查看文章

Objectives

Mapping allows the prediction of utilities using data from questionnaires without preference-based scoring systems. Datasets to calculate mapping algorithms can include multiple (repeated) observations from the same individual, and it is assumed that these are more similar to each other than measures from different individuals. However, standard statistical techniques used in the development of mapping algorithms require observations to be independent. Not accounting for repeated measurements can lead to inaccurate variance-covariance matrices. It is unclear if and how authors of mapping studies take into account interdependence between observations. This work aimed to investigate how published studies deal with this methodological issue and provide potential insights for future work.

Methods

We carried out a systematic literature search of databases MEDLINE, Cumulative Index to Nursing …