Data on the costs of

Data on the costs of mental disorders could motivate changes in mental health services. In the UK, economists and clinical researchers showed that increasing access to evidence-based psychological services (eg, cognitive behavioural therapy) would be nearly cost-neutral for the UK Government because of increased productivity and tax revenue and reduced social service costs (mainly health care and welfare benefits). The findings were used in a large public advocacy campaign that resulted in the UK Government committing publicly to investment in psychosocial treatments and, subsequently, nationwide scale-up of cognitive behavioural therapy. Similar, detailed country-specific economic evidence could affect policy in low-income and middle-income countries.
Child undernutrition in developing countries remains a persistent problem. It contributes importantly to child mortality and carries long-term consequences for malnourished children, including reduced cognitive development, worse economic outcomes, and lower offspring birthweight. In 2011, an estimated 165 million children in developing countries were stunted and 101 million children were underweight. In their analysis in , Sebastian Vollmer and colleagues assess the (often implicit) claim that economic growth will automatically lead to declines in child undernutrition: if poor nutritional outcomes are a manifestation of the poverty of nations, then surely economic growth would offer a remedy? Marshalling individual-level data from 121 Demographic and Health Surveys, Vollmer and colleagues document at best a weak, and often absent, association between economic growth and reductions in stunting, underweight, or wasting. This latest paper is not the first to URMC-099 cost doubt on the optimistic view that economic growth will necessarily translate into better anthropometric outcomes; indeed, even the current authors have previously presented similar evidence in India and sub-Saharan Africa. The merit of the present paper rather is in the range of the empirical exercise, specifically the extensiveness of the data and the range of statistical tests implemented. The null to quantitatively weak association between economic growth and child undernutrition documented is not for want of statistical precision or due to model and sample choice: Vollmer and colleagues report narrow confidence bounds of their estimates and robustness to a wide range of data concerns, thus increasing confidence in their results. This disappointing, but perhaps expected, result raises the crucial question: why is it that growth and reductions in child undernutrition are so weakly correlated? Vollmer and colleagues posit three possible channels: (i) perhaps households do not spend incomes in a way that is effectively targeted at improving nutrition; (ii) perhaps unequal distribution of growth within countries leaves poorer households, in which children are most at risk of undernutrition, unaffected; and (iii) rising national incomes might be poorly correlated with the public investments that are necessary to reduce child undernutrition. Each of these channels seems intuitively plausible. From a policy perspective, however, distinguishing between them is essential for the design of effective interventions. In the rest of this Comment, I seek to do this through the lens of India, where high-income growth has been accompanied by a slow decline in undernutrition. Results from India might have broader implications, much as Vollmer and colleagues\' latest analysis echoes previous results across Indian states. Rising incomes might indeed not be accompanied by increases in nutritional intake. In 1987, a notable study had reported low nutrient-income elasticities in rural Indian communities; in the period since, until 2005, per capita consumption of calories and protein in India declined in absolute terms, despite a manifold increase in incomes and despite a steep decline in the proportion of households reporting a lack of food; although anthropometric indicators have improved, the decline has been slow and absolute levels remain high.