Indirect technique to estimate neonatal, infant, and under-five mortality rates for India and its states: A population-based cross-sectional study
- 1Department of Epidemiology and Biostatistics, KAHER, Belgaum, Karnataka, India
- 2Department of Sciences (Mathematics), Indian Institute of Information Technology Ranchi, Ranchi, Jharkhand, India
- 3Department of Statistics, Institute of Science, Banaras Hindu University, Varanasi, Uttar Pradesh, India
Abstract
Background: Accurate assessment of child mortality rates is essential for policy formulation, resource allocation, and tracking progress toward Sustainable Development Goal-3 (SDG-3). The study of child mortality provides useful information to know the demographic situation of country. The death is vital event and recorded through the civil registration system. But in many developing countries, the quality of registered death is not very much reliable due to illiteracy and ignorance in population. So, the lack of accurate registration of death has forced demographers to explore the indirect techniques for estimating child mortality. In this paper, authors have utilized an indirect technique for estimating neonatal, infant and under-five mortality rate by using data on proportion of dead children and proportion of 4 + birth-order. Methods: The method is mainly based on technique of linear line regression analysis, where the proportion of dead infants among all children born to currently married females (15–49 years) and proportion of 4 + birth order are taken as the independent variables and the neonatal mortality rate, or infant mortality rate, or under-five mortality rate are used as the dependent variable. This study is based on data collected in fifth round of National Family Health Survey (NFHS), 2019−21. After applying the inclusion criteria, a total of 512,408 currently married women aged 15–49 years were included in the analysis. Result: For above mentioned child mortalities are calculated for India as well its major states. The actual and predicted mortality rates overlapped substantially when both predictors were used together, indicating the suitability of the proposed model. The combined-predictor models (proportion of dead children and proportion of 4 + birth-order) achieved higher R² values (0.91 for neonatal mortality rate (NMR), 0.92 for infant mortality rate (IMR), and 0.96 for under-five mortality rate (U5MR) along with the minor percentage differences between observed and estimated rates (mostly within ±10%) than single-predictor models. Conclusion: The findings suggest that integrating both the predictors, proportion of dead children and the proportion of higher birth order children improves the accuracy of indirect estimates of NMR, IMR, and U5MR. This approach may serve as a practical tool for generating timely mortality estimates in contexts where civil registration and survey data are incomplete or infrequent.
Related
- A Recipe for a Good π. How to Properly Estimate Population Genetics Summary Statistics and Why we Should Systematically Report Them
- First density and abundance estimate for the only East African Population of Western bongo (Tragelaphus eurycerus eurycerus) in Semuliki National Park, Uganda
- International consensus guidance for general population screening for islet autoantibodies to diagnose early-stage type 1 diabetes: a nominal group technique process
- Do Harbour Porpoise Mortality Records Reflect Living Population Structure? A Matrix Population Model Diagnostic
- Climate and Other Human Causes of Spatial Variation in Mallard Mortality and Population Dynamics
BibTeX
@article{singh2026indirect,
title = {{Indirect technique to estimate neonatal, infant, and under-five mortality rates for India and its states: A population-based cross-sectional study}},
author = {Mayank Singh and Anuj Singh and Abhay Kumar Tiwari},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0353613},
url = {https://doi.org/10.1371/journal.pone.0353613},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Indirect technique to estimate neonatal, infant, and under-five mortality rates for India and its states: A population-based cross-sectional study
AU - Mayank Singh
AU - Anuj Singh
AU - Abhay Kumar Tiwari
PY - 2026
DA - 2026/09/15
JO - PLOS One
DO - 10.1371/journal.pone.0353613
UR - https://doi.org/10.1371/journal.pone.0353613
AB - Background: Accurate assessment of child mortality rates is essential for policy formulation, resource allocation, and tracking progress toward Sustainable Development Goal-3 (SDG-3). The study of child mortality provides useful information to know the demographic situation of country. The death is vital event and recorded through the civil registration system. But in many developing countries, the quality of registered death is not very much reliable due to illiteracy and ignorance in population. So, the lack of accurate registration of death has forced demographers to explore the indirect techniques for estimating child mortality. In this paper, authors have utilized an indirect technique for estimating neonatal, infant and under-five mortality rate by using data on proportion of dead children and proportion of 4 + birth-order. Methods: The method is mainly based on technique of linear line regression analysis, where the proportion of dead infants among all children born to currently married females (15–49 years) and proportion of 4 + birth order are taken as the independent variables and the neonatal mortality rate, or infant mortality rate, or under-five mortality rate are used as the dependent variable. This study is based on data collected in fifth round of National Family Health Survey (NFHS), 2019−21. After applying the inclusion criteria, a total of 512,408 currently married women aged 15–49 years were included in the analysis. Result: For above mentioned child mortalities are calculated for India as well its major states. The actual and predicted mortality rates overlapped substantially when both predictors were used together, indicating the suitability of the proposed model. The combined-predictor models (proportion of dead children and proportion of 4 + birth-order) achieved higher R² values (0.91 for neonatal mortality rate (NMR), 0.92 for infant mortality rate (IMR), and 0.96 for under-five mortality rate (U5MR) along with the minor percentage differences between observed and estimated rates (mostly within ±10%) than single-predictor models. Conclusion: The findings suggest that integrating both the predictors, proportion of dead children and the proportion of higher birth order children improves the accuracy of indirect estimates of NMR, IMR, and U5MR. This approach may serve as a practical tool for generating timely mortality estimates in contexts where civil registration and survey data are incomplete or infrequent.
SN - 1932-6203
ER -