Clinical Trials Research Unit (CTRU), University of Leeds, Leeds, UK.
BMC Med Res Methodol. 2014 Feb 12;14:22. doi: 10.1186/1471-2288-14-22.
When developing new measuring instruments or deciding upon one for research, consideration of the 'best' method of administration for the target population should be made. Current evidence is inconsistent in differentiating superiority of any one method in terms of quantity and quality of response. We trialed a novel mixed methods approach in early scale development to determine the best administration method for a new patient-reported outcome instrument for people with pressure ulcers (the PU-QOL).
Cognitive interviews were undertaken with 35 people with pressure ulcers to determine appropriateness of a self-completed version of the PU-QOL instrument. Quantitative analysis, including Rasch analysis, was carried out on PU-QOL data from 70 patients with pressure ulcers, randomised to self-completed or interview-administered groups, to examine data quality and differential item functioning (DIF).
Cognitive interviews identified issues with PU-QOL self-completion. Quantitative analysis supported these findings with a large proportion of self-completed PU-QOLs returned with missing data. DIF analysis indicated administration methods did not impact the way patients from community care settings responded, supporting the equivalence of both administration versions.
Obtaining the best possible health outcomes data requires use of appropriate methods to ensure high quality data with minimal bias. Mixed methods, with the inclusion of Rasch, provided valuable evidence to support selection of the 'best' administration method for people with PUs during early PRO instrument development. We consider our approach to be generic and widely applicable to other elderly or chronically ill populations or suitable for use in limited samples where recruitment to large field tests is often difficult.
在开发新的测量工具或为研究选择工具时，应考虑针对目标人群的“最佳”管理方法。目前的证据在区分任何一种方法在数量和质量响应方面的优势方面并不一致。我们在早期量表开发中尝试了一种新的混合方法，以确定一种新的用于压疮患者的患者报告结局工具（PU-QOL）的最佳管理方法。
对 35 名压疮患者进行认知访谈，以确定 PU-QOL 仪器的自我完成版本是否合适。对 70 名压疮患者的 PU-QOL 数据进行定量分析，包括 Rasch 分析，将这些患者随机分为自我完成组或访谈管理组，以检查数据质量和差异项目功能（DIF）。
认知访谈发现了 PU-QOL 自我完成存在的问题。定量分析支持了这些发现，即大量自我完成的 PU-QOL 数据存在缺失。DIF 分析表明管理方法不会影响来自社区护理环境的患者的反应方式，支持两种管理版本的等效性。
获得尽可能好的健康结果数据需要使用适当的方法，以确保高质量的数据和最小的偏差。混合方法，包括 Rasch，提供了有价值的证据，支持在早期 PRO 工具开发过程中为 PU 患者选择“最佳”管理方法。我们认为我们的方法是通用的，可广泛应用于其他老年或慢性病患者群体，或者适用于招募大量现场测试通常很困难的有限样本。
