The current use and application of thresholds for clinical importance of the EORTC QLQ-C30, the EORTC CAT core and the EORTC QLQ-C15-PAL- a systematic scoping review.

BACKGROUND

Thresholds for clinical importance (TCIs) were previously established for the cancer-specific patient reported outcome (PRO) measures EORTC QLQ-C30, EORTC QLQ CAT Core, and EORTC QLQ-C15-PAL. TCIs aim to aid the interpretation of scores for individual patients at a single point in time. They intend to indicate whether a symptom or functional health limitation is of clinical relevance, i.e., requires to be discussed with healthcare professionals. In this systematic scoping review, we aimed to describe the uptake of TCIs by the research community and discuss opportunities and threats in their application to PRO data.

METHODS

We systematically searched PubMed and Web of Science databases that contained search terms on the respective PRO measures and TCIs. Additionally, we performed a hand search on citations of the original TCI articles on Google Scholar. Articles were included if they applied TCIs in the analysis or the interpretation of PRO data or in clinical practice. Data concerning the study design, the use of TCIs, the terminology, and the application of TCIs were extracted.

RESULTS

A total of 512 articles were identified. After title, abstract and full-text screening, data extraction was performed on 117 of these articles. Most articles reported on longitudinal-observational (n = 55) or cross-sectional observational (n = 49) studies, whereby the most frequent cancer populations having mixed diagnoses (n = 25), breast cancer (n = 23), haematological malignancies (n = 18), or colorectal cancer (n = 11). Various terms were used to refer to the concept of TCIs, with "thresholds for clinical importance" being the most frequently used term (n = 63; 50.8%). Strikingly, 41 of the 117 articles (35.0%) reported that TCIs were applied to group-level data (e.g. mean scores), which is a clearly unintended application of the TCIs.

CONCLUSION

TCIs are frequently used by the research community and thus enhanced the interpretability of PRO data in oncology. While most studies correctly applied TCIs in their analysis and interpretation, further guidance and clarification on their use are required. This article aims to contribute to this endeavour.