Karen Eggleston speaking to staff at Zhejiang Provincial CDC, China

Addressing Health Disparities in China

Insights from Machine Learning

Peering into China’s longevity strategies

A collaborative research project looks to Tongxiang county in Zhejiang, one of China’s most developed provinces, as a case study instructive for the country’s responses to healthcare inequalities and population aging challenges in rural and urban areas.

Research Focus

Tongxiang is a mostly rural county in Zhejiang, one of China’s most developed provinces. It is home to a population of about 700,000 registered residents, of whom the majority are rural (agricultural) hukou residents and the minority are urban (non-agricultural) hukou residents. The hukou system of household registration strongly marks and shapes socio-economic status and access to social programs in China. Hukou status is determined at birth, cannot easily be changed, and dictates many of the social programs to which a resident is entitled, such as education, healthcare, and old-age support. For example, historically, firms do not pay social security benefits for agricultural hukou workers but are required to pay for non-agricultural hukou workers.
 
How is China poised to tackle inequalities in healthcare spending and health outcomes among rural and urban populations? How will the country develop strategies for healthy aging in rural versus urban areas? Studying Tongxiang residents with urban and rural hukou may offer insights into these questions.
 
The Addressing Health Disparities in China project, led by Shorenstein APARC’s Asia Health Policy Program (AHPP), focuses on two areas:
 
First, it investigates the different health and medical care experiences of residents with urban and rural hukou. Tongxiang citizens have two main healthcare options available to them: the clinic and the regional hospital, and patients are free to choose which source of healthcare they utilize. The study will describe the different patterns of healthcare utilization by urban versus rural patients. It will also evaluate the efficiency of the current system by conducting a cost-benefit analysis of a hypothetical move to a “gatekeeper system,” a structure in which patients would have to see a doctor at the clinic in order to get a referral for a specialist.
 
This analysis requires an evaluation of how patients in Tongxiang make healthcare choices, which may differ according to their life-course exposures to urban and rural environments and their associated privileges and constraints. What do they value and what constraints do they face? Varying commute length, varying severity of illness, age, education, credit constraints, and individual insurance design all might influence a patient’s choice of visiting a clinic or a specialist. Evaluating these values may be the work product of creating general “maps” of how patients navigate the healthcare system. By analyzing how patients with chronic illnesses navigate their treatment in the current system, this study can help evaluate the system and make predictions about a hypothetical new one.
 
Second, the study assesses the impact of catastrophic coverage for inpatient services among those covered by resident insurance. This 2014 insurance expansion had two related effects on benefits: it removed the cap on how much insurance can cover and it significantly lowered the out-of-pocket payment requirements for resident insurance beneficiaries. The study examines the impact of this insurance expansion on Tongxiang citizens. It uses machine learning techniques and extensive data from medical claims as well as the local population health management system—frequently an unobserved, confounding factor in such studies.
 
Tongxiang’s health system features two social health insurance schemes: the employee health insurance scheme and the resident health insurance scheme. Most agricultural hukou residents in Tongxiang only qualify for the resident insurance scheme. The study also uses a difference-in-difference approach to explore the spending and health outcome changes before and after the 2014 insurance expansion.
 
The project builds upon a previous initiative of the Zhejiang Provincial CDC, AHPP Director Karen Eggleston, and Stanford economics student Yiwei Chen. The Zhejiang CDC-Eggleston collaboration compiled and linked several important individual-level and household-level datasets related to the health and survival of Tongxiang residents, resulting in the most comprehensive micro health data for studying health insurance schemes in China.
 

Lead Researchers and Collaborators

fsi_bio

Karen Eggleston

APARC Deputy Director and Asia Health Policy Program Director
close
fsi_bio

Karen Eggleston

APARC Deputy Director and Asia Health Policy Program Director
Senior Fellow, Freeman Spogli Institute for International Studies
Fellow at the Center for Health Policy and the Center for Primary Care and Outcomes Research
fsi_bio

Stefan Wager

Assistant Professor of Operations, Information, and Technology at Stanford’s Graduate School of Business
close
fsi_bio

Stefan Wager

Assistant Professor of Operations, Information, and Technology at Stanford’s Graduate School of Business
Assistant Professor of Statistics (by courtesy), School of Humanities and Sciences