A pilot discrete choice experiment for understanding breast cancer survivors' preferences for post-operative physiotherapy care

As many breast cancer survivors (BCS) in British Columbia report difficulties accessing physiotherapy (PT), which increases their risk of developing chronic shoulder and arm issues, strategies are needed to inform effective PT program design. This thesis project used sequential mixed methods to develop a pilot discrete choice experiment (DCE) to understand the feasibility of using DCE methodology to elicit BCSs’ preferences for PT services. Methods: First, focus groups were held with BCS (n=35) to identify key attributes and levels of PT care: timing of education, method of PT referral, timing of first PT appointment, physiotherapist expertise level, treatment frequency, treatment format, and out-of-pocket cost. Second, think aloud interviews were conducted with BCS (n=5) to explore attribute and level relevance and task comprehension. Third, an experimental design of 32 choice tasks (separated into two blocks) was created from the identified attributes and levels, quantitatively pilot tested with BCS, and fit using an error-component mixed logit model to estimate preference data. Feasibility was assessed based on: (i) response rate; (ii) completion rate; and (iii) respondent diversity. Results: The pilot DCE had a 21% response rate and 17% of respondents completed the entire questionnaire and provided usable data. All respondents identified as female and were mean of 9.1 years post-diagnosis. Most received PT treatment for breast cancer-related issues (69%), completed post-secondary education (77%), and had annual family incomes of at least $80,000 (69%). Preliminary preference data indicate, on average, respondents preferred breast surgeon referral to PT, less time between surgery completion and receipt of first PT appointment, a physiotherapist with greater breast cancer-specific expertise, and to pay less out-of-pocket for PT treatment. Conclusion: Given the low response and completion rates and respondent homogeneity, alternative recruitment strategies and choice task instruction and presentation should be explored before beginning a full-scale DCE. As this DCE was a pilot study, preference data should be interpreted with caution and not inferred to reflect the preferences of the wider population of BCS. Preference data from the full-scale DCE may be incorporated into exploratory policy analyses and programming for breast cancer-specific PT services in British Columbia.Medicine, Faculty ofGraduat