Comparison of Rating Scale, Time Tradeoff, and Conjoint Analysis Methods for Assessment of Preferences in Prostate Cancer.

Background. Conjoint analysis is widely used in studies of consumer preference but has only recently been applied to measure patient utilities for health outcomes. We compared the reliability, feasibility, and internal and predictive validity of conjoint scaling methods against better established rating scale and time tradeoff methods for assessing prostate cancer utilities in men at risk for prostate cancer. Methods. In total, 194 men who were biopsy negative for prostate cancer were randomly assigned to complete 2 preference assessment modules, either conjoint analysis and a rating scale module or conjoint analysis and a time tradeoff module. Each participant's most important attribute was identified and evaluated in relation to age group (age <65, age 65 and older), education (high school, some college, college graduate), race/ethnicity (white, black, Latino), and relationship status (in significant relationship v. not). The methods were also evaluated in terms of ease of use and satisfaction. Results. Rating scales were rated as easiest to use and respondents were more satisfied with rating scales and conjoint in comparison to time tradeoffs. Rating scales and conjoint measures demonstrated significantly higher internal validity compared to time tradeoff when evaluated through R2 of the fitted utility function. The 3 methods were similar in terms of predictive validity, but conjoint analysis outperformed the rating scale method when patients were presented with novel combinations of attribute levels (68% correct v. 43%, P = 0.003). Conclusions. Rating scales and conjoint analysis exercises offer greater ease of use and higher satisfaction when measuring patient preferences in men biopsied for prostate cancer in comparison to time tradeoff exercises. Conjoint analysis may be a more robust approach to preference measurement for men at risk for prostate cancer.

A Bayesian hierarchical model for discrete choice data in health care.

In discrete choice experiments, patients are presented with sets of health states described by various attributes and asked to make choices from among them. Discrete choice experiments allow health care researchers to study the preferences of individual patients by eliciting trade-offs between different aspects of health-related quality of life. However, many discrete choice experiments yield data with incomplete ranking information and sparsity due to the limited number of choice sets presented to each patient, making it challenging to estimate patient preferences. Moreover, methods to identify outliers in discrete choice data are lacking. We develop a Bayesian hierarchical random effects rank-ordered multinomial logit model for discrete choice data. Missing ranks are accounted for by marginalizing over all possible permutations of unranked alternatives to estimate individual patient preferences, which are modeled as a function of patient covariates. We provide a Bayesian version of relative attribute importance, and adapt the use of the conditional predictive ordinate to identify outlying choice sets and outlying individuals with unusual preferences compared to the population. The model is applied to data from a study using a discrete choice experiment to estimate individual patient preferences for health states related to prostate cancer treatment.

Does Patient Preference Measurement in Decision Aids Improve Decisional Conflict? A Randomized Trial in Men with Prostate Cancer.

BACKGROUND: Shared decision making (SDM) has been advocated as an approach to medical decision making that can improve decisional quality. Decision aids are tools that facilitate SDM in the context of limited physician time; however, many decision aids do not incorporate preference measurement. OBJECTIVES: We aim to understand whether adding preference measurement to a standard patient educational intervention improves decisional quality and is feasible in a busy clinical setting. METHODS: Men with incident localized prostate cancer (n = 122) were recruited from the Greater Los Angeles Veterans Affairs (VA) Medical Center urology clinic, Olive View UCLA Medical Center, and Harbor UCLA Medical Center from January 2011 to May 2015 and randomized to education with a brochure about prostate cancer treatment or software-based preference assessment in addition to the brochure. Men undergoing preference assessment received a report detailing the relative strength of their preferences for treatment outcomes used in review with their doctor. Participants completed instruments measuring decisional conflict, knowledge, SDM, and patient satisfaction with care before and/or after their cancer consultation. RESULTS: Baseline knowledge scores were low (mean 62%). The baseline mean total score on the Decisional Conflict Scale was 2.3 (±0.9), signifying moderate decisional conflict. Men undergoing preference assessment had a significantly larger decrease in decisional conflict total score (p = 0.023) and the Perceived Effective Decision Making subscale (p = 0.003) post consult compared with those receiving education only. Improvements in satisfaction with care, SDM, and knowledge were similar between groups. CONCLUSIONS: Individual-level preference assessment is feasible in the clinic setting. Patients with prostate cancer who undergo preference assessment are more certain about their treatment decisions and report decreased levels of decisional conflict when making these decisions.

The Voice of the Patient Methodology: A Novel Mixed-Methods Approach to Identifying Treatment Goals for Men with Prostate Cancer.

BACKGROUND: Many guidelines advocate the use of shared decision making for men with newly diagnosed prostate cancer. Decision aids can facilitate the process of shared decision making. Implicit in this approach is the idea that physicians understand which elements of treatment matter to patients. Little formal work exists to guide physicians or developers of decision aids in identifying these attributes. We use a mixed-methods technique adapted from marketing science, the 'Voice of the Patient', to describe and identify treatment elements of value for men with localized prostate cancer. METHODS: We conducted semi-structured interviews with 30 men treated for prostate cancer in the urology clinic of the West Los Angeles Veteran Affairs Medical Center. We used a qualitative analysis to generate themes in patient narratives, and a quantitative approach, agglomerative hierarchical clustering, to identify attributes of treatment that were most relevant to patients making decisions about prostate cancer. RESULTS: We identified five 'traditional' prostate cancer treatment attributes: sexual dysfunction, bowel problems, urinary problems, lifespan, and others' opinions. We further identified two novel treatment attributes: a treatment's ability to validate a sense of proactivity and the need for an incision (separate from risks of surgery). CONCLUSIONS: Application of a successful marketing technique, the 'Voice of the Customer', in a clinical setting elicits non-obvious attributes that highlight unique patient decision-making concerns. Use of this method in the development of decision aids may result in more effective decision support.