Factors and a configurational pathway associated with lower turnover intention among high-level medical professionals in county-level hospitals in China: a multicenter cross-sectional study.
Dong Shuxin S, Tai Qiuyuan Q, Yao Yiming Y, Shen Junlong J et al.
This study developed a context-specific framework of perceived conditions related to lower turnover intention among high-level medical professionals in county-level hospitals and examined the configurational pattern associated with a high retention-oriented score. A multicenter cross-sectional study combined three Delphi rounds, psychometric evaluation, and fuzzy-set qualitative comparative analysis (fsQCA). The Delphi process progressed from 33 items in round 1-37 in round 2 and 34 assessed in round 3, with 33 retained for the survey. From July 2024 to February 2025, online and paper questionnaires were distributed in 10 county-level public hospitals in Jiangsu Province. Of 635 questionnaires distributed, 611 were returned and 579 were valid. The sample was randomly split for exploratory factor analysis (EFA; n = 289) and confirmatory factor analysis (CFA; n = 290). fsQCA used the full sample, a frequency threshold of 10, consistency of 0.85, and proportional reduction in inconsistency (PRI) of 0.70. Thirty experts participated in Delphi round 1 and 29 in rounds 2 and 3 after one expert retired. The final 31 condition items showed high internal consistency (alpha = 0.971), suggesting both reliability and possible redundancy. Parallel analysis in the EFA half-sample suggested two factors, whereas a theory-constrained five-factor solution remained interpretable. In the CFA half-sample, the correlated five-factor model fit was marginal (Satorra-Bentler chi-square = 1350.48, df = 424, CFI = 0.903, TLI = 0.893, RMSEA = 0.087, 90% CI 0.082-0.092, SRMR = 0.042) and was similar to the second-order and bifactor models. No condition met the 0.90 necessity threshold. The corrected truth-table analysis identified one high-score configuration, PD*IES*CIS*OIE (consistency = 0.953; PRI = 0.917; coverage = 0.631). PD and CIS were core present conditions; IES and OIE were peripheral present conditions; IEPS was not included. A single sufficient configuration combined favorable perceptions of personal development, institutional/environmental support, compensation/incentive support, and organizational/interpersonal empowerment. The result does not support multiple substitutable pathways. Because the outcome was a retention-oriented score derived from a turnover-intention measure, the findings concern lower turnover intention rather than directly observed retention. The cross-sectional, individual-level results indicate associations and should not be interpreted causally or at the hospital level.