The Sickest Older Adults Are the Least Likely to Get Community Care, Study Finds

One of the most troubling paradoxes in modern eldercare has a name that sounds like it belongs in a dairy, not a hospital: cream skimming. The term describes a practice long recognized in insurance markets, where providers quietly select the cheapest, healthiest customers and avoid the expensive, high-need ones. Now a new study suggests the same logic may be operating in China’s rapidly growing market for community-based services for older adults, with consequences that fall hardest on the very people who need help the most.

The research, published in BMC Geriatrics by Hang Li and Huijun Liu of Xi’an Jiaotong University’s School of Public Policy and Administration and Weihong Zeng of the university’s Jinhe Center for Economic Research, examines whether health deprivation, a measurable deficit in an older person’s physical or cognitive wellbeing, is linked to unmet needs for community services. The answer is stark: older adults experiencing health deprivation were significantly more likely to report that their needs for community services went unmet. In other words, the people with the greatest need were the least likely to be reached by the system designed to serve them.

To reach that conclusion, the team drew on one of the most valuable longitudinal resources in aging research: the Chinese Longitudinal Health Longevity Survey, or CLHLS. The study used an unbalanced panel of 14,762 respondents aged 60 or older, tracked across three survey waves in 2011, 2014, and 2018. An unbalanced panel means not every participant appears in every wave, which is an unavoidable reality when following people deep into old age, but it also allows researchers to capture far more of the aging process than a single snapshot ever could.

The analytical engine behind the findings was a set of fixed effects models, a statistical technique that deserves a moment of explanation because it is central to why the results matter. In observational data, older adults who are healthier differ from those who are deprived in countless ways: income, education, family structure, where they live. A naive comparison would tangle all of those differences together with the effect of health itself. Fixed effects models solve part of this problem by comparing each person with themselves over time. By focusing on within-person changes, the models effectively absorb every stable characteristic of an individual, observed or not, and ask whether periods in which a person experiences greater health deprivation are also periods in which their community service needs go unmet.

That design gives the finding considerable weight. The association between health deprivation and unmet service needs persisted even after stripping out the confounding influence of stable personal characteristics. And the study documented a second, downstream consequence: unmet needs for community services were themselves associated with an increased care burden on families. When formal services fail to arrive, the work does not disappear. It shifts, usually onto spouses, children, and other relatives, many of whom are aging themselves or balancing caregiving with employment.

The authors interpret their results through the lens of risk selection, a concept borrowed from economics and insurance theory. In a well-functioning long-term care market, services should flow toward need. But providers of home and community-based services face real costs, and serving frail, medically complex clients is more expensive and more difficult than serving healthier ones. When providers can identify who is likely to be costly, they have an incentive to select the profitable cases, the cream, and skim away from the rest. Health deprivation, in this framework, functions as a signal that providers use, deliberately or not, to sort potential service recipients. Those who carry the signal of high need end up with lower accessibility.

What makes this study distinctive is its demand-side framing. Much of the existing literature on cream skimming focuses on the supply side, on the behavior of insurers or providers. The authors argue that existing theories have not adequately explained why risk selection continues to plague what they call the premature long-term care market in China, a market still in an early stage of development where regulation, standardization, and financing mechanisms remain incomplete. Their findings support what they describe as an explanatory interpretation in which health deprivation serves as a selection signal, corresponding to lower accessibility to community services and, crucially, a pattern they call re-familization from the demand side.

Re-familization is a term worth unpacking, because it captures the quiet policy reversal the study documents. Over recent decades, many welfare states, China included, have worked to de-familize care, shifting responsibility from families toward formal institutions and public services. The findings here suggest a counter-movement operating not through legislation but through market behavior: when community services systematically bypass the most deprived older adults, the burden of their care snaps back to the household. Families become the residual claimants of a system that failed to select their relatives. The result is a double penalty, in which poor health both reduces access to formal support and increases the informal demands placed on loved ones.

The scale of the demographic backdrop makes these findings urgent. China’s aging population has dramatically increased demand for community services among older adults, and the country has invested heavily in home and community-based care as an alternative to institutional care, which is costlier and culturally less favored for many families. Community services, ranging from daily assistance to health-related support, are meant to be the connective tissue that allows people to age in place. If the allocation of those services follows a cream-skimming logic, the policy instrument designed to relieve family burden may instead be reinforcing it precisely where frailty is greatest.

Methodologically, the study’s reliance on three waves of a large national survey spanning seven years gives it a robustness that cross-sectional studies cannot match, and the fixed effects approach rules out a large class of alternative explanations. Still, the authors are careful about interpretation. The work uses secondary data from the publicly available CLHLS database, and the analysis establishes an association between health deprivation and unmet needs rather than a fully identified causal mechanism. The selection signal that providers respond to may operate through many channels, from visible frailty at the point of intake to the practical difficulties that deprived older adults face in navigating application processes. Disentangling those channels is a task for future research.

The implications, however, reach well beyond China. Rapidly aging societies across East Asia, Europe, and North America are all expanding home and community-based services, and all face versions of the same incentive problem: providers who can choose their clients will tend to choose the easy ones. The Chinese evidence offers a quantitative warning of what happens when that tendency goes unchecked. The oldest and sickest citizens, the ones for whom community care was conceived, drift back into the arms of their families, and the promise of a modern long-term care system quietly inverts itself. Recognizing health deprivation as a selection signal, the authors suggest, is the first step toward designing financing and oversight mechanisms that make the highest-need older adults impossible to skim away.

Subject of Research: Health deprivation and unmet needs for community services among older adults in China

Article Title: Cream skimming: how does health deprivation affect older adults’ access to community services?

Article References: Li, H., Liu, H., & Zeng, W. (2026). Cream skimming: how does health deprivation affect older adults’ access to community services?. BMC Geriatrics. https://doi.org/10.1186/s12877-026-08396-5

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08396-5

Keywords: cream skimming, health deprivation, unmet needs, community services, long-term care, older adults, aging population, re-familization, risk selection, China, CLHLS, gerontology