Temporal changes in health resource allocation across 11 prefecture-level cities of Jiangxi Province, China, 2015–2024: a descriptive ecological panel study using entropy-weighted TOPSIS and weighted rank-sum ratio methods

Assessments of health resource allocation in a Chinese province are usually cross-sectional, and entropy-weighted composite scores are rarely comparable across years because weights and normalisation ranges are re-estimated annually. We describe changes in the level and inter-city distribution of four per-capita indicators across the 11 prefecture-level cities of Jiangxi Province, 2015–2024. The design is explicitly descriptive and cannot identify the effect of COVID-19. Descriptive ecological panel study of 110 city-year observations. Indicators were beds, practicing (assistant) physicians and registered nurses per 1,000 population, and budget expenditure on health per 1,000 population. Inter-city inequality was summarised by the coefficient of variation (CV), Gini coefficient and Theil index, and σ-convergence was tested by regressing the annual CV on year. To make scores comparable across years, entropy weights and ideal solutions were estimated once from the pooled 110 × 4 matrix (longitudinal variant); the conventional year-by-year variant was retained only for comparison. Cities were ordered by entropy-weighted TOPSIS and by weighted rank-sum ratio (WRSR), agreement between them was measured by Kendall’s W and Spearman’s ρ, and imprecision by 1,000 Dirichlet bootstrap replications of the weights. Robustness was assessed across equal, CRITIC and entropy weights, three normalisation rules, leave-one-indicator-out specifications and the period definition. All four indicators rose in all 11 cities. Mean beds rose from 5.10 to 7.07 per 1,000 (+ 38.60%), physicians from 1.92 to 2.71 (+ 40.93%), nurses from 2.34 to 3.40 (+ 44.87%) and health expenditure from 102.71 to 144.63 ten-thousand CNY per 1,000 (+ 40.81%) (paired Wilcoxon P = 0.001 for all four). Inter-city inequality narrowed for the three workforce and facility indicators: CV fell from 13.42% to 6.26% for beds, 18.92% to 8.29% for physicians and 22.81% to 12.77% for nurses (σ-convergence slope − 1.47 to − 1.88% points per year, all P < 0.001), with Gini coefficients falling correspondingly. Health expenditure did not converge (CV 8.89% to 9.01%; slope + 0.14, P = 0.665). Under the pooled specification city scores were far more stable than under the year-specific one, indicating that the large annual swings of the conventional analysis arose mainly from re-standardisation. Agreement between the two ranking methods was high (Kendall W 0.918–0.995) but declined over time (Spearman ρ 0.973 in 2015, 0.836 in 2024). Bootstrap intervals spanned two or more positions for most mid-table cities, and rankings were sensitive to indicator selection: excluding nurses reduced the mean annual ρ with the base case to 0.811. Per-capita beds, physicians, nurses and health expenditure increased in every city, and the inter-city distribution of the three workforce and facility indicators became markedly more even, whereas that of health expenditure did not. Changes in relative standing were small and, for mid-table cities, statistically imprecise; they are not evidence of convergence of health need or of equitable access. The increases predate 2020 and coincided with nationwide health-system reforms, so no effect of the pandemic can be inferred. Policy attention should focus on expenditure and on nurse density relative to need rather than on movements in composite scores.

Authors

Institutions

Publication Details

Journal
BMC Health Services Research
Published
2026-11-18
DOI
https://doi.org/10.1186/s12913-026-16193-4
Primary Topic
Healthcare Systems and Practices
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Temporal changes in health resource allocation across 11 prefecture-level cities of Jiangxi Province, China, 2015–2024: a descriptive ecological panel study using entropy-weighted TOPSIS and weighted rank-sum ratio methods

Junjie Tao, Yvying Wan, Hongyan Yan
4 links
BMC Health Services Research
Healthcare Systems and Practices
article

Temporal changes in health resource allocation across 11 prefecture-level cities of Jiangxi Province, China, 2015–2024: a descriptive ecological panel study using entropy-weighted TOPSIS and weighted rank-sum ratio methods

Junjie Tao, Yvying Wan, Hongyan Yan
article en

Abstract

Assessments of health resource allocation in a Chinese province are usually cross-sectional, and entropy-weighted composite scores are rarely comparable across years because weights and normalisation ranges are re-estimated annually. We describe changes in the level and inter-city distribution of four per-capita indicators across the 11 prefecture-level cities of Jiangxi Province, 2015–2024. The design is explicitly descriptive and cannot identify the effect of COVID-19. Descriptive ecological panel study of 110 city-year observations. Indicators were beds, practicing (assistant) physicians and registered nurses per 1,000 population, and budget expenditure on health per 1,000 population. Inter-city inequality was summarised by the coefficient of variation (CV), Gini coefficient and Theil index, and σ-convergence was tested by regressing the annual CV on year. To make scores comparable across years, entropy weights and ideal solutions were estimated once from the pooled 110 × 4 matrix (longitudinal variant); the conventional year-by-year variant was retained only for comparison. Cities were ordered by entropy-weighted TOPSIS and by weighted rank-sum ratio (WRSR), agreement between them was measured by Kendall’s W and Spearman’s ρ, and imprecision by 1,000 Dirichlet bootstrap replications of the weights. Robustness was assessed across equal, CRITIC and entropy weights, three normalisation rules, leave-one-indicator-out specifications and the period definition. All four indicators rose in all 11 cities. Mean beds rose from 5.10 to 7.07 per 1,000 (+ 38.60%), physicians from 1.92 to 2.71 (+ 40.93%), nurses from 2.34 to 3.40 (+ 44.87%) and health expenditure from 102.71 to 144.63 ten-thousand CNY per 1,000 (+ 40.81%) (paired Wilcoxon P = 0.001 for all four). Inter-city inequality narrowed for the three workforce and facility indicators: CV fell from 13.42% to 6.26% for beds, 18.92% to 8.29% for physicians and 22.81% to 12.77% for nurses (σ-convergence slope − 1.47 to − 1.88% points per year, all P < 0.001), with Gini coefficients falling correspondingly. Health expenditure did not converge (CV 8.89% to 9.01%; slope + 0.14, P = 0.665). Under the pooled specification city scores were far more stable than under the year-specific one, indicating that the large annual swings of the conventional analysis arose mainly from re-standardisation. Agreement between the two ranking methods was high (Kendall W 0.918–0.995) but declined over time (Spearman ρ 0.973 in 2015, 0.836 in 2024). Bootstrap intervals spanned two or more positions for most mid-table cities, and rankings were sensitive to indicator selection: excluding nurses reduced the mean annual ρ with the base case to 0.811. Per-capita beds, physicians, nurses and health expenditure increased in every city, and the inter-city distribution of the three workforce and facility indicators became markedly more even, whereas that of health expenditure did not. Changes in relative standing were small and, for mid-table cities, statistically imprecise; they are not evidence of convergence of health need or of equitable access. The increases predate 2020 and coincided with nationwide health-system reforms, so no effect of the pandemic can be inferred. Policy attention should focus on expenditure and on nurse density relative to need rather than on movements in composite scores.

BMC Health Services Research
Second Affiliated Hospital of Nanchang University (CN), Gannan Medical University (CN), Gannan Normal University (CN), Jiangxi College of Applied Technology (CN), Gannan Health Vocational College (CN)
Reduced inequalities, Good health and well-being
Healthcare Systems and Practices
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon