Abstract
Ratio-based prevalence and absolute headcounts are the two most commonly accepted metrics to measure the burden of various socioeconomic phenomenon. However, ratio-based prevalence, calculated as the number of cases with certain conditions relative to the total population, is by far the most widely used to rank burden and consequently for targeting, across different populations, often defined in terms of geographical areas. In this regard, targeting areas exclusively based on prevalence-based metric poses certain fundamental difficulties with some serious policy implications. Drawing the data from the National Family Health Survey 2015–2016, and Census 2011, this paper takes four indicators of child undernutrition in India as an example to examine two contextual questions: first, does the choice of metric matter for targeting areas for reducing child undernutrition in India? and second; which metric should be used to facilitate comparisons and targeting across variable populations? Our findings suggest a moderate correlation between prevalence estimates and absolute headcounts implying that choice of metric does matter when targeting child undernutrition. Huge variations were observed between prevalence-based and absolute count-based ranking of the districts. In fact, in various cases, districts with the highest absolute number of undernourished children were ranked as relatively lower-burden districts based on prevalence. A simple comparison between the two approaches—when applied to targeting undernourished children in India—indicates that prevalence-based prioritization may miss high-burden areas where substantially higher number of undernourished children are concentrated. For developing populous countries like India, which is already grappling with high levels of maternal and child malnutrition and poor health infrastructure along with intrinsic socioeconomic inequalities, it is critical to adopt an appropriate metric for effective targeting and prioritization.
Highlights
Identification and prioritization of target areas across populations is a fundamental task in policymaking
The mean P of child stunting at the national level was 38.37% and it ranged from 13.12 to 65.20% across 640 districts in India with a standard deviation (SD) of 9.94% (Table 1; district estimates provided in Supplementary Tables S1-S5)
Based on A measure, we found on average A(stunting)= 93,394 children, A(underweight)= 84,976 children, A(LBW)= 40,062 and A(anemia)= 143,148 children, with substantial variation across districts
Summary
Identification and prioritization of target areas across populations is a fundamental task in policymaking. Ratio-based prevalence (P from hereafter), calculated as the number of cases with certain conditions relative to the total population, is by far the most widely used metric to rank the burden across different populations, often defined in terms of geographical areas (Davis and Hertz 1951; Minhas 1974; Dandekar and Rath 1971; Sen 1976; Arriaga 1970). Based on P, areas with a relatively higher prevalence of poor conditions, related to measures ranging from poverty (Minhas 1974; Dandekar and Rath 1971; Sen 1976), illiteracy (Carr-Hill and Pessoa 2008), unemployment, urbanisation (Arriaga 1970; Davis and Hertz 1951), morbidity, disability, and malnutrition (United Nations Children’s Fund 2012), are more likely to be prioritized for immediate interventions and greater resource allocation. Comparisons of disease burden across geographical regions and population groups are predominantly based on P metric (James et al 2018)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
