Equipercentile Equating of Scores From Common Patient-Reported Outcome Measures of Role, Social, and Emotional Functioning in Patients With Cancer.
Patient-reported outcome measures (PROMs) are fundamental in evaluating cancer treatments within clinical research and practice. The existence of various PROMs that assess similar health domains presents challenges for comparing or aggregating study results and data. Score linking is a well-established method for addressing this issue by converting scores from the metric of one PROM to another.
To provide equipercentile equating of psychosocial scales from commonly used PROMs in cancer patients to facilitate data pooling and comparisons.
Cancer patients with diverse diagnoses from five European countries completed the EORTC CAT Core, EORTC QLQ-C30, FACT-G, SF-36, and HADS Emotional, Role, and Social functioning (EF; RF; SF) scales. For scales with Spearman correlations exceeding 0.70, scores were linked using equipercentile equating as provided by the R package "equate". To evaluate linking precision, Standardized Mean Absolute Errors (SMAEs) were calculated.
A total of 945 cancer patients were included in the analysis (mean age 58.4 years; 54.7% male). Bidirectional conversion tables were established for three pairs of SF/RF scales and 10 pairs of EF scales. Overall, linking was possible for most scales from the EORTC Measures, SF-36, and HADS, whereas the FACT-G allowed linking only for its Emotional Well-Being (EWB) scale.
The findings suggest greater linking ability for EF scales compared to RF and SF scales, which seem to be more heterogeneous in terms of content. The established crosswalks support the combined evaluation of data from different PROMs.
To provide equipercentile equating of psychosocial scales from commonly used PROMs in cancer patients to facilitate data pooling and comparisons.
Cancer patients with diverse diagnoses from five European countries completed the EORTC CAT Core, EORTC QLQ-C30, FACT-G, SF-36, and HADS Emotional, Role, and Social functioning (EF; RF; SF) scales. For scales with Spearman correlations exceeding 0.70, scores were linked using equipercentile equating as provided by the R package "equate". To evaluate linking precision, Standardized Mean Absolute Errors (SMAEs) were calculated.
A total of 945 cancer patients were included in the analysis (mean age 58.4 years; 54.7% male). Bidirectional conversion tables were established for three pairs of SF/RF scales and 10 pairs of EF scales. Overall, linking was possible for most scales from the EORTC Measures, SF-36, and HADS, whereas the FACT-G allowed linking only for its Emotional Well-Being (EWB) scale.
The findings suggest greater linking ability for EF scales compared to RF and SF scales, which seem to be more heterogeneous in terms of content. The established crosswalks support the combined evaluation of data from different PROMs.
Authors
Rothmund-Grenier Rothmund-Grenier, Pilz Pilz, Schlosser Schlosser, Arraras Arraras, Groenvold Groenvold, Holzner Holzner, Van Leeuwen Van Leeuwen, Petersen Petersen, Schmidt Schmidt, Young Young, Rose Rose, Cella Cella, Giesinger Giesinger,
View on Pubmed