Studying physician effects on patient outcomes: Physician interactional style and performance on quality of care indicators
Studying physician effects on patient outcomes: Physician interactional style and performance on quality of care indicators
复制标题
DOI：
10.1016/j.socscimed.2005.05.027
复制
发表时间：
2006-01-01
影响因子：
5.4
通讯作者：
Epstein, RM
中科院分区：
文献类型：
--
作者：
Franks, P;Jerant, AF;Epstein, RM
许多先前的研究表明，医生的互动方式和病人的结果之间的关系可能被混淆，因为仅仅依赖于病人的报告，每个医生检查很少的病人，或者没有证明医生对结果的影响。我们研究了医生的互动方式是否与护理质量指标的表现有关，通过患者报告和客观遭遇评分来衡量。我们还测试了医生对绩效指标的影响。使用100名美国初级保健医生（PCP）的数据，包括121,606名管理护理患者的索赔数据，4746名来访患者的调查数据，以及2名标准化患者与每位医生的录音会面，我们研究了基于索赔的护理质量指标与调查得出的患者对医生的看法和录音标准化患者会面中互动方式的客观评级之间的关系。多层次模型检验了医生对护理指标的影响（方差成分）是否由患者感知或互动风格的客观评分介导。我们发现与糖血红蛋白和胆固醇检测相关的显著医师效应。患者对医生更好的认知和更多的糖蛋白检测之间也有临床意义的关联。然而，多层次分析显示，医生对糖蛋白检测的影响并不是由患者感知的医生互动方式介导的。总之，与之前的研究类似，我们发现了患者对医生的看法与患者预后之间存在明显关系的证据。然而，在本研究中发现的患者对医生的看法与患者护理过程之间的明显关系并不能反映医生的风格，而可能反映了未测量的患者混淆。多层次建模有助于更好地理解医生风格与患者预后之间的关系。(c) 2005 Elsevier Ltd版权所有。
Many prior studies which suggest a relationship between physician interactional style and patient outcomes may have been confounded by relying solely on patient reports, examining very few patients per physician, or not demonstrating evidence of a physician effect on the outcomes. We examined whether physician interactional style, measured both by patient report and objective encounter ratings, is related to performance on quality of care indicators. We also tested for the presence of physician effects on the performance indicators. Using data on 100 US primary care physician (PCP) claims data on 1,21,606 of their managed care patients, survey data on 4746 of their visiting patients, and audiotaped encounters of 2 standardized patients with each physician, we examined the relationships between claims-based quality of care indicators and both survey-derived patient perceptions of their physicians and objective ratings of interactional style in the audiotaped standardized patient encounters. Multi-level models examined whether physician effects (variance components) on care indicators were mediated by patient perceptions or objective ratings of interactional style. We found significant physician effects associated with glycohemoglobin and cholesterol testing. There was also a clinically significant association between better patient perceptions of their physicians and more glycohemoglobin testing. Multi-level analyses revealed, however, that the physician effect on glycohemoglobin testing was not mediated by patient perceived physician interaction style. In conclusion, similar to prior studies, we found evidence of an apparent relationship between patient perceptions of their physician and patient outcomes. However, the apparent relationships found in this study between patient perceptions of their physicians and patient care processes do not reflect physician style, but presumably reflect unmeasured patient confounding. Multi-level modeling may contribute to better understanding of the relationships between physician style and patient outcomes. (c) 2005 Elsevier Ltd. All rights reserved.