Localized prostate cancer patients' preferences for hypofractionated radiotherapy: A discrete choice experiment

Abstract

Introduction: While external beam radiotherapy (EBRT) has always been a standard of care for localized prostate cancer patients, hypofractionated EBRT has now become a standard EBRT approach. We aimed to elicit the relative weight patients place on various aspects of EBRT and to compare these findings to clinical trials results, with a goal of informing clinical practice.

Methods: A discrete choice experiment was designed in which men had to choose between two EBRT profiles that differed in five treatment attributes: effectiveness, acute toxicity, late toxicity, length, and fiducial marker implantation. We recruited prostate cancer patients with localized disease who had completed or were undergoing EBRT at our institution. Conditional (multinomial) logit regression (MNL), mixed multinomial logit (MXL), and latent class (LC) models were used to analyze the relative importance of treatment attributes, and the trade-offs this patient population were willing to make between attributes. These results were compared to explicit ranking and rating scores the participants completed for internal validation, as well as to clinical trials findings for interpretation. The MXL and LC models were used to compare realistic EBRT scenarios that corresponded to conventional fractionation, moderate hypofractionation, and stereotactic body radiotherapy, and determine what characteristics resulted in individuals preferring certain scenarios.

Results: All treatment attributes’ parameter estimates were significantly different from zero, and across the MNL and MXL models the most important attributes were risks of PSA recurrence and late toxicity. The rating exercises agreed with the quantitative results. Preference heterogeneity was present, largely due to age and distance patients traveled to the cancer centre. One latent class (the “convenience group”) were not willing to lengthen EBRT to reduce risk of toxicity at any level within the model.

Conclusions: Prostate cancer patients with localized disease place importance on convenience of EBRT. Older participants tend to be willing to accept a risk of increased toxicity to be treated with hypofractionated regimens. This risk is well within the expected risk based on clinical trial results. Participants had a good understanding of the DCE, and the responses were internally valid.