Medical practice variations in three public insurance schemes in Thailand

Variations in medical practice can be expected because patients have different health needs and preferences. However, variations may also occur because patient groups have different health insurance arrangements which can affect quality of care, equity in health, and financial burden on health systems and individuals.

Thailand is recognised by the global health policy and research community as a standout example among low- and middle-income countries (LMIC) in advancing health system development based on primary health care. The country's introduction of Universal Health Coverage in 2002 through multiple public health insurance schemes led to substantial enhancements in healthcare access and outcomes. This PhD thesis compares healthcare utilisation and selected outcomes in specialised care settings across Thailand's three public health insurance schemes: Civil Servant Medical Benefit Scheme (CSMBS), Social Health Insurance (SHI), and Universal Coverage Scheme (UCS). While all schemes offer comprehensive benefits, differences exist in their eligibility criteria, treatment options, provider choices and access rules, cost-sharing policies, and provider payment methods.

Employing mixed methods, the thesis includes a realist review and four quantitative studies. The review synthesises existing literature from Asian countries and Thai expert interviews into seven insurance features (benefit package, cost-sharing policies, beneficiaries, contracted providers, provider payment methods, budget size, and administration and management) influencing healthcare variation through 22 mechanisms.

Quantitative tracer conditions analyse national insurance claims data to explore variations in coronary revascularisation and mortality in ST elevation myocardial infarction (STEMI) patients, caesarean section (CS), benign hysterectomy, and total knee arthroplasty (TKA) across different insurance schemes. Poisson regression models estimate procedure rates in the population, and generalised linear models estimate absolute differences by insurance scheme. After adjusting for demographics, comorbidities, admission year, and health regions, significant differences in service utilisation across schemes were found. CSMBS-insured patients generally exhibited higher utilisation rates, followed by their counterparts in SHI and UCS. The variations were more pronounced in elective procedures compared to emergency and life-threatening services. Additionally, STEMI patients insured through CSMBS showed higher survival rates, followed by those insured through SHI, and then UCS.

Realist review findings shed light on some of the reasons for variations including mechanisms such as proximity to healthcare, socioeconomic status, and financial incentives. However, certain key mechanisms available only to CSMBS beneficiaries, notably direct access to specialists and greater opportunity to bypass public hospital waiting lists, may also influence their higher utilisation rates. These differing access rules have not been highlighted previously as key sources of variation in the Asian literature.

Policy recommendations include unification of public insurance schemes or harmonisation of specific insurance features to mitigate disparities, enhancing effective care coverage, short-term strategies to reduce access barriers, promoting health literacy among less advantaged populations, improving effectiveness in policy communication, and establishing routine monitoring of medical practice variation.

Future research could broaden the scope and depth of tracer conditions, notably by including other clinical and service utilisation outcomes, assessing appropriateness of preference-sensitive care, qualitatively exploring sources and mechanisms of variation, analysing variation across different hospital types, investigating cross-subsidisation between schemes within hospitals, and reinvestigating healthcare variation following changes in insurance scheme design. The design and theoretical framework of the thesis are well-suited for examining variations in other services, reassessing variations following significant changes in insurance features, and exploring other equity dimensions.

This thesis pioneers the adaptation and application of theoretical frameworks of practice variation research in LMIC contexts to explore variations in specialised care settings across diverse public insurance schemes using national datasets. The research findings add knowledge on how well universal health coverage is benefiting specific patient populations, and can guide policy not only for Thailand but also other countries facing similar challenges in healthcare service provision and health financing.