India’s under-five child mortality rate dropped 22 per cent in the last 10 years, reputed international journal Lancet has said — a drop steeper than government and UN estimates. But this hasn’t been welcomed as good news as experts have found fault with the model used to come to this figure.
On May 25, Lancet unveiled a study that said under-five child mortality in the world dropped from 11.9 million deaths in 1990 to 7.7 million in 2010. The figure for India was 1.64 million annual deaths, way below the government’s Sample Registration System (SRS) and UN figure of 1.83 million.
The study covered 187 countries. The Institute for Health Metrics and Evaluation at the University of Washington used the Gaussian Process Regression model, not commonly used to make health predictions. “The SRS sample size (71.0 mn), arguably the world’s largest, helps provide robust estimates...” said Thomas Chandy, CEO, Save the Children.
“The huge drop is a result of changes in the figure for half-a-dozen countries such as India, Afghanistan, Angola and Nigeria where the data was inconclusive,” said Harvard University data analyst Kenneth Hill in the Economist.
“Modelling cannot replace actual data,” said India’s NSSO official. World Bank population expert Eduard Bos said the IHME had made a similar claim in 2007 using another model, which too had been challenged.
The study’s lead author and IHME head Christopher Murray said the data was valid.