Inequality in the distribution of health resources and health services in Chinahospitals versus primary care institutions

Here is a number that should stop you: China's Gini coefficient for the per capita distribution of hospital beds is around 0.17 — close to the kind of equality most health systems can only aspire to. Now here is the other number: the Gini coefficient for the geographic distribution of those same beds exceeds 0.82. Same country, same year, same hospitals. Two completely different pictures of equality. The question that Zhang and colleagues set out to answer is: which number is telling you the truth — and what happens to real people depending on which one a government decides to optimize for? China launched a major health system reform in two thousand nine with equity as a central goal. Despite achieving near-universal social health insurance coverage, the reform left a fundamental structural tension in place. Primary care institutions — community clinics and township health centers — make up more than ninety-six percent of all health facilities in China. In theory, that density should make them the backbone of equitable access. In practice, hospitals attract the bulk of quality resources: the beds, the trained specialists, and the technology. And forty-six point sixty-eight percent of tertiary hospitals are concentrated in the eastern provinces alone. Consumers know this. They routinely bypass primary care and seek hospital care even for minor illnesses because they distrust the alternative. Between two thousand ten and two thousand fourteen, hospital inpatient volume grew much faster than outpatient volume — a market signal that the gravitational pull toward hospitals was strengthening, not weakening. To measure how deep this problem runs, Zhang and colleagues used two complementary tools. The Gini coefficient runs from zero to one: below 0.3 is considered equitable, above 0.4 signals concern, and above 0.6 reflects severe inequality. They calculated it two ways — against population size and against geographic size — and that methodological choice turns out to be the crux of everything. The concentration index, or CI, is a different instrument designed to capture wealth-related inequality in service use. It also runs from negative one to positive one, where zero means no wealth tilt, a positive value means richer people use more of a service, and a negative value means poorer people use more. Together, these two tools let Zhang and colleagues separate the question of who has resources nearby from the question of who actually seeks care and where. The population map looks, at first, reasonably fair. Against population size, hospital sector Gini coefficients ranged from 0.17 to 0.44 across institutions, health workers, and beds between two thousand ten and two thousand fourteen. Some of those values fell modestly over the period — hospital beds dropped from a Gini of 0.26 in two thousand ten to 0.17 in two thousand fourteen, which is genuine progress. Primary care looked similar for institutions and beds, though health workers in primary care showed a slightly higher Gini, around 0.43 to 0.45, flagging that the distribution of clinical staff is less even than the distribution of facilities. Then you look at the geographic map, and the picture collapses. Gini coefficients for the distribution of institutions, health workers, and beds against physical area exceeded 0.7 in both the hospital and primary care sectors — every year, every indicator. Hospital health workers sat at 0.90 in two thousand ten and 0.88 in two thousand fourteen. Primary care institutions ranged from 0.77 to 0.89. Those are not mild inequalities. By the threshold Zhang and colleagues use, anything above 0.6 reflects a highly inequitable state. These numbers stayed stubbornly high across the entire study window. No meaningful improvement. What this tells you in physical terms is that enormous stretches of rural China — low population density and large geographic footprint — are nearly empty of doctors, hospitals, and beds. When resources follow people rather than land, and people cluster in cities, the map of care looks like a coastline: dense in the east, sparse in the west, with vast interior areas receiving almost nothing. This is where the concentration index sharpens the story. Zhang and colleagues found that the CI for outpatient visits to hospitals ranged from 0.16 to 0.21 across the study period — a consistent, moderate tilt toward wealthier populations. Richer people are choosing hospitals for routine outpatient care. Now look at the other end: the CI for inpatient care in primary care institutions ranged from negative 0.24 to negative 0.22. Poorer people are disproportionately being admitted to primary care facilities for inpatient care — the most acute episodes, the situations where resources matter most — in the settings that, by the paper's own data, lack the bulk of health workers and beds. Two other services — hospital inpatient care and outpatient visits to primary care — showed CI values near zero, between negative 0.02 and 0.02, indicating little wealth-related concentration. Sit with that asymmetry for a second. For high-acuity inpatient care, the poor end up in the lower-resourced tier. For routine outpatient care, the wealthy seek out the higher-resourced tier. The services are crossing, but not in a compensatory way — in a compounding one. The regional picture adds another layer. Zhang and colleagues divided China into three economic zones: the eastern developed region, the central developing region, and the western undeveloped region. The east had far higher resource density — hospitals per thousand square kilometers were roughly eight times those in the west — and per capita GDP in two thousand fourteen was seventy-one thousand seven hundred fifty-three yuan in the east versus thirty-eight thousand seven hundred eighty-eight yuan in the west. You might expect that wealth to smooth out internal inequalities. It didn't. The eastern region showed the highest internal inequality in service use, and concentration indices there tended to rise between two thousand ten and two thousand fourteen while those in the central and western zones declined. Decentralized budgeting and the fiscal advantage of wealthy provinces have continued to draw quality personnel and tertiary facilities toward the east — concentrating resources within an already advantaged region. Meanwhile, the western zone depends disproportionately on primary care for inpatient services. In two thousand fourteen, inpatient admissions in primary care institutions were three point three seven six percent in the western zone compared with one point four nine percent in the east. So the region with the least-resourced primary care is the one using it most for inpatient needs. Greater development has produced more resources overall in the east, but those resources are distributed more unevenly within the region, while the least developed west depends on poorly resourced primary care to meet inpatient needs. Zhang and colleagues draw four clear policy implications from all of this. First, more resources — especially quality health workers — need to move into primary care institutions to close the capacity gap. Second, regional disparities require central coordination through financial transfers because decentralized budgeting alone cannot redistribute across provincial lines. Third, the two-tier risk is real: the current dynamic, left unaddressed, builds a system where hospitals serve the wealthy and primary care serves everyone else — at lower quality. Fourth, a tiered delivery system needs to be developed so that patients access services by clinical need, not by ability to pay. The study has honest limitations. It covers two thousand ten to two thousand fourteen, using aggregate national yearbook data rather than individual-level records, which restricts causal interpretation. Some measures the authors wanted — like premature hospital discharge driven by financial barriers — were simply unavailable. But the core finding does not depend on those limitations. The per capita Gini and the geographic Gini are not measuring the same thing, and they are not producing the same answer. A reform that hits its population equity targets while ignoring geographic distribution will look successful on one metric while leaving rural and poor populations progressively further behind. Which number you optimize for is, in the end, a political choice. And the data suggest that China's reform, at least through two thousand fourteen, was optimizing for the number that looks better on paper. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.