Staggs Vincent S, Cramer Emily
Health Services and Outcomes Research, Children's Mercy Hospitals and Clinics, School of Medicine, University of Missouri-Kansas City, 2401 Gillham Road, Kansas City, MO, 64108.
School of Nursing, University of Kansas Medical Center, Kansas City, KS.
Res Nurs Health. 2016 Aug;39(4):298-305. doi: 10.1002/nur.21727. Epub 2016 May 25.
Hospital performance reports often include rankings of unit pressure ulcer rates. Differentiating among units on the basis of quality requires reliable measurement. Our objectives were to describe and apply methods for assessing reliability of hospital-acquired pressure ulcer rates and evaluate a standard signal-noise reliability measure as an indicator of precision of differentiation among units. Quarterly pressure ulcer data from 8,199 critical care, step-down, medical, surgical, and medical-surgical nursing units from 1,299 US hospitals were analyzed. Using beta-binomial models, we estimated between-unit variability (signal) and within-unit variability (noise) in annual unit pressure ulcer rates. Signal-noise reliability was computed as the ratio of between-unit variability to the total of between- and within-unit variability. To assess precision of differentiation among units based on ranked pressure ulcer rates, we simulated data to estimate the probabilities of a unit's observed pressure ulcer rate rank in a given sample falling within five and ten percentiles of its true rank, and the probabilities of units with ulcer rates in the highest quartile and highest decile being identified as such. We assessed the signal-noise measure as an indicator of differentiation precision by computing its correlations with these probabilities. Pressure ulcer rates based on a single year of quarterly or weekly prevalence surveys were too susceptible to noise to allow for precise differentiation among units, and signal-noise reliability was a poor indicator of precision of differentiation. To ensure precise differentiation on the basis of true differences, alternative methods of assessing reliability should be applied to measures purported to differentiate among providers or units based on quality. © 2016 The Authors. Research in Nursing & Health published by Wiley Periodicals, Inc.
医院绩效报告通常会包含各科室压力性溃疡发生率的排名。基于质量对各科室进行区分需要可靠的测量方法。我们的目标是描述并应用评估医院获得性压力性溃疡发生率可靠性的方法,并评估一种标准的信噪比可靠性指标,以此作为各科室区分精度的指标。我们分析了来自美国1299家医院的8199个重症监护、逐步降级护理、内科、外科以及内科 - 外科护理科室的季度压力性溃疡数据。使用贝塔二项式模型,我们估计了各科室年度压力性溃疡发生率的科室间变异性(信号)和科室内部变异性(噪声)。信噪比可靠性通过科室间变异性与科室间和科室内部变异性总和的比值来计算。为了评估基于压力性溃疡发生率排名对各科室进行区分的精度,我们模拟数据以估计在给定样本中,一个科室观察到的压力性溃疡发生率排名落在其真实排名的5%和10%百分位数范围内的概率,以及溃疡发生率处于最高四分位数和最高十分位数的科室被正确识别的概率。我们通过计算信噪比指标与这些概率的相关性,来评估其作为区分精度指标的情况。基于单一年度季度或每周患病率调查得出的压力性溃疡发生率对噪声过于敏感,无法对各科室进行精确区分,且信噪比可靠性并非区分精度的良好指标。为了基于真实差异确保精确区分,对于旨在基于质量区分提供者或科室的指标,应采用其他评估可靠性的方法。© 2016作者。《护理与健康研究》由威利期刊公司出版。