.
.
. 2023;6 (6) :e2317055.
Introduction
The COVID-19 pandemic and subsequent national lockdowns in many countries disrupted access to basic services.1 Several welfare programs were put in place, even in resource-limited settings, to mitigate the socioeconomic and health consequences.2 To assess the overall implication of COVID-19 for population health, data ideally should be collected immediately before and after the outbreak. We used the 2019 to 2021 National Family Health Survey (NFHS)3 in India, a country with the second-highest number of COVID-19 cases and the third-highest death tolls in the world as of January 2023,4 to examine the systematic differences in various child health outcomes before vs after the outbreak.
Methods
The cross-sectional data of the 2019 to 2021 NFHS provided a unique opportunity to perform an empirical assessment, as the data were collected both before and after March 2020, when the national lockdown was declared in 13 of the 36 states or Union Territories in India, facilitating a natural comparison in health outcomes. For this cross-sectional study, data collected from June 17, 2019, to February 29, 2020, were defined as before COVID-19 and those from March 1, 2020, to May 20, 2021, were defined as after COVID-19. Further details on the survey design are available elsewhere.3 The Harvard Longwood Campus Institutional Review Board deemed this study exempt from ethics review because it was a secondary use of anonymized information. This study followed the STROBE reporting guideline.
Child health outcomes with a short reference period and deemed most likely to be affected by disruptions in health services were selected (Table 1). Twenty-six indicators related to pregnancy and child health and health care, feeding and nutrition, anthropometric failures, and vaccination were included. Absolute differences (in percentage points) were calculated by comparing the prevalence of outcomes before vs after the outbreak (eg, prevalence [stunting] after COVID-19 − prevalence [stunting] before COVID-19).
Statistical significance was determined using logistic regression models adjusted for child age, sex, and state fixed effects. To account for the multistage, stratified cluster-sampling design, survey weights were applied to all statistical analyses. Two-sided P = .05 indicated statistical significance. Analyses were performed between October 2022 and January 2023, using Stata 17 (StataCorp LLC).
Results
The sample size for the most complete outcome was 125 812 (65 574 boys [52.1%], 60 238 girls [47.9%]; mean [SD] age, 30.0 [17.6] months) (Table 1). Compared with before-COVID-19 data, after-COVID-19 data showed small but significant deterioration in neonatal mortality (0.49 percentage points), feeding and nutrition (eg, 4.22 percentage points reduction in solid or semisolid food intake), and anthropometric failures (eg, 1.87 percentage points increase in underweight) (Table 2). The most substantial difference was found in vaccination indicators, with 7.74 percentage points and 6.51 percentage points reduction in first dose of DPT (diphtheria, pertussis, tetanus) and polio, respectively. Other indicators, including many related to health services, either remained constant or marginally improved during the outbreak.
Discussion
Mixed results from this analysis suggested that adverse consequences of COVID-19 and national lockdown were countered, to some extent, by emergency relief programs. For example, the Indian government launched Pradhan Mantri Garib Kalyan Ann Yojana in 2020 to distribute 5 kg of food grains and 1 kg of pulses per month to approximately 800 million individuals (approximately two-thirds of India’s population).5 This initiative may explain the relatively constant or minimally worsened patterns in child nutrition status before and after the outbreak. It also underscored the need to sustain relief programs in nonpandemic times to promote children’s health. Improvements in child health outcomes, such as diarrhea and acute respiratory infection rates, may be attributed to the wider promotion of interpersonal hygiene during the pandemic.6
Study limitations included the cross-sectional design, which prohibited any causal inferences from being drawn, and the inability to distinguish COVID-19’s implications from those of longer-term exposures to harmful conditions. Nevertheless, the results showed that nationally representative surveys, even with COVID-19-related disruptions in data collection, can aid in understanding the pandemic’s outcome.