Mental Illness and Mental Health in Later Life: Testing the Two Continua Model in India

Background/Purpose: Mental health and human development are increasingly conceptualized as the presence and impact of subjective well-being, not merely the absence of mental illness. Subjective evaluations, referred to as positive mental health, comprise three core domains: emotional, psychological and social well-being.  Yet public mental health policy and services still focus almost exclusively on prevention and treatment of mental illness.  Keyes et al. (2005, 2010) have proposed a “two continua model,” which holds that mental health and mental illness are related but distinct dimensions of well-being. The model has good empirical support in Western settings. The purpose of this study is to test this model with a national probability sample of older adults in India. We hypothesized that: 1) mental health and mental illness will be related but independent constructs, and 2) sociodemographic characteristics and health and functional status will influence mental illness and emotional, psychological and social well-being differently.

Methods:  Data are from the India component of the WHO Study on Global AGEing and Adult Health (SAGE).  The survey used a multi-stage, stratified clustered sample design, with household clusters sampled to reflect age, sex, level of wealth/local economic development, and urban/rural status in the India Census. Data are based on interviews with 7,150 adults aged ≥ 50 years in Wave 1 (2007- 2010). Measures are standardized, well-validated and widely used in world health surveys.  Independent variables (IV) include demonstrated sociodemographic correlates of mental illness and mental health, 7 chronic health conditions, cognitive functioning, and a general physical and functional health measure (World Health Survey ‘Health State’ (α=.90)). In hypothesis 1, we examined zero-order correlations of mental illness (12-month major depressive episode using CIDI, ver. 3), emotional well-being (Day Reconstruction Method); psychological well-being (WHO- Quality of Life Index (α = .86)); and social well-being (9-item Social Cohesion scale (α=.75)). To test hypothesis 2, we regressed the IVs on depression in a logistic model and regressed the IVs on each of the three measures of positive mental health using OLS.

Results: Data supported the hypotheses. Correlations of depression with positive mental health domains were significant (p<.01), but low to modest (range = .02 -.30). Likewise, all regression models were significant, but the only common correlates in the depression and positive mental health models were educational attainment, health state and cognitive functioning (p<.001). Additional significant correlates were: 1) depression: unemployed and angina and chronic lung disease; 2) emotional well-being: older age groups, employed and arthritis (negative); 3) psychological well-being: male and stroke, COPD or asthma (negative); and 4) social well-being (female, not employed, married, oldest age group, urban residence, and asthma (negative).

Conclusions and Implications:  The study extends the "two continua model" of mental illness and mental health theoretically by demonstrating its applicability to older adults in a different cultural setting and a developing country. It also contributes empirical knowledge and evidence for policy and practice by indicating specific foci for prevention and intervention strategies to address mental illness while promoting and preserving positive mental health.