Outcome-based payment linked to better patient value in Taiyuan

A multicenter evaluation in Taiyuan, China, found that an outcome-based payment pathway for selected traditional Chinese medicine diseases was associated with a 4.85-point improvement on a patient-centered public value index. The difference-in-differences analysis, conducted among 7,985 inpatient cases, also found 520 RMB lower total hospitalization expenditure and 300 RMB lower out-of-pocket payment per admission under the new payment pathway. Patient satisfaction and perceptions of insurance benefit were also higher. The study, however, did not establish superior therapeutic efficacy for traditional Chinese medicine, statistically significant improvements in generic quality of life, savings for the insurance fund, or equivalent short-term safety.

The research compared the designated outcome-based pathway with conventional diagnosis-related group and diagnosis-intervention packet settlement, commonly abbreviated DRG/DIP, in a quasi-experimental study covering eight hospitals, four tertiary and four secondary, between January 2022 and December 2025. January 1, 2024, was treated as the policy cutoff, providing two years of observations before and after implementation. The five eligible cohorts were vertigo, cervical Bi disease, lumbar Bi disease, frozen shoulder and hemorrhoidal disease. These conditions belonged to a local payment catalogue, but the designation “TCM-advantage disease” did not itself demonstrate that traditional Chinese medicine was clinically superior for them.

Researchers linked medical insurance settlement records, discharge abstracts, electronic medical records and patient questionnaires using encrypted identifiers. The administrative cohort included 3,976 cases settled through the outcome-based pathway and 4,009 through conventional DRG/DIP, with 3,922 cases recorded before implementation and 4,063 afterward. A patient-reported survey included 545 respondents, while semi-structured interviews involved three patients, three traditional Chinese medicine clinicians, two hospital insurance managers and two medical insurance administrators. Propensity score matching produced 3,342 cases in each payment group and reduced all measured baseline standardized mean differences below 0.10, improving comparability before the policy-effect estimates were calculated.

The study’s primary measure was a newly constructed Patient-Centered Public Value Index, or PCPVI, combining five domains: health benefit, financial protection, care quality and safety, TCM service provision, and patient experience. A modified Delphi process involving 15 experts assigned weights of 30% to health benefit, 20% each to financial protection and care quality and safety, and 15% each to TCM-specific value and patient experience. All component indicators were placed on a 0–100 scale, with negatively framed measures reverse-coded. Internal consistency was reasonably strong, with a Cronbach’s alpha of 0.82, although the authors cautioned that such statistics alone do not validate the index or establish its clinical interpretability. No minimum important difference has yet been established for the composite score.

After matching, the main difference-in-differences estimate attributed to the payment pathway was a 4.85-point increase in PCPVI, with a 95% confidence interval of 3.32 to 6.38 and p<0.001. Total hospitalization expenditure was 520 RMB lower per admission, with a 95% confidence interval of negative 843 to negative 197 RMB, while out-of-pocket payment was 300 RMB lower, with an interval of negative 480 to negative 120 RMB. The reimbursement ratio rose by 2.40 percentage points, and mean length of stay fell by 0.26 days. Overall satisfaction increased by 0.18 points on the study’s scale. These associations indicate improved measured patient-centered value, but the design cannot eliminate every possible source of unmeasured or time-varying confounding.

Not every outcome changed significantly. The estimated effect on insurance fund payment was negative 130 RMB per admission, but its 95% confidence interval ranged from a 361 RMB reduction to a 101 RMB increase, and p=0.270. The result therefore does not demonstrate insurer savings, budget neutrality or long-term fiscal sustainability. The EQ-5D-5L estimate was 0.021, with a 95% confidence interval of negative 0.001 to 0.043 and p=0.061, so the analysis did not detect a statistically significant change in generic health-related quality of life. Nor were differences in 30-day readmission or in-hospital complications statistically significant. These non-significant findings do not prove equivalence, non-inferiority or the absence of clinically important effects.

Among the 545 survey respondents, those in the outcome-based group reported higher perceived insurance benefit, communication quality, acceptability of traditional Chinese medicine treatment and overall satisfaction. Mean satisfaction was 4.20 versus 4.02 in the conventional group, perceived insurance benefit was 4.06 versus 3.86, communication quality was 4.18 versus 4.02, and treatment acceptability was 4.25 versus 3.95. The groups did not differ significantly on the EQ-5D-5L index, which averaged 0.85 and 0.83, respectively, or on categorical measures of symptom improvement, functional recovery and willingness to choose the pathway again. The EQ visual analogue scale was higher in the outcome-based group, 76.6 versus 74.0, but questionnaires were administered at discharge, limiting conclusions about sustained recovery.

The cost of traditional Chinese medicine rose by 4.90 percentage points as a proportion of total hospitalization expenditure in the outcome-based group, compared with a smaller increase in the conventional group. This shift is a measure of service composition, not proof that additional traditional Chinese medicine produced better clinical outcomes. It could reflect more treatment volume, fewer non-TCM services, price changes or some combination of these. An exploratory analysis excluding TCM cost share from the index still produced a positive policy estimate of 3.72 points, while intensity-defined subgroup results were larger among cases with high TCM spending than among those with low spending. Because payment reform may itself influence spending intensity, and that measure contributes to the index, the contrast cannot establish a dose-response relationship or treatment efficacy.

PCPVI estimates were positive in all five disease cohorts, both age groups, both insurance categories and both hospital levels. The disease-specific estimates ranged from 3.20 points for hemorrhoidal disease to 5.80 points for lumbar Bi disease. Results remained similar under inverse probability weighting, entropy balancing, equal weighting, removal of extreme costs and inclusion of repeated admissions. A placebo policy date produced an estimate of 0.52 points, with a 95% confidence interval of negative 1.11 to 2.15. Although these checks supported the main finding’s stability, they did not prove that the groups followed parallel trends or that the policy cutoff was the only relevant change. National DRG/DIP version 2.0 guidance was also released in 2024, creating a potential source of concurrent policy effects.

Interviews offered context for the measured changes. Six of 10 stakeholders described reduced perceived out-of-pocket burden, seven reported greater recognition of TCM pathway value and five mentioned more standardized clinical pathway management. These accounts were consistent with several quantitative results but did not independently verify them. Implementation also brought burdens: five interviewees raised coding or documentation demands, seven called for dynamic adjustment of payment standards and four expressed concern about hospital cost-control pressure. The conclusions remain limited by the single-city setting, incomplete TCM documentation, possible questionnaire selection bias, 294 matched cases without complete 30-day follow-up, and the absence of long-term clinical and fiscal evaluation. The findings support further testing of outcome-based payment, but not broad claims of TCM-specific effectiveness, insurance savings or transferable policy benefits.

Subject of Research: Evaluation of outcome-based insurance payment for traditional Chinese medicine disease cohorts in Taiyuan, China

Article Title: Aligning insurance payment with public value: outcome-based payment for traditional Chinese medicine advantage diseases under China’s DRG/DIP reform

Article References: Pei, Z., Hu, A., & Duan, Z. (2026). Aligning insurance payment with public value: outcome-based payment for traditional Chinese medicine advantage diseases under China’s DRG/DIP reform. Frontiers in Public Health, 14, Article 1902781. https://doi.org/10.3389/fpubh.2026.1902781

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1902781

Keywords: outcome-based payment, DRG/DIP reform, traditional Chinese medicine, patient-centered value, medical insurance, financial protection, difference-in-differences, health services research, patient satisfaction, Aligning, insurance, payment