Rosa Derek, MacDermid Joy, Klubowicz Dorota
Physical Therapy, Western University, London, ON, Canada.
Roth McFarlane Hand and Upper Limb Center, St. Joseph's Hospital, London, ON, Canada.
Health Qual Life Outcomes. 2021 Apr 23;19(1):132. doi: 10.1186/s12955-021-01774-0.
Standardized coding of the content presented in patient reported outcome measures can be achieved using classification frameworks, and the resulting data can be used for ascertaining content validity or comparative analyses. The International Classification of Functioning (ICF) is a framework with a detailed conceptual structure that has been successfully utilized for such purposes through established coding procedures. The Item Perspective Classification (IPC) framework is a newly developed relational coding system that classifies the respondent perspective and conceptual domains addressed in items. The purpose of this study was to compare and describe the performance of these two frameworks when used alone, and in conjunction, for the generation of data pertaining to the content of patient reported outcome measures.
Six health-related quality of life questionnaires with a total of 159 items were classified by two raters using the Item Perspective Classification framework in conjunction with the International Classification of Functioning. Framework performance indicators included: classification capacity (percent of items amenable to successful classification), coding efficiency (number of codes required to classify items), and content overlap detection (percent of items sharing identical classification codes with at least one other item). Inter-rater reliability of item coding was determined using Krippendorff's alpha.
Classification capacity of the IPC framework was 97%, coding efficiency 26, and content overlap detection was 95%; whereas respective values for the ICF were 68%, 114, and 58%. When used in conjunction values were 63%, 129, and 30%. Krippendorff's alpha exceeded 0.97 for all 3 classification indices.
Inter-rater agreement on classification data was excellent. The IPC framework provided a unique classification of the respondent's judgment during item response and classified more items using fewer categories, indicated greater content overlap across items and was able to describe the relationship between multiple concepts presented within the context of a single item. The ICF provided a unique classification of item content relating to aspects of disability and generated more detailed and precise descriptions. A combined approach provided a rich description (detailed codes) with each framework providing complementary information. The benefits of this approach in instrument development and content validation require further investigation.
使用分类框架可以实现患者报告结局测量中呈现内容的标准化编码，所得数据可用于确定内容效度或进行比较分析。国际功能分类（ICF）是一个具有详细概念结构的框架，已通过既定的编码程序成功用于此类目的。项目视角分类（IPC）框架是一种新开发的关系编码系统，可对项目中涉及的受访者视角和概念领域进行分类。本研究的目的是比较和描述这两个框架单独使用以及联合使用时在生成与患者报告结局测量内容相关的数据方面的表现。
两名评分者使用项目视角分类框架并结合国际功能分类，对总共159个项目的六份健康相关生活质量问卷进行分类。框架性能指标包括：分类能力（可成功分类的项目百分比）、编码效率（对项目进行分类所需的代码数量）和内容重叠检测（与至少一个其他项目共享相同分类代码的项目百分比）。使用Krippendorff's alpha确定项目编码的评分者间信度。
IPC框架的分类能力为97%，编码效率为26，内容重叠检测为95%；而ICF的相应值分别为68%、114和58%。联合使用时的值分别为63%、129和30%。所有三个分类指标的Krippendorff's alpha均超过0.97。
评分者间在分类数据上的一致性非常好。IPC框架在项目回答过程中对受访者的判断提供了独特的分类，使用较少的类别对更多项目进行了分类，表明项目间的内容重叠更大，并且能够描述单个项目背景下呈现的多个概念之间的关系。ICF对与残疾方面相关的项目内容提供了独特的分类，并生成了更详细和精确的描述。联合方法提供了丰富的描述（详细代码），每个框架都提供了补充信息。这种方法在工具开发和内容验证中的益处需要进一步研究。
