Discrete Choice Experiment Design Research Articles

Population Preferences for Treatments When Critically Ill: A Discrete Choice Experiment.

Many patients in the intensive care unit are too unwell to participate in shared decision making or have not previously documented their wishes. In these situations, understanding the values of the general population could help doctors provide appropriate guidance to surrogate decision makers. Using a discrete choice experiment design, we conducted an online survey using an Australian panel. Participants were asked about their willingness to accept treatments, faced with a variety of possible outcomes and probabilities (low, moderate or high). The outcomes were across four domains: loss of functional autonomy, pain, cognitive disability and degree of burden on others. Demographic details, prior experience of intensive care unit and current health conditions were also collected. Data were analysed using logistic regression, predicting whether respondents choose to continue active treatment or not. Nine hundred and eighty-four respondents, representative of age and sex completed the web-based survey. With the increasing likelihood of negative post-intensive care unit sequelae, there was a higher probability of the respondent preferring to stop ongoing active treatment, with the largest coefficients being on caring assistance and the need for full-time residential care. Those who identified as very religious, were younger or who had children under 5years of age were more likely to choose to continue active treatment. Respondents valued their independence as the most important factor in deciding whether to receive ongoing medical treatments in the intensive care unit. When clinicians are unable to obtain specific patient information, they should consider framing their decision making around the likelihood of the patient achieving functional independence rather than survival.

Generating Optimal Designs for Discrete Choice Experiments in R: The idefix Package

Discrete choice experiments are widely used in a broad area of research fields to capture the preference structure of respondents. The design of such experiments will determine to a large extent the accuracy with which the preference parameters can be estimated. This paper presents a new R package, called idefix, which enables users to generate optimal designs for discrete choice experiments. Besides Bayesian D-efficient designs for the multinomial logit model, the package includes functions to generate Bayesian adaptive designs which can be used to gather data for the mixed logit model. In addition, the package provides the necessary tools to set up actual surveys and collect empirical data. After data collection, idefix can be used to transform the data into the necessary format in order to use existing estimation software in R.

Understanding incentive preferences of community health workers using discrete choice experiments: a multicountry protocol for Kenya, Uganda, Bangladesh and Haiti

IntroductionThere is a renewed global interest in improving community health worker (CHW) programmes. For CHW programmes to be effective, key intervention design factors which contribute to the performance of CHWs...

Consumers Demand for Social Farming Products: An Analysis with Discrete Choice Experiments

This paper analyses the demand for social farming (SF) products. In particular, we investigate the preferences of consumers who buy their products from large retailers, rather than from solidarity purchasing groups or other niche markets using a sample of 225 consumers. In this regard, a discrete choice experiment (DCE) was carried out to estimate the willingness to pay (WTP) a premium price for the purchase of a common product (i.e., eggs) from farms that employ disabled people. The attributes considered in our DCE design are the employment of disabled people and two additional attributes which may have ethical implications for the choices. The results indicate that consumers are interested in buying SF products, with about 74% of the sample willing to buy the eggs produced by social farms and the average WTP being equal to €1.36 for a pack of six eggs. Moreover, the average WTP for the use of labour of disabled people attribute amounted to €0.69 for a pack of six eggs.

PCV127 PATIENT PREFERENCES REGARDING ORAL ANTICOAGULANTS: RESULTS OF A DISCRETE CHOICE EXPERIMENT IN SPANISH PATIENTS WITH ATRIAL FIBRILLATION TREATED WITH VITAMIN-K ANTAGONISTS

Oral anticoagulation is pivotal in the management of thromboembolic risk in non-valvular atrial fibrillation (NVAF) patients. An effective anticoagulation therapy also relates on good medication adherence to achieve good anticoagulation control. One key element for reaching or sustaining a good adherence and persistence is to prescribe treatments that patients prefer. The objective of this study was to elucidate patient preferences regarding attributes that describe different anticoagulation alternatives. A multi-center discrete choice experiment (DCE) was conducted in Spain. Study sites enrolled patients aged 18+ years with NVAF, who initiated a Vitamin-K antagonist (VKA) treatment. The DCE design considered four treatment attributes: need of bridging (yes/no), interactions with food/alcohol (yes/no), need of regular international normalized ratio (INR) assessments/blood controls and subsequent dose adjustment (yes/no) and frequency of intake (once/twice daily). Distance to treating practitioner represented a neutral attribute to express patients’ overall utility in terms of a comprehensible unit. Conditional logit regression model was applied to estimate patients’ utility for each attribute and to evaluate the influence of different levels on the probability of a patient of choosing a specific OAC alternative. In total, 286 patients were included (mean age: 72.9 years, 42.0% female, mean CHA2DS2-VASc: 3.2). Patients preferred an anticoagulation treatment option characterized by “no need for bridging” (β=0.530, p<0.001), “no need for regular INR controls” (β=0.652, p<0.001), “less interactions with food/alcohol” (β=0.503, p<0.001) and “once daily intake” (β=0.360, p<0.001). To receive such a treatment option compared to a treatment that offers none of the favourable characteristics, the average patient would be willing to travel additional 57 kilometres to the treating practitioner. In our study, NVAF patients currently treated with VKA showed a strong preference for features that characterize non-VKA anticoagulation options.

On the Optimization of Bayesian D-Efficient Discrete Choice Experiment Designs for the Estimation of QALY Tariffs That Are Corrected for Nonlinear Time Preferences

ObjectivesThis article explains how to optimize Bayesian D-efficient discrete choice experiment (DCE) designs for the estimation of quality-adjusted life year (QALY) tariffs that are unconfounded by respondents' time preferences. MethodsThe calculation of Bayesian D-errors is explained for DCE designs that allow for the disentanglement of respondents' time and health-state preferences. Time preferences are modelled via an exponential, hyperbolic, or power discount function and the performance of the proposed DCE designs is compared with that of several conventional DCE designs that do not take nonlinear time preferences into account. ResultsBased on the achieved D-error, asymptotic standard error, and estimated sample size to obtain statistically significant estimates of the discount rate parameters, the proposed designs outperform the conventional DCE designs. ConclusionsWe recommend that applied researchers use appropriately optimized DCE designs for the estimation of QALY tariffs that are corrected for time preferences. The TPC-QALY software package that accompanies this article makes the recommended designs easily accessible for health-state valuation researchers.

Fractional factorial and D-optimal design for discrete choice experiments (DCE)

Discrete choice experiments (DCE) are a method to measure people’s preferences in various fields of application that used to estimate the importance of attributes of the product or service based on the respondent’s choice. The design of DCE consists of several choices called profiles in each choice set. The best DCE’s design depends on how to divide the profiles into each choice set so it needs a strategy to construct the appropriate DCE’s design. So, the purpose of our study is looking for the most appropriate strategy from two strategies used namely fractional factorial design as the first strategy and D-optimal design with a point-exchange algorithm as the second strategy. The quality of the two strategies is evaluated with the D-optimality criterion value. Both of them are simulated in the case that involves four attributes with two levels, respectively, so all possible profiles are 24 = 16 profiles. The first strategy gained DCE’s design that consisted of two choice sets and eight profiles, respectively. The second strategy gained two different DCE’s designs. The first design consisted of four choice sets and four profiles, respectively. The second design consisted of eight choice sets and two profiles, respectively. Both the first and second strategies have the same D-optimality criterion values but a difference in structure. In terms of statistics, both strategies have the same quality. However, in terms of application, the second strategy is more appropriate in the structure than the first strategy. Hence, the D-optimal design is an alternative to construct DCE’s design based on applications.

Are Healthcare Choices Predictable? The Impact of Discrete Choice Experiment Designs and Models

BackgroundLack of evidence about the external validity of discrete choice experiments (DCEs) is one of the barriers that inhibit greater use of DCEs in healthcare decision making. ObjectivesTo determine whether the number of alternatives in a DCE choice task should reflect the actual decision context, and how complex the choice model needs to be to be able to predict real-world healthcare choices. MethodsSix DCEs were used, which varied in (1) medical condition (involving choices for influenza vaccination or colorectal cancer screening) and (2) the number of alternatives per choice task. For each medical condition, 1200 respondents were randomized to one of the DCE formats. The data were analyzed in a systematic way using random-utility-maximization choice processes. ResultsIrrespective of the number of alternatives per choice task, the choice for influenza vaccination and colorectal cancer screening was correctly predicted by DCE at an aggregate level, if scale and preference heterogeneity were taken into account. At an individual level, 3 alternatives per choice task and the use of a heteroskedastic error component model plus observed preference heterogeneity seemed to be most promising (correctly predicting >93% of choices). ConclusionsOur study shows that DCEs are able to predict choices—mimicking real-world decisions—if at least scale and preference heterogeneity are taken into account. Patient characteristics (eg, numeracy, decision-making style, and general attitude for and experience with the health intervention) seem to play a crucial role. Further research is needed to determine whether this result remains in other contexts.

Accounting for model uncertainty in individualized designs for discrete choice experiments

We develop a method for finding optimal designs in discrete choice experiments (DCEs). More specifically, we deal with individualized designs, which are sequentially generated for each person undertaking a survey, using, at each stage, responses from previous choice sets in order to select the next best set. Currently, the optimal design for a DCE is predicated on a specific mixed logit model that represents the probability of choosing a specific alternative in a choice set through a linear predictor combining several attributes or factors deemed to be relevant in each choice. Using a unique model represents a major limitation, which we overcome by allowing for a collection of different models characterized by distinct linear predictors. Our approach is fully Bayesian and incorporates a prior distribution on the model space as well as a prior on the parameter space specific to each respondent. To perform our optimal design strategy for DCEs, we specify two distinct utility functions, one being targeted to model discrimination while the other aims at the dual purpose of model discrimination and parameter estimation. We implement our methodology using a sequential Monte Carlo algorithm suitably tailored to account for model uncertainty and individual-specific parameters. To evaluate our methodology, several simulation settings are considered in detail. Comparisons with alternative methods are also investigated.

Treatment decision-making in chronic lymphocytic leukaemia: Key factors for healthcare professionals. PRELIC study

ObjectiveTo explore the preferences of Spanish healthcare professionals (haematologists and hospital pharmacists) for the treatment selection of active Chronic Lymphocytic Leukaemia (CLL) patients at first relapse, condition that mainly afflicts older adults. MethodsA discrete choice experiment (DCE) was conducted among haematologists and hospital pharmacists. A literature review and a focus group informed the DCE design. CLL treatment settings were defined by seven attributes: four patient/disease-related attributes (age, functional status, comorbidities, and risk of the disease) and three treatment-related attributes (efficacy [hazard ratio of progression-free survival, HR-PFS], rate of discontinuations due to adverse events and cost). A mixed-logit model was used to determine choice-based preferences. Relative importance (RI) of attributes was calculated and compared between stakeholders. Willingness-to-pay (WTP) was estimated through the DCE. Besides, nine ad-hoc questions were posed, to explore more in depth CLL treatment decision making. ResultsA total of 130 participants (72 haematologists and 58 hospital pharmacists) answered the DCE. All attributes were significant predictors of preferences (p < 0.05) in the multinomial model. Higher RI was obtained for treatment-related attributes: the highest rated being ‘cost’ (23.8%) followed by ‘efficacy’ (20.9%). Regarding patient-related attributes, the highest RI was obtained for ‘age’ (18.1%). No significant differences (p > 0.05) in RI between haematologists and pharmacists were found. WTP for the treatment was higher for younger CLL patients. Ad-hoc questions showed that patient age and functional status influence treatment decisions. ConclusionsFor healthcare professionals, ‘cost’ and ‘efficacy’ (treatment-related attributes) and age (patient-related attribute) are the main factors that determine CLL treatment selection at first relapse. WTP decreases as patient's age increases.

Adoption and Cooperation Decisions in Sustainable Energy Infrastructure: Evidence From a Sequential Choice Experiment in Germany

In this paper, we propose and apply the design of a sequential discrete choice experiment to examine homeowner preferences regarding the adoption of micro-generation systems and willingness to cooperate in sustainable energy infrastructure. Adoption and cooperation decisions of private households in the energy sector are complex, interlinked, and assumably sequential. A common design with single choice tasks reflecting both adoption and cooperation decisions is assumed as cognitively too burdensome for survey respondents. The objective of the proposed sequential choice task design is twofold. Firstly, reducing complexity for respondents. Secondly, reflecting a step-wise decision process as is appropriate for the studied decisions. Our application from the energy sector is motivated by the need for innovative business models for non-industrial prosumers providing flexibility services in (local) distribution grids, due to an increasing amount of volatile and decentrally generated electricity. Results indicate that respondents reveal more pronounced preferences when dealing with their decision in sequential steps and that the task design has a lasting effect on respondents’ choices. By estimating latent class logit models, five consumer classes are identified and labeled by their distinguished motivational foci: costs (1), climate protection (2), self-supply (3), local reference (4), and other (5).

Using a stated preference discrete choice experiment to assess societal value from the perspective of decision-makers in Europe. Does it work for rare diseases?

Objectives: To pilot the feasibility of using a discrete choice experiment (DCE) design to investigate individual preferences from the decision-maker perspective regarding the use of public funding for orphan drugs and generate prior information for future experimental designs.Methods: A DCE was used on a convenience sample of participants from five European countries (England, France, Germany, Italy and Spain), exploring their preferences in distinct healthcare scenarios involving orphan drugs. A preliminary review of the empirical literature on distributive preferences informed the selection of attributes and their levels in the design. An online questionnaire was used to conduct the DCE survey.Results: A total of 199 questionnaires were completed. The five country model showed relative preference for some attributes over others: cost of treatment, improvement in health, value for money and availability of treatment alternatives received the greatest attention. However, disease severity, beginning of life, waiting times and side effects were also shown to be important social values that should not be ignored.Conclusions: The findings presented in this study provide insight about the preferences that can influence decisions on orphan drugs in different countries. This study also provides valuable prior information that could inform future DCE designs in this area.

Treatment of relapsed refractory multiple myeloma: which new PI-based combination treatments do patients prefer?

Background and objectivesThis study describes preferences of German relapsed refractory multiple myeloma (RRMM) patients with novel proteasome inhibitor-based combination treatments.MethodsPatients with a minimum age of 18 years and a diagnosis of RRMM were included. Their preferences were assessed using a discrete choice experiment design, which was developed based on a literature review and two patient focus group discussions. The final discrete choice experiment design consisted of four attributes, namely “therapy application regimen,” “time without progression of disease,” “possibility of grade ≥3 adverse events (AEs) affecting the blood,” and “possibility of grade ≥3 AE heart failure.”ResultsAnalysis was based on 84 patients (36.9% females, mean age 62.7 years, mean multiple myeloma disease duration 5.5 years). Among the tested attributes, “therapy application regimen” was assigned the highest importance for treatment decisions (38.8%), the second important attribute was “time without progression of disease” (38.7%), followed by “possibility of AE heart failure” (13.9%) and “possibility of AEs affecting the blood” (8.6%). Patients preferred oral intake once a day and once a week over other application modes such as oral intake once a day and once a week plus twice-weekly infusions. Furthermore, they preferred longer disease progression-free time and lower risk of grade ≥3 AEs. The highest overall utility was derived for ixazomib + lenalidomide + dexamethasone (utility: 3.218), compared with lenalidomide + dexamethasone (2.769), and carfilzomib + lenalidomide + dexamethasone (1.928).ConclusionRRMM patients prefer treatments with an all-oral application, a longer disease-progression-free time, and a lower probability of AEs. If patients face tradeoffs, they accept a lower progression-free time and/or higher AE rates to get an all-oral therapy.

Attitudes toward prenatal screening for chromosomal abnormalities: A focus group study

BackgroundWhile discrete choice experiments (DCEs) are well established methods to ascertain patient preferences, there is limited literature describing use of qualitative methods in DCE design. AimThis article provides a case study of the qualitative research process for developing the conceptual attributes for a DCE for prenatal screening and diagnosis. MethodsParticipants were recruited through posters and social media. Four in-depth, semi-structured focus groups with pregnant women and their partners/support people were conducted in Metro Vancouver. FindingsOur analysis indicates that choosing prenatal screening and diagnosis involves four intertwined decisions: whether to undergo screening and testing, which screening test to take, which diagnostic test to take, and what to do with a positive diagnosis. The factors that are important to women and their partners vary depending on the decision and include: time of diagnosis, information on conditions tested, false positives, cost, the invasiveness of the test, and potential harm to woman and baby. DiscussionFindings suggest that certain attributes were more salient for screening versus diagnostic tests. Preferences were often shaped by a woman’s perceived ability to care for a child with a genetic anomaly, personal risk factors, parity, views on termination, and perceptions on public or private coverage. Participants valued mental well-being and demonstrated a willingness to trade-off on certain attributes in order to minimize stress or anxiety during pregnancy. ConclusionStudy findings will be used to inform DCE attributes, levels, and choice questions. Findings will be important for policy decisions surrounding prenatal testing.

Severity-Stratified Discrete Choice Experiment Designs for Health State Evaluations

BackgroundDiscrete choice experiments (DCEs) are increasingly used for health state valuations. However, the values derived from initial DCE studies vary widely. We hypothesize that these findings indicate the presence of unknown sources of bias that must be recognized and minimized. Against this background, we studied whether values derived from a DCE are sensitive to how well the DCE design spans the severity range.MethodsWe constructed an experiment involving three variants of DCE tasks for health state valuation: standard DCE, DCE-death, and DCE-duration. For each type of DCE, an experimental design was generated under two different conditions, enabling a comparison of health state values derived from current best practice Bayesian efficient DCE designs with values derived from ‘severity-stratified’ designs that control for coverage of the severity range in health state selection. About 3000 respondents participated in the study and were randomly assigned to one of the six study arms.ResultsImposing the severity-stratified restriction had a large effect on health states sampled for the DCE-duration approach. The unstratified efficient design returned a skewed distribution of selected health states, and this introduced bias. The choice probability of bad health states was underestimated, and time trade-offs to avoid bad states were overestimated, resulting in too low values. Imposing the same restriction had limited effect in the DCE-death approach and standard DCE.ConclusionVariation in DCE-derived values can be partially explained by differences in how well selected health states spanned the severity range. Imposing a ‘severity stratification’ on DCE-duration designs is a validity requirement.Electronic supplementary materialThe online version of this article (10.1007/s40273-018-0694-6) contains supplementary material, which is available to authorized users.

Numeracy and Understanding of Quantitative Aspects of Predictive Models: A Pilot Study.

The assessment of user preferences for performance characteristics of patient-oriented clinical prediction models is lacking. It is unknown if complex statistical aspects of prediction models are readily understandable by a general audience. A pilot study was conducted among nonclinical audiences to determine the feasibility of interpreting statistical concepts that describe the performance of prediction models. We conducted a cross-sectional electronic survey using the Amazon Mechanical Turk platform. The survey instrument included educational modules about predictive models, sensitivity, specificity, and confidence intervals (CIs). Follow-up questions tested participants' abilities to interpret these characteristics with both verbatim and gist knowledge. Objective and subjective numeracy were assessed using previously validated instruments. We also tested understanding of these concepts when embedded in a sample discrete choice experiment task to establish feasibility for future elicitation of preferences using a discrete choice experiment design. Multivariable linear regression was used to identify factors associated with correct interpretation of statistical concepts. Among 534 respondents who answered all nine questions, the mean correct responses was 95.9% (95% CI, 93.8-97.4) for sensitivity, 93.1% (95% CI, 90.5-95.0) for specificity, and 86.6% (95% CI, 83.3-89.3) for CIs. Verbatim interpretation was high for all concepts, but significantly higher than gist only for CIs (p < 0.001). Scores on each discrete choice experiment tasks were slightly lower in each category. Both objective and subjective numeracy were positively associated with an increased proportion of correct responses (p < 0.001). These results suggest that a nonclinical audience can interpret quantitative performance measures of predictive models with very high accuracy. Future development of patient-facing clinical prediction models can feasibly incorporate patient preferences for model features into their development.

Substitution Effects and Spatial Preference Heterogeneity in Single- and Multiple-Site Choice Experiments

A novel discrete choice experiment (DCE) design allows testing of substitution effects based on the framing of, and distances to, substitute sites. The same ecosystem services and biodiversity provided at different sites are valued independently in single- site DCEs and simultaneously in a multiple- site DCE, using an identical experimental design. Site-specific utility functions are estimated with and without controlling for spatial preference heterogeneity. Framing choices as alternative projects at single or multiple sites significantly influences substitutability between sites, choice behavior, and welfare estimates. Distance decay displays significant heterogeneity in different directions from the valued sites and between urban and rural areas.

Advocating a Paradigm Shift in Health-State Valuations: The Estimation of Time-Preference Corrected QALY Tariffs

BackgroundDespite evidence of nonproportional trade-offs in time trade-off exercises and the explicit incorporation of exponential discounting in health technology assessment calculations, quality-adjusted life-year (QALY) tariffs are currently still established under the assumption of linear time preferences. ObjectivesThe aim of this study was to introduce a general method of accommodating for nonlinear time preferences in discrete choice experiment (DCE) duration studies and to evaluate its impact on estimated QALY tariffs. MethodsA parsimonious utility function is proposed that accommodates any discounting function and preserves linear time preferences as a special case. Based on an efficient DCE design and 1775 respondents from a nationally representative scientific household panel, preferences and QALY tariffs for the Dutch SF-6D were estimated while accommodating for nonlinear time preferences via exponential and hyperbolic discounting functions. ResultsWhen the discount rate was estimated directly, we found strong evidence of nonlinear time preferences (with an exponential and hyperbolic discount rate of 5.7% and 16.5%, respectively). When the discount rate was estimated as a function of health state severity, we found that years lived in better health states are discounted minus years lived in impaired health states. Finally, the best statistical fit was obtained when using a hyperbolic discount function, which resulted in smaller QALY decrements and fewer health states classified as worse than immediate death. ConclusionsOur results highlight the relevance and even necessity of a paradigm shift in health valuation studies in favor of time-preference corrected QALY tariffs, with potentially important implications for health technology assessment calculations and regulatory decisions.

Use of discrete choice experiments to facilitate design of effective environmentally friendly agricultural policies

Appropriate environmentally friendly agricultural technologies and practices must be implemented in China (and many other countries) to reduce farmland non-point source pollution and meet sustainability objectives. A major hindrance to their implementation is that Chinese farmers have very low willingness to accept such practices. Discrete choice experiments (DCEs) have known value for identifying goods or policies with attractive attributes, and thus exploring target groups’ choice behavior and addressing such problems. Sound questionnaire design is of paramount importance for robust results. This article describes the nature and theoretical foundations of DCEs and considers their use in diverse applications abroad, and their limited but growing applications in China. It then presents a case study, focused on the detailed design of DCEs intended to gauge paddy farmers’ willingness to accept several agriculturally friendly techniques and related compensation methods, problems encountered, and predictions for future studies.

Workshop Synthesis: New directions in experimental design

Abstract This paper summarizes the discussion of the workshop “new directions in experimental design”, in which the evolution of experimental designs was discussed in the context of the rapid changes and challenges that the transportation field is experiencing. Because an experimental setting in travel behavior analysis is associated with stated preference methods, the workshop had a clear focus on the design of discrete choice experiments. However, one of the main conclusions is that we need an expanded notion of experiments to achieve diversity in the participation of experimentalists beyond standard stated preference analysts. In this broader, more holistic definition of experiment, workshop participants made clear that good designs need to explicitly address treatments and control, as well as the identification of a good, appropriate baseline for valid hypothesis testing. In fact, the group also discussed the idea of “experiment design of the experiment design” meaning that we should control in a methodical way the parameters of the design so that analysts can assess and formalize design protocols, while also examining the combined use of conventional and new tools.