PURPOSE

Patient-reported outcomes (PROs) provide valuable information on adverse effects of cancer treatment. The aim of this study was to describe the feasibility of a population-based PROs collection and identify factors affecting participation.

METHODS

A total of 13 595 patients diagnosed with prostate cancer (PCa) 2017-2019, along with 10 653 age- and region-matched men with no history of PCa (controls), were invited to a health survey collecting PROs shortly after diagnosis and again one and three years after the initial invitation. Invitations were sent by digital or regular mail. The survey included Norwegian translations of the EQ-5D-5L, EORTC QLQ-C30, and EPIC-26 instruments. We assessed participation rates in all three survey rounds.

RESULTS

We invited 90% of newly diagnosed PCa patients to round 1 of the survey. Participation was higher among patients (58%) than controls (34%), with the highest response rate among digitally invited patients aged 60-69 (66%). Among patients, 44% participated in both rounds 1 and 2. Almost one third (32%) of the invited patients participated in all three rounds. Factors associated with higher participation included digital invitation, higher education, lower age, and being treated with curative intent. Among participants treated with curative intent, 51% completed the first questionnaire before starting local treatment.

CONCLUSIONS

National registry-based PROs data collection is feasible for both PCa patients and controls. However, both non-responder and attrition biases are likely to influence questionnaire results. Institutions conducting population-based health surveys should provide information on non-responders and attrition to aid interpretation of questionnaire data. The Cancer Registry of Norway invited all Norwegian prostate cancer patients diagnosed between 2017 and 2019 to participate in a three-year health and quality of life survey. This study examines the success of the set-up of this nationwide survey. Patients received the first questionnaire at least six weeks after diagnosis, followed by the second and third questionnaires one and three years later. A control group consisting of men with no history of prostate cancer was also invited. We sent invitations as digital or regular (paper) mail. We used widely recognised questionnaires to assess health and quality of life in prostate cancer patients. Of the 13 595 patients invited, 58% participated in the first round. Among the 10 653 controls invited, 34% participated. The combined participation rate for patients in the first two rounds was 44%, and 32% for all three rounds. We found that digital invitation, higher education, lower age, and curative treatment were associated with higher response rates. We have shown that it is possible to conduct a registry-based national health survey over time among patients and controls. However, questionnaire results may not be representative due to non-responders and responders being different regarding, e.g., education, age and treatment. Therefore, those conducting population-based health surveys should provide information on non-responders and attrition to aid interpretation of questionnaire data.