India’s new Annual Health Survey: the largest survey sample in the world?

Given India’s extensive intra- and inter-district variability, district-level planning is critical for health systems delivery and strengthening.  In the absence of strong district-level vital statistics data, state estimates are used for planning and impact measurement, thereby diluting information about some of the high-focus areas in the state.  Given that the national District-level Household Survey provides information only every five years, and focuses on maternal and child health services and uptake, there has been growing demand for vital statistics that are more regularly and locally reported, for use in planning and evaluation.   Four of our five Model Districts are included in these states; Andhra Pradesh is not because it is not an EAG state or in the high-priority Northeast region.

India’s Annual Health Survey (AHS) is the first survey initiative to provide district-level fertility and mortality statistics, by annual panel, in nine of high-focus states.  These include all eight of the Empowered Action Group (EAG) states, and Assam. The survey is governed by the Registrar General’s office, in collaboration with the Ministry of Health and Family Welfare.  The AHS baseline sampled 18.2 million people (3.6 million households) in all 284 districts of the selected nine states.  District-level sample sizes were derived from the state’s known Infant Mortality Rates (IMR).  The survey selected over 20,000 sample units – villages in rural areas, and Census Enumeration Blocks in urban areas.  Given the size, it is said to be the largest sample survey in the world.  The survey reports on vital indicators, including: total crude birth and death rates, sex ratios (at birth, 0-4, and all ages), and mortality rates (maternal, infant, neonatal, and under-five).  All indicators are disaggregated by urban/rural and male/female.

Some highlights from the national data:

• CBR: 82% of districts fall between the range of 19-28 childbirths per 1000 people, per year

• IMR:  the top 100 districts in IMR constitute about 50% of total infants deaths across the sampled states.  79% of districts range between 45 and 74 infant deaths per 1000 live births; 6-7 of these deaths are in neonates.  68% of districts report a range of 30-49 neonatal births per 1000 live births.

• U5MR:  66% of districts fall in the range of 61-100 deaths of under fives per 1000 live births; the top 100 districts make up 63% of child deaths in the sampled states

• 44% of districts have a sex ratio at birth lower than 900 females to 1000 males; in 33 districts the SRB is below 850

Al indicators demonstrated wide variability between districts and states (for example, 7 times the variability across states in NNMR), and poorer indicators in rural areas across the board.

So, how did the Model Districts report?

Select indicators are demonstrated below.  Of note, 57 ‘hot spot’ districts were designated both in the top 100 IMR districts and top 25 administrative divisions for MMR – our Model District Morigaon is included here.

We applaud the government’s effort to capture key and timely district-level data, given the methodological and logistical feat. This is critical information for planning district-targeted interventions and scale-up.  We imagine the survey will gain momentum; this year’s release didn’t get much publicity, despite (or maybe because it is) reporting poorer sex ratios and vital statistics than other survey efforts in the country.

You can read more about the AHS methodology, and see each state’s report, here.

For updates from India, follow us on Twitter @modeldistricts.