医院管理质量调查中调查对象的利益冲突对调查结果的影响

The impact of the respondents'conflict of interests on the hospital management survey results

目的:探究医院管理质量调查中调查对象利益冲突对调查结果的影响。方法:利用来自国家卫生计生委项目《急性冠脉综合征临床路径研究III期(CPACS-3)研究——医院管理子课题》81家医院管理基线调查数据及在此期间住院的4996例急性冠脉综合征患者资料,选择每家医院涉及2个或以上人员作为被询问对象的医院管理问题(共10个),比较不同被调查对象所选择答案间的一致性;进一步分析其答案与患者住院MACE发生率的关系,以探究调查对象利益冲突对调查结果的影响。结果:10个问题中,4个问题不同被调查者的回答存在显著统计学差异(P<0.10),且均观察到其中一方与所调查问题结果间存在明显利益冲突,而另一方则没有。在存在利益冲突者中,做出肯定回答组患者发生MACE的风险高于做出否定回答组(OR>1),不符合常理;而不存在利益冲突者中,做出肯定回答组患者发生MACE的风险则低于做出否定回答组(OR<1),符合常理。结论:被调查者对于所问问题有潜在利益冲突时,调查结果会被歪曲,应选择没有利益冲突的被调查者或选择客观的信息来源来避免此类偏倚。

Objective： To explore the impact of the respondents＇ conflict of interests on the hospital management survey results. Methods： Information which came from the study of Acute Coronary Syndrome Clinical Pathways Ⅲ and its hospital management sub-study, including 81 hospitals and 4996 ACS patients were analyzed, in which, all the questions that were administrated to more than 2 persons with different roles in the hospitals were selected and compared their answers to the same questions. Furtherly, the relationship of the answers with risk of in-hospital MACE were analyzed. Results： For 4 of the 10 questions analyzed, significant difference in the answers was found between different roles（P〈0.05）. Further, for these 4 questions, one of the respondents had showed apparent conflict of interest on the question, the others didn＇t. The one who had conflict of interest showed that the risk of in-hospital MACE for patients in the group with affirmative answers was higher than those in the group with negative answer（OR 〉1）, which did not meet the common sense; those without conflict of interest showed that the risk of in-hospital MACE for patients in the group with affirmative answers was lower than that with negative answer（OR〈 1）, consistent with common sense. Conclusion： When respondents are asked questions with potential conflict of interest, the survey results will be distorted. So we should selected respondents with no conflict of interest to answer the question, or choose an objective source of information to avoid such bias.