OBJECTIVES Patients diagnosed with prostate cancer (PCa) have similar survival rates across treatments, making treatments choices based on health related quality of life (HRQoL) more important. Objective was to increase HRQoL usage, sustainability and reliability measurement over time for use in cost–effectiveness analyses by investigating the occurrence of response shift  (RS) in PCa patients. Never before has RS been measured over this long a time in PCa patients, not limited by few available treatment options. METHODS A prospective cross sectional cohort study was started in January of 2012 with 1,720 CAPSURE patients using the SF-36 and UCLA-Prostate Cancer Index (PCI). In January of 2012 patient were asked to fill out questionnaires asking about their perceived HRQoL at time of diagnosis (then score) and current HRQoL (post-score). These scores were matched to previous collected scores at baseline (pre-score), ranging from 3 months to 20 years. RS (then-pre score ), True Change (TC=post-then score) and Felt Change (FC=post-then) were calculated and compared with t-tests in different questionnaire domains. Linear regression was used to explore relations between scores and patient characteristics. RESULTS RS and FC are found to be negative overall for SF-36 (RS:-3.6/-14.5 TC:-8.1/1.8 FC:-3.0/-22.7) and PCI (RS:-4.0/-30.2 TC:-20.2/0.07 FC:-3.6/-49.8), showing significant difference (P<0.05). Significant difference was also seen over the whole time range between post-scores and RS-adjusted-post-scores for SF-36 and PCI. A difference was found between recurrence and non-recurrence in both questionnaires (ranging 0/-4.0), although not significant. RS over time did not show change. CONCLUSIONS Mean negative RS  scores were found in this population using SF-36 and PCI, indicating over reporting in retrospective collected data. It is recommended physicians and researchers adjust found HRQoL scores with RS values found in this study, to increase usage, sustainability and reliability of retrospective collected HRQoL-scores. Further research is needed to investigate RS dependency on other variables or characteristics.