Investigating the Equivalence of the EORTC QLQ-C30 and the QLQ-F17

The EORTC QLQ-C30, a patient-reported outcome measure (PROM) that is available in 110 different languages, has been used in thousands of clinical cancer trials worldwide. The QLQ-C30 is composed of six functioning scales (including a measure for global quality of life), three symptom scales, and six single items.

Researchers and regulatory bodies came up with the idea to construct a shortened EORTC questionnaire that solely consists of functioning scales and to use additional symptom items according to the side effect profile of the specific medication under investigation.

This shortened form termed QLQ-F17 includes the Physical Functioning (PF), Role Functioning (RF), Emotional Functioning (EF), Cognitive Functioning (CF), and Social Functioning (SF) scales as well as the Global Health Status/Quality of Life (QL) scale in their original wording. This functioning questionnaire can be amended with symptom-specific items taken from the EORTC item library. This method provides a flexible and economic testing strategy that fits the demands of regulators and users in industry and academia.

It is an open question, however, whether the QLQ-F17 is equivalent to the QLQ-C30 in terms of measurement properties. Based on empirical research on response biases and order effects, one might argue that elimination of the symptoms between RF and CF-Dyspnea (DY), Pain (PA), Fatigue (FA), Insomnia (SL), Appetite (AP), Nausea/Vomiting (NV), Constipation (CO), and Diarrhea (DI)-and of the Financial difficulties (FI) item between SF and QL may alter the manner in which subsequent items are completed.

Thus, from a methodological point of view, it is essential to confirm the psychometric properties of the QLQ-F17 and to present evidence that scale values derived from the QLQ-F17 are equivalent to those of the QLQ-C30. Only in the case of equivalence may studies using either of the two basic questionnaires be compared directly. The present project is designed to address this research question.

This is an international multicenter survey study that will include respondents with cancer from Australia, Finland, France, Germany, Italy, Poland, Romania, Spain, Sweden, and United Kingdom. A randomized cross-over design will be applied, enabling between-patients as well as within-patients comparisons of the QLQ-C30 and the QLQ-F17. One group of respondents will first fill in the QLQ-C30 followed by the QLQ-F17, the other group will start with the QLQ-F17 followed by the QLQ-C30.

A sample size of 1.500 cancer patients is sufficient to get precise estimates with narrow confidence intervals regarding item and scale-level agreement. Thresholds and margins to be used for the analyses in this study will be consented by a statistical advisory group. Reliability and psychometric properties can also be precisely estimated with a sample of this size. The present study is based on the hypothesis that the QLQ-F17 and the QLQ-C30 questionnaires are equivalent.