Quality of mortality data in India

Assessment of Quality of Mortality Data from Large Surveys in India

About Us Objectives  Other_Information

In India, demographic changes and diversity across geographic space and social groups are well documented. The available research shows that the demographic diversity has widened over transition period.  Indian children born alive were expected to live hardly 20-23 years at the onset of the 20th Century. The decade of 1911-21 was particularly miserably as influenza killed millions of the Indian children. As against of 20-23 years of eoo in the early 20th century a new born in India can now live for more than 65 years; although still at a low level, but great improvement indeed.  Before 1911, IMR for India was more than 300 and under–5 mortality (U5MR) was more than 450 per 1000 live births; meaning nearly half of the children born did not survive beyond age 5. The latest estimate of IMR and U5MR for 2016 is 34 and 39 respectively; improved greatly, yet too high and far from the goal of SDG. The gaps in U5MR over geographies currently varied from lowest of 11 in Kerala to 55 in Madhya Pradesh.

Civil Registration System (CRS) is still not matured to provide complete count of vital events – births and deaths especially for children below age 5 years. CRS counted only around six million of total deaths against the expected little more than 8 million (completeness of only 65%). The system miserably fails as it could count only fewer than 2,00,000 infant deaths as against of one million infant deaths in India. Subsequently, Sample Registration System (SRS) was launched in the late 1960s and the system has been regularly providing the vital rates since 1970. Two points related to this are noteworthy. First, fertility and mortality were not published until 1981 for West Bengal and Bihar, two states sharing 17% of the Indian total population in 2011, and for smaller states the data quality has been under suspicion due to difficult areas and sample size. Second, during initial years of SRS (or before 1976) the stabilization of the system was very unlikely, and thus there was a greater fluctuation in annual vital rates.

Nonetheless, SRS was the only source of data on fertility and mortality condition in the country and for bigger states until the emergence of regular large sample surveys like National Family Health Survey (NFHS) in early 1990s and District Level Household Survey (DLHS) in the late 1990s. The SRS provides only level and trends in mortality and lack data on socioeconomic differentials in mortality. On the other hand, NFHS and DLHS have issues of sampling as well as non-sampling errors and have not been scrutinized, analyzed with rigours of mathematical modelling to detect inconsistency and make adjustment. These surveys, especially NFHSs have been used mainly to understand early child mortality using birth history data. The information on death/s in a household in last three years or so has been rarely used.

In the first and second NFHS conducted in 1992–93 and in 1998–99, death/s in a household was obtained but was discontinued in NFHS3 of 2005–06 and was reintroduced in the NFHS4 conducted in 2015–16. Similar questions were canvassed in DLHS2 (2002–04), DLHS3 (2007–08) and DLHS4 (2012–13). These data set provide unique opportunity to explore quality of mortality data and examine mortality condition over a period of nearly 25 years.

Funding Agency: Ministry of Health and Family Welfare, Government of India

Team of the Project:

Dr. Usha Ram
International Institute for Population Sciences, Govandi Station Road, Deonar
Mumbai – 400 088, Maharashtra, India

Dr. Manas Ranjan Pradhan
Assistant Professor
International Institute for Population Sciences, Govandi Station Road, Deonar
Mumbai – 400 088, Maharashtra, India


Back to Top
inner pages