Beyond the labels: Classifying countries by child health outcomes – A cluster analysis of child mortality and child-health data

TL;DR

This study classifies countries based on child health outcomes using cluster analysis to identify patterns and inform policy improvements.

Contribution

The study introduces a novel approach to classifying health systems using child health outcomes rather than traditional organisational components.

Findings

Six main groups of countries were identified based on child and maternal mortality rates and other health indicators.

Countries performing better or worse than expected were identified through regression analysis.

The study highlights the complexity of child health and social care provision patterns.

Abstract

Most health service classification systems are based on organisational components such as service provision, financing, and regulation. This study considers health systems using data focusing on child health outcomes, service provision, and selected social characteristics. This more accurately reflects the reality of health service provision for children, young people, and their families.

To classify health systems based on child health data through cluster analysis and exploratory and descriptive data analysis.

Data were extracted from the current version of the UNICEF (2023) State of the World’s Children full dataset, concentrating on outcomes related to mortality. Cluster analyses were conducted, and a heatmap was produced to identify patterns and groups among countries and child health indicators. Row and column distances were calculated using the Euclidean distance, and clustering was performed using the complete linkage method. Each variable was centred and scaled using the scale command, allowing variables measured on different scales to be compared without those with large values being weighted more heavily. Countries that performed better or were less healthy than expected were identified through linear regression analysis using the ggplot2 package.

Analysis of countries by cluster reveals six main groups, characterised by child and maternal mortality rates, vaccination levels, access to maternal and child healthcare, access to water and sanitation, and population migration levels.

Identifying patterns in outcomes and identifying countries that perform above or below expectations concerning child health can inform a more nuanced approach to improving a country’s child health outcomes.

Main findings: Understanding child health and social care provision patterns and related outcomes is complex. Studies such as this may indicate further enquiry for identifying combinations of interventions that maximise health benefits.

Added knowledge: This study demonstrates that understanding the patterns of health provision and child health outcomes is instructive in developing policies related to child health.

Global health impact for policy and action: Policymakers and leaders within Health Systems must understand the reasons behind countries that have better-than-expected child health outcomes compared to current health expenditures. What are the nuances within countries that cause these patterns?

Figures