Effect of Health Alliances on the Quality of Primary Care in Urban China: a Coarsened Exact Matching Difference-in-differences Analysis
Min Su,Zhongliang Zhou,Yafei Si,Xiaolin Wei
DOI: https://doi.org/10.1016/s0140-6736(19)32422-5
2019-01-01
Abstract:BackgroundTo improve the quality of primary care and guide patients flowing to primary facilities, the Chinese Government called for piloting of health alliances—associations led by tertiary hospitals and composed of primary and secondary hospitals within a catchment area. We assessed the effects of health alliances, and compared loose versus tight health alliances, with respect to improving the quality of primary care provided through community health centres in urban China.MethodsTrained simulated patients presented to community health centres in Xi'an with one of two common outpatient diseases (unstable angina or asthma). Data from recorded patient–physician interactions were used to assess quality of primary care with respect to adherence to clinical checklists, accuracy of diagnoses, and appropriateness of treatment (each graded against checklists of recommended questions, examinations, diagnoses, and treatments), as well as proportions of unnecessary examinations and of harmful prescriptions, diagnosis and waiting times, and patient-centred communication scores (9-item Patient Perception of Patient-Centeredness scale). Difference-in-difference analyses combined with a coarsened exact matching method were used to identify the direct causal effects of health alliances and their various modes on primary care quality over time, after controlling for patient gender, physician gender, physician age, and health centre characteristics, as well as the fixed effects of disease, district, health centre, and survey time.FindingsBetween Aug 17 and Aug 28, 2017, and July 30 and Aug 10, 2018, eight simulated patients completed 484 interactions with physicians across 63 community health centres. Centres included in health alliances, compared with those not included, saw improvements in adherence to clinical checklists (0·36 percentage points increase, p<0·0001), correct treatments (4·18 percentage points increase, p=0·0020), proportions of interactions involving unnecessary examinations (0·81 percentage points decrease, p<0·0001) or harmful drug prescriptions (2·67 percentage points decrease, p=0·0360), and patient-centred communication scores (0·80-point increase, p<0·0001); whereas diagnosis times increased by 4·03 min (p<0·0001), and waiting times (1·11 min decrease, p=0·7910) and correct diagnoses (0·19 percentage points decrease, p=0·8650) were similar. Compared with loose health alliances, tight health alliances saw significantly improved adherence to clinical checklists (1·35 percentage points increase, p<0·0001) and correct diagnoses (0·61 percentage points increase, p<0·0001), decreased proportions of unnecessary examinations (1·65 percentage points decrease, p=0·0440), increased diagnosis times (0·36 min increase, p=0·0470), and improved patient-centred communication (1·35-point increase, p<0·0001), while other quality aspects were similar.InterpretationHealth alliances can improve the quality of some aspects of primary care by helping to sink the high-quality health resources of tertiary hospitals and secondary hospitals to primary health facilities. Tight health alliances are more likely to improve the quality of primary care, probably because of the increased collaboration between primary care facilities and larger hospitals and the unification of human resources, finances, and materials.FundingChina Medical Board (grant number 15-277 ), National Natural Science Foundation of China (grant number 71874137 ), Shaanxi Social Science Foundation (grant number 2017S024 ), Research Program of Shaanxi Soft Science (grant number 2015KRM117 ), National High-Level Talents Special Support Plan (Thousands of People plan), Shaanxi Provincial Youth Star of Science and Technology in 2016.