VAS) score (0, worst to 100, best) measured perceived health in the previous 30 days. Disability was assessed using a modified WHO Disability Assessment Schedule (WHODAS), including: cognition, mobility, self-care, getting along, role functioning (life activities), family burden, stigma, and discrimination. Path analysis was used to estimate total effects of conditions on perceived health VAS and their separate direct and indirect (through the WHODAS dimensions) effects. Twelve-month prevalence was 14.4% for any mental and 51.4% for any physical condition. 31.7% of respondents reported difficulties in role functioning, 11.4% in mobility, 8.3% in stigma, 8.1% in family burden and 6.9% in cognition. Other difficulties were much less common. Mean VAS score was 81.0 (SD = 0.1). Decrements in VAS scores were highest for neurological conditions (9.8), depression (8.2) and bipolar disorder (8.1). Across conditions, 36.8% (IQR: 31.2-51.5%) of the total decrement in perceived health associated with the condition were mediated by WHODAS disabilities (significant for 17 of 19 conditions). Role functioning was the dominant mediator for both mental and physical conditions. Stigma and family burden were also important mediators for mental conditions, and mobility for physical conditions., Conclusions: More than a third of the decrement in perceived health associated with common conditions is mediated by disability. Although the decrement is similar for physical and mental conditions, the pattern of mediation is different. Research is needed on the benefits for perceived health of targeted interventions aimed at particular disability dimensions.