Aveyard P
Medical School, University of Birmingham, UK.
J Eval Clin Pract. 1997 Nov;3(4):275-81. doi: 10.1046/j.1365-2753.1997.t01-1-00004.x.
Performance indicators for general practice which reduce complex processes to simple counts can have little validity. Additionally, performance indicators are often statistically unreliable in small populations like general practices. Instead, it is possible to combine these measures of performance by using multiple regression to predict the outcome from a set of processes. This allows one to adjust the outcome for differences in the practice populations. It also improves the statistical reliability, because data from all practices are used to predict the outcome. This approach has statistical problems, because it is an ecological analysis, and does not pick out the poor performers ('bad apples'). The regression approach is similar to the concepts of continuous quality improvement (CQI). It is arguable that using CQI to improve quality is more likely to lead to cooperation from general practices than trying to pick out the poor performers.
将复杂流程简化为简单计数的全科医疗绩效指标可能没什么效度。此外，在像全科医疗这样的小群体中，绩效指标往往在统计上不可靠。相反，通过使用多元回归从一组流程预测结果，可以将这些绩效衡量指标结合起来。这使得人们能够针对不同的执业人群调整结果。它还提高了统计可靠性，因为所有执业的数据都被用于预测结果。这种方法存在统计问题，因为它是一种生态分析，无法挑出表现不佳者（“坏苹果”）。回归方法类似于持续质量改进（CQI）的概念。可以说，与试图挑出表现不佳者相比，使用CQI来提高质量更有可能得到全科医疗的合作。
