Introduction:

Through collaborative shared decision-making, we expect that patients with prostate cancer will make treatment-related choices that are most consistent with their values. However, prior evidence suggests that actual treatment choice is more influenced by the type of clinician a patient sees (radiation oncologist vs. urologist) rather than a patient’s originally stated preference. Additionally, traditionally underserved patients often experience more trouble communicating their treatment preferences with clinicians during visits. To mitigate this risk, clinicians may benefit from greater insight in two key areas: (1) which aspects of prostate cancer treatment matter most to patients, and (2) how these preferences vary across demographic groups. We developed a prospective, patient-centered discrete choice experiment to better address these gaps in understanding. Discrete choice experimentation is a quantitative technique used commonly in economics and marketing research to assess how respondents assign relative importance to various attributes of a product or treatment.

Methods:

We recruited patients with Gleason Grade Group 2 or 3 prostate cancer from a single-center academic urology clinic. All patients had biopsy-confirmed prostate cancer but had yet to begin treatment. We designed the discrete choice experiment to include eight attributes with 3-4 levels each to describe the travel burden and treatment harms associated with intermediate-risk prostate cancer. Each survey included 12 choice tasks, and each choice task consisted of two theoretical alternatives for each attribute. For the primary analysis, we estimated patient preferences for alternative treatment options using a conditional logit model. We then conducted a latent class analysis to identify unobserved subgroups within the study population characterized by distinct patterns in choice behaviors. We used a multivariable linear probability model to assess the association between patient demographic characteristics and latent class membership.

Results:

A total of 87 patients completed the survey. In our primary analysis, patients had a significant preference towards avoiding bowel symptoms, erectile dysfunction, use of a urinary catheter, or frequent trips to the facility (Table 1). Bowel symptoms were considered the most important of all attributes (very frequent loose bowel movements: odds ratio (OR)= 0.29; 95% confidence interval (CI), 0.21-0.39, compared to no symptoms), while urinary symptoms did not significantly influence patients’ choices (leaking urine: OR=1.00; 95% CI, 0.80-1.24, compared to no symptoms). Patients placed a low weight on facing prolonged travel times or undergoing a major surgery with overnight hospitalization. LCA revealed 2 groups based on survey response patterns: (1) patients averse to surgery and catheter and (2) patients averse to erectile and bowel symptoms. Older patients and racial and ethnic minorities were more likely to be members of the surgery and catheter-averse class (Table 2).

Conclusion:

We found that patients generally prioritized avoiding bowel symptoms over avoiding urinary symptoms when choosing between treatment options. We also found that travel time and treatment duration did not significantly influence patients’ treatment choices. Finally, a subset of older patients placed the greatest emphasis on avoiding surgery and a urinary catheter. Ultimately, these results can be used to inform shared decision-making conversations regarding treatment choice for surgery versus radiation. 

Funding: N/A