Preference Elicitation Research Articles

Research methodologies for eliciting patients' preferences in invasive procedures: a scoping review.

Endpoints that patients and clinicians consider important may differ based on patients' preferences and values. Several methods are available to elicit patient preferences in a succinct and methodologically valid manner. We conducted a scoping review of methods used to elicit patient preferences in invasive procedures to provide a framework for researchers and clinicians to incorporate these measures into future efforts. MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials and health and psychological instruments database were searched from inception until September 2020. Articles that examined patient preferences for any invasive procedure were eligible for inclusion. Selection and extraction were completed in duplicate. Preference elicitation methods were identified and summarized. Three hundred ninety-four articles (n = 76,921 patients) were included representing several surgical specialties. Of included studies, 11.7% (n = 46) used both quantitative and qualitative methods, 81.2% (n = 320) used quantitative methods only, and 7.1% (n = 28) used qualitative methods only to elicit preferences. The most frequently employed quantitative method to elicit preferences was simple choice selection, while one-on-one interviews with participants was the most frequently used qualitative method. Preference elicitation was the primary outcome in 74.6% (n = 294) of included studies. There are several methods to elicit patient preferences in surgical research. Qualitative methods are valuable for exploring views and generating consensus statements. Quantitative methods are better suited for assessing relative preferences, establishing preference thresholds, or ascertaining the presence of preferences. The choice of method should align with the specific research objectives.

Comparison of four approaches in eliciting health state utilities with SF-6Dv2.

To empirically compare four preference elicitation approaches, the discrete choice experiment with time (DCETTO), the Best-Worst Scaling with time (BWSTTO), DCETTO with BWSTTO (DCEBWS), and the Standard Gamble (SG) method, in valuing health states using the SF-6Dv2. A representative sample of the general population in Quebec, Canada, completed 6 SG tasks or 13 DCEBWS (i.e., 10 DCETTO followed by 3 BWSTTO). Choice tasks were designed with the SF-6Dv2. Several models were used to estimate SG data, and the conditional logit model was used for the DCE or BWS data. The performance of SG models was assessed using prediction accuracy (mean absolute error [MAE]), goodness of fit using Bayesian information criterion (BIC),t-test, Jarque-Bera (JB) test, Ljung-Box (LB) test, the logical consistency of the parameters, and significance levels. Comparison between approaches was conducted using acceptability (self-reported difficulty and quality levels in answering, and completion time), consistency (monotonicity of model coefficients), accuracy (standard errors), dimensions coefficient magnitude, correlation between the value sets estimated, and the range of estimated values. The variance scale factor was computed to assess individuals' consistency in their choices for DCE and BWS approaches. Out of 828 people who completed SG and 1208 for DCEBWS tasks, a total of 724 participants for SG and 1153 for DCE tasks were included for analysis. Although no significant difference was observed in self-reported difficulties and qualities in answers among approaches, the SG had the longest completion time and excluded participants in SG were more prone to report difficulties in answering. The range of standard errors of the SG was the narrowest (0.012 to 0.015), followed by BWSTTO (0.023 to 0.035), DCEBWS (0.028 to 0.050), and DCETTO (0.028 to 0.052). The highest number of insignificant and illogical parameters was for BWSTTO. Pain dimension was the most important across dimensions in all approaches. The correlation between SG and DCEBWS utility values was the strongest (0.928), followed by the SG and BWSTTO values (0.889), and the SG and DCETTO (0.849). The range of utility values generated by SG tended to be shorter (-0.143 to 1) than those generated by the other three methods, whereas BWSTTO (-0.505 to 1) range values were shorter than DCETTO (-1.063 to 1) and DCEBWS (-0.637 to 1). The variance scale factor suggests that respondents had almost similar level of certainty or confidence in both DCE and BWS responses. The SG had the narrowest value set, the lowest completion rates, the longest completion time, the best prediction accuracy, and produced an unexpected sign for one level. The BWSTTO had a narrower value set, lower completion time, higher parameter inconsistency, and higher insignificant levels compared to DCETTO and DCEBWS. The results of DCEBWS were more similar to SG in number of insignificant and illogical parameters, and correlation.

Virtual Service Delivery in Mental Health and Substance Use Care: A Systematic Review of Preference Elicitation Studies.

Mental health and substance use disorders affect the lives of many people worldwide. Prevention and treatment of these conditions is important for optimal health and wellbeing, yet service access barriers are common. Virtual models of care may help to reduce barriers to receiving care. However, to facilitate uptake and use of virtual services, they need to appeal to patients and clinicians. This systematic review aimed to synthesise preference elicitation studies to determine what features of virtual mental health and substance use care are preferred by service users and service providers. Following the PRISMA guidelines for systematic reviews, we searched PubMed, PsycINFO, EconLit, MEDLINE, CINAHL, Academic Search Ultimate, and ProQuest Central for all available studies from database inception until May 2023. The Mixed Methods Appraisal Tool was used to assess the methodological quality of included studies. Nineteen studies met the eligibility criteria. However, none examined preferences for elements of different models of virtual care. Across the included studies, we identified 41 unique features that mapped to four themes of mental health and substance use care ('service', 'treatment', 'clinician' and 'additional supports'). Participant preferences were for individual, in-person, effective, flexible, and low-cost treatment. These preferences varied based on demographic factors, such as culture, gender, and participant type (e.g., patients, clinicians, general population). A user-centred approach should be adopted when designing and implementing mental health and substance use services. While preferences for features of mental health and substance use services more broadly are known, preferences for different models of virtual care remain unexplored. Future research should examine what features of virtual services would lead to optimal uptake and use across different users and stakeholders.

Abstract P332: High Shared Decision-Making Scores Associated with Blood Pressure Control: The Office Guidelines Applied to Practice (Office-GAP) IMPACT Program

Introduction: Cardiovascular disease (CVD) mortality is higher in people with diabetes mellitus due to the increased prevalence of CVD risk factors. The Office-Gap IMPACT Program (Office-Guidelines Applied to Practice Improving Diabetic Patients’ Adherence to Treatment and Prevention of Cardiovascular Disease) is a patient activation cluster-randomized trial (intervention: Office-GAP + text vs. control: text alone) seeking to improve communication and patient-provider partnerships through patient and provider training in shared decision-making (SDM), use of a guideline-based checklist, and text messaging to improve medication adherence in Federally Qualified Health Care Centers. Hypothesis: We assessed the hypothesis that individual level demographic and clinical characteristics of patients enrolled in Office-GAP IMPACT would be associated with patient-reported SDM using the CollaboRATE 3-item instrument, which measures three core SDM domains: (1) explanation of health issues, (2) elicitation of patient preferences and (3) integration of patient preferences into decisions with a 10-point response scale for each domain ranging from 0 “No effort was made” to 9 “Every effort was made.” Methods: The CollaboRATE was completed three months after baseline. A top-scoring method (a rating of 9 for all three items) was used to dichotomize CollaboRATE responses into two categories - participants who reported the highest levels of SDM on each of the three items versus those who did not. Accounting for the cluster-randomized design, generalized estimating equations (GEEs) were used to compare demographics and baseline factors that may be associated with the outcome. Results: A total of 288 patients were enrolled (intervention, N=158 and control, N=130). Among those with non-missing CollaboRATE scores, 152 (73% ± 3%) had high SDM. The inverse probability of non-missing weighting method was used to account for the missing data in CollaboRATE. Results from the weighted univariable models suggested older patients, history of smoking, and those with BP under control (< 140/90 mmHg) at baseline were more likely to be satisfied with the SDM process. In multivariable logistic regressions, only BP control was statistically significant after adjusting for other factors. Conclusions: In conclusion, participants with BP control had higher SDM scores, which could lead to better health outcomes. Providers should include the practice of SDM principles in every clinical encounter.

Identifying pregnant and postpartum women’s priorities for enhancing nutrition support through social needs programmes in a resource-constrained urban community in South Africa

BackgroundMalnutrition remains a pressing public health concern for mothers and children in South Africa. Despite the government’s multisectoral response, unaddressed social needs prevent some mothers getting full benefit from interventions, spanning financial planning, income stability, housing, access to government services, social support, and provision of affordable, nutritious foods. Engaging with mothers and prioritising their concerns is important if we wish to overcome obstacles to women benefiting from government nutrition interventions. This study aimed to identify the programmes that women perceived as a priority in addressing the social needs of mothers of young infants and pregnant women to enhance nutrition in a resource-constrained urban township in South Africa.MethodsA cross-sectional study employed a quantitative preference elicitation survey, administered to 210 mothers and pregnant women from five primary healthcare facilities in Soweto. The survey tool was developed with the community to identify unmet social needs and potential solutions, which were synthesised with findings from the literature. The survey described 15 programmes, grouped into three delivery levels: clinics, community, and government. Participants were required to rank programme options in two stages. First, they selected their top two programmes within each delivery level. Subsequently, they allocated stickers to indicate the strength of their preference among the top programmes across the levels. Rankings were analysed using descriptive statistics.ResultsThe highest priority was given to five programmes. Two delivered at the community level: Women’s economic empowerment groups and Job search assistance, two at the clinic level: Social needs assessment and referral, and Prescription-based food, and one at the government level: Free quality childcare. The lowest-ranked programmes were two clinic-based programmes, specifically Maternal nutrition groups and Couple antenatal education.ConclusionWomen expressed strong views about which programmes should be prioritised to support mothers and pregnant women in addressing their social needs and improving nutrition. Key areas included providing support with job searching and entrepreneurship, accessing childcare and the healthy foods recommended at clinics, as well as finding information on available community and government services. Leveraging multisectoral collaboration, aligned policy objectives, efficient public financing, and strengthened implementation capacity will be pivotal in delivering these programmes.

Pay all subjects or pay only some? An experiment on decision-making under risk and ambiguity

We investigate the validity of a double random incentive system where only a subset of subjects is paid for one of their choices. By focusing on individual decision-making under risk and ambiguity, we show that using either a standard random incentive system, where all subjects are paid, or a double random system, where only 10% of subjects are paid, yields similar preference elicitation results. These findings suggest that adopting a double random incentive system could significantly reduce experimental costs and logistic efforts, thereby facilitating the exploration of individual decision-making in larger-scale and higher-stakes experiments.

Personalized Language Model Selection Through Gamified Elicitation of Contrastive Concept Preferences.

Language models are widely used for different Natural Language Processing tasks while suffering from a lack of personalization. Personalization can be achieved by, e.g., fine-tuning the model on training data that is created by the user (e.g., social media posts). Previous work shows that the acquisition of such data can be challenging. Instead of adapting the model's parameters, we thus suggest selecting a model that matches the user's mental model of different thematic concepts in language. In this article, we attempt to capture such individual language understanding of users. In this process, two challenges have to be considered. First, we need to counteract disengagement since the task of communicating one's language understanding typically encompasses repetitive and time-consuming actions. Second, we need to enable users to externalize their mental models in different contexts, considering that language use changes depending on the environment. In this article, we integrate methods of gamification into a visual analytics (VA) workflow to engage users in sharing their knowledge within various contexts. In particular, we contribute the design of a gameful VA playground called Concept Universe. During the four-phased game, the users build personalized concept descriptions by explaining given concept names through representative keywords. Based on their performance, the system reacts with constant visual, verbal, and auditory feedback. We evaluate the system in a user study with six participants, showing that users are engaged and provide more specific input when facing a virtual opponent. We use the generated concepts to make personalized language model suggestions.

The effect of preference elicitation methods on the user experience in conversational recommender systems

The prevalence of conversational interfaces is rapidly rising, since improved algorithms allow for remarkable proficiency in understanding and generating natural language. This also holds for Conversational Recommender Systems (CRS), that benefit from information being provided by the user in the course of the dialogue to offer personalized recommendations. However, the challenge remains eliciting the user's characteristics and preferences in a way that leads to the most optimal user experience. Hence, the current research was aimed at investigating the effect of different Preference Elicitation (PE) methods on the user experience of a CRS. We introduce two axes across which PE methods can be classified, namely the degree of system prompt guidance and the level of user input restriction. We built three versions of a CRS to conduct a between-subjects experiment which compared three conditions: high guidance-high restriction, high guidance-low restriction and low guidance-low restriction. We tested their effect on ten constructs of user experience measures on 66 European participants, all working in agriculture or forestry.The study did not find any significant effects of the three preference elicitation methods on all user experience constructs collected through questionnaires. However, we did find significant differences in terms of the objective measures chat duration (Speed), response time (Cognitive Demand) and recommendation performance (Accuracy of Recommended Items). Regarding the recommendation performance, it was found that the preference elicitation methods with high guidance led to a higher match score than the condition with low guidance. The certainty score was highest in the condition with high guidance and high input restriction. Finally, we found through a question at the end of the conversation that users who were satisfied with the recommendation responded more positively to six out of ten user experience constructs. This suggests that satisfaction with the recommendation performance is a crucial factor in the user experience of CRSs.

Client preferences for the design and delivery of injectable opioid agonist treatment services: Results from a best-worst scaling task.

Clinical trials support injectable opioid agonist treatment (iOAT) for individuals with opioid use disorder (OUD) for whom other pharmacological management approaches are not well-suited. However, despite substantial research indicating that person-centered care improves engagement, retention and health outcomes for individuals with OUD, structural requirements (e.g. drug policies) often dictate how iOAT must be delivered, regardless of client preferences. This study aimed to quantify clients' iOAT delivery preferences to improve client engagement and retention. Cross-sectional preference elicitation survey. Metro Vancouver, British Columbia, Canada. 124 current and former iOAT clients. Participants completed a demographic questionnaire package and an interviewer-led preference elicitation survey (case 2 best-worst scaling task). Latent class analysis was used to identify distinct preference groups and explore demographic differences between preference groups. Most participants (n = 100; 81%) were current iOAT clients. Latent class analysis identified two distinct groups of client preferences: (1) autonomous decision-makers (n = 73; 59%) and (2) shared decision-makers (n = 51; 41%). These groups had different preferences for how medication type and dosage were selected. Both groups prioritized access to take-home medication (i.e. carries), the ability to set their own schedule, receiving iOAT in a space they like and having other services available at iOAT clinics. Compared with shared decision-makers, fewer autonomous decision-makers identified as a cis-male/man and reported flexible preferences. Injectable opioid agonist treatment (iOAT) clients surveyed in Vancouver, Canada, appear to prefer greater autonomy than they currently have in choosing OAT medication type, dosage and treatment schedule.

Elicitation of societal preferences for chronic lymphocytic leukemia’s treatments: a discrete choice experiment

The overall value of treatments for chronic lymphocytic leukemia (CLL) depends on several factors, including preferences of the general population, who contributes to the financing of health systems. This study investigated societal preferences for attributes of CLL treatments in Italy. An online large-scale survey was designed using a discrete choice experiment (DCE) methodology and delivered to the Italian adult general population. Ten treatment attributes were identified, covering efficacy, safety, operational aspects and (hypothetical) out-of-pocket cost. DCE data were analyzed using a mixed logit regression model, estimating the willingness-to-pay for attribute levels’ change. The general population significantly preferred more effective treatments, with shorter duration, administered orally rather than orally + intravenously. Changes in therapy duration, frequency of checkups and organ damage risk had the greatest impact on preferences. The integration of societal preferences in the value judgments of CLL therapies may help health authorities in establishing priority setting and taking pricing-reimbursement decisions.

Exploring Anesthesia Provider Preferences for Precision Feedback: Preference Elicitation Study.

Health care professionals must learn continuously as a core part of their work. As the rate of knowledge production in biomedicine increases, better support for health care professionals' continuous learning is needed. In health systems, feedback is pervasive and is widely considered to be essential for learning that drives improvement. Clinical quality dashboards are one widely deployed approach to delivering feedback, but engagement with these systems is commonly low, reflecting a limited understanding of how to improve the effectiveness of feedback about health care. When coaches and facilitators deliver feedback for improving performance, they aim to be responsive to the recipient's motivations, information needs, and preferences. However, such functionality is largely missing from dashboards and feedback reports. Precision feedback is the delivery of high-value, motivating performance information that is prioritized based on its motivational potential for a specific recipient, including their needs and preferences. Anesthesia care offers a clinical domain with high-quality performance data and an abundance of evidence-based quality metrics. The objective of this study is to explore anesthesia provider preferences for precision feedback. We developed a test set of precision feedback messages with balanced characteristics across 4 performance scenarios. We created an experimental design to expose participants to contrasting message versions. We recruited anesthesia providers and elicited their preferences through analysis of the content of preferred messages. Participants additionally rated their perceived benefit of preferred messages to clinical practice on a 5-point Likert scale. We elicited preferences and feedback message benefit ratings from 35 participants. Preferences were diverse across participants but largely consistent within participants. Participants' preferences were consistent for message temporality (α=.85) and display format (α=.80). Ratings of participants' perceived benefit to clinical practice of preferred messages were high (mean rating 4.27, SD 0.77). Health care professionals exhibited diverse yet internally consistent preferences for precision feedback across a set of performance scenarios, while also giving messages high ratings of perceived benefit. A "one-size-fits-most approach" to performance feedback delivery would not appear to satisfy these preferences. Precision feedback systems may hold potential to improve support for health care professionals' continuous learning by accommodating feedback preferences.

Systematic Review of Patient Preference Studies in Non-metastatic Breast Cancer Adjuvant Medication Therapy: Attribute Selection

Context: Breast cancer poses significant challenges due to its high incidence and prevalence, necessitating heightened attention. Understanding how patients prioritize different treatment options based on various attributes can assist healthcare decision-makers in maximizing patient utility. The discrete choice experiment, a conjoint method, facilitates preference elicitation by presenting different attributes and choices. This systematic review aims to identify key factors in patient preference research related to adjuvant treatment for early breast cancer characterized by hormone receptor-positive, HER2-negative status. Evidence Acquisition: PubMed, Embase, Web of Science, and Scopus were searched from 01.01.2000 to 31.03.2023. Original English articles reporting patient preferences in adjuvant breast cancer treatment were retrieved based on predefined inclusion and exclusion criteria. Included studies were examined through a narrative synthesis approach, with descriptive statistics employed for analysis. Results: Out of 1163 articles reviewed, four met the inclusion criteria and were conducted in the USA, Canada, and the Netherlands. Attributes extracted from all studies included alopecia, sensory neuropathy, motor neuropathy, myalgia/arthralgia, nausea, vomiting, fatigue, neutropenia, mucositis/stomatitis, hand-foot syndrome, diarrhea, prevention of breast cancer recurrence, osteoporosis, risk of endometrial cancer, joint and muscle pain, fluid retention, libido decrease, hot flashes, ECG monitoring, efficacy, treatment regimen, 5-year invasive disease-free survival (iDFS), dosing schedule, and treatment duration. The most frequently reported attributes were side effects, efficacy, and treatment regimen. Systematic review was commonly used to determine which attributes and levels to include. The minimum number of attributes identified per study was seven, and the maximum was 12. Sample sizes ranged from 102 to 300, with none of the studies mentioning the method of sample size estimation. Ordinary Least Squares, logistic regression, and hierarchical Bayes regression were the most frequent analysis methods. Conclusions: Side effects, 5-year iDFS, and treatment regimen are three attributes identified for conducting discrete choice experiment studies. Utilizing conjoint analysis to assess patient preferences for breast cancer treatment can aid in selecting optimal treatment regimens and improving patient adherence. Moreover, adhering to guidelines for developing experimental designs and conducting data analysis is essential for yielding robust results when employing preference elicitation methods.

Health state utility assessment for low-grade glioma.

e14039 Background: Low-grade gliomas (LGGs, WHO Grade II gliomas) undergo malignant transformation, portending a median survival of 5-15 years. Though resection and adjuvant therapies have significantly improved survival, there are increased risks of long-term neurological decline and treatment-related side effects. As a result, not only quantity but also quality of life (QoL) must be considered when appraising therapeutic paradigms. Health state utility assessments can quantify society’s perceptions of given health states, supporting quality-adjusted life years (QALYs) and cost-effectiveness analyses. The primary and secondary objectives were to calculate LGG health states utility values (UVs) and interpret differences between health states, respectively. Methods: Health state descriptions were developed from a comprehensive literature review, expert input, and focus groups involving neuro-oncology providers and patient advocates. Healthy, adult participants were surveyed through an online REDCap form. Two preference elicitation techniques, visual analog scale (VAS) and time trade-off (TTO), were used to derive UVs of ten LGG health states. UVs are anchored at 0 representing death and 1 for perfect health. Mann-Whitney U tests compared UVs across health states. Results: Ninety-eight participants were included. Most participants were male (52%), aged 18-29 (81.6%), and White or Asian (78%). Stable LGG had the highest UVs (Mean [SD]: VAS- 0.64 [0.21]; TTO- 0.54 [0.42]) compared to Progressive LGG (VAS- 0.44 [0.26], p<0.0001; TTO- 0.36 [0.37], p<0.0001), Stable LGG with Chemoradiation (VAS- 0.43 [0.23], p<0.0001; TTO- 0.32 [0.34], p<0.0001), and Stable LGG with Motor (VAS- 0.45 [0.23], p<0.0001; TTO- 0.23 [0.32], p<0.0001) and Language (VAS- 0.52 [0.22], p<0.0001; TTO- 0.43 [0.40], p=0.04) Deficits. Progressive LGG UVs were not statistically different from those in Progressive LGG with Motor (VAS- 0.43 [0.23], p=0.69; TTO- 0.32 [0.34], p=0.73), Language (VAS- 0.43 [0.23], p=0.62; TTO- 0.32 [0.34], p=0.60), or Visual (VAS- 0.43 [0.23], p=0.78; TTO- 0.32 [0.34], p=0.60) Deficits. Radiation negatively impacted UVs, as UVs of Stable LGG with Chemoradiation (VAS- 0.43 [0.23]; TTO- 0.32 [0.34]) were significantly lower than UVs of Stable LGG with Chemotherapy alone (VAS- 0.43 [0.23], p=0.018; TTO- 0.32 [0.34], p=0.28) on VAS. Conclusions: These are the first-ever reported LGG health state UVs. These data will enable future QALY and cost-effective analyses, which have increasing importance as new investigational therapies become available for LGG.

How voting rules impact legitimacy

Collective action is essential for addressing the grand challenges of our time. However, for such action to be successful, decision-making processes must be perceived as legitimate. In this study, we investigate the legitimacy of different voting methods. Using a pre-registered human subject experiment, 120 participants cast their votes using four voting methods: majority voting, combined approval voting, range voting, and the modified Borda count. These methods represent a range of preference elicitation designs, from low to high complexity and flexibility. Furthermore, we developed a legitimacy scale upon which the participants rate the voting methods. The experiment was conducted in a non-political setting (voting on color preferences) and a political context (voting on COVID-19-related questions). Our findings suggest that the perceived legitimacy of a voting method is context-dependent. Specifically, preferential voting methods are seen as more legitimate than majority voting in a political decision-making situation, but only for individuals with well-defined preferences. Furthermore, preferential voting methods are more legitimate than majority voting in a highly polarized situation.

Monetary value of health-a practical decision-making framework combining equity considerations and WTP.

We estimate the first monetary value of a health gain in Croatia to inform the debate about the appropriate "demand-side" cost-effectiveness thresholds in Croatia but also Central and Eastern Europe, where such debates are still uncommon. We test the empirical support for two equity considerations: age and severity operationalized as proportional shortfall (PS), and propose a pragmatic framework for combining equity considerations with the monetary value of health into a single threshold. We used the contingent valuation method to elicit the willingness to pay per Quality-Adjusted Life Year (QALY) in Croatia, using a representative sample of the population (N = 1,500, online survey). 29 EQ-5D health states were valued using payment scales and open-ended question as payment vehicles. To test the hypotheses, we used both parametric tests and non-parametric tests. Multilinear regression was employed to test the theoretical validity of the results. The monetary value of a health gain in Croatia is equivalent to 1.15 of GDP per capita (equaling €17,000). Age of patients seems to be an important equity-related characteristic. The WTP per QALY in the age-neutral risk group (€11,900) was nearly equivalent to the WTP per QALY in the adult (neutral) risk group (€11,700) but lower by 16% compared to the WTP per QALY estimated in children (€14,200; p = 0.00). WTP estimates are theoretically valid and to, a small degree, scale sensitive. There is a positive association between the level of proportional shortfall and willingness to pay. To increase the usefulness of our results for the policy-makers, we combine the elicited preferences into a single decision-making framework and construct several cost-effectiveness thresholds based on willingness to pay and equity-related preferences. Based on empirical results, cost-effectiveness thresholds could range up to €20,308 for the most severe health conditions in children or could be lowered to €16,777 for less severe health conditions. In Central and Eastern Europe, in spite of a growing understanding of the importance of further developing value-based assessment frameworks there has been very little empirical research to guide, inform and promote this development. Countries in this region use mainly GDP-based thresholds without empirical evidence to support such important decisions. This may lead to thresholds that are too high, with detrimental consequence for the pricing and reimbursement systems.

Putting the Choice in Choice Tasks: Incorporating Preference Elicitation Tasks in Health Preference Research.

Choice-based preference elicitation methods such as the discrete choice experiment (DCE) present hypothetical choices to respondents, with an expectation that these hypothetical choices accurately reflect a 'real world' health-related decision context and that consequently the choice data can be held to be a true representation of the respondent's health or treatment preferences. For this to be the case, careful consideration needs to be given to the format of the choice task in a choice experiment. The overarching aim of this paper is to highlight important aspects to consider when designing and 'setting up' the choice tasks to be presented to respondents in a DCE. This includes the importance of considering the potential impact of format (e.g. choice context, choice set presentation and size) as well as choice set content (e.g. labelled and unlabelled choice sets and inclusion of reference alternatives) and choice questions (stated choice versus additional questions designed to explore complete preference orders) on the preference estimates that are elicited from studies. We endeavoure to instil a holistic approach to choice task design that considers format alongside content, experimental design and analysis.

Increased farmer willingness to pay for quality cassava (Manihot esculenta Crantz) planting materials: evidence from experimental auctions in Cambodia and Lao PDR

AbstractVarietal turnover is a major mechanism for farmer adaptation to rapidly changing environmental and market conditions. Understanding factors related to varietal turnover is critical for safeguarding food security in the face of increasingly rapid changes to production contexts. Preference elicitation methods like experimental auctions reveal farmers’ willingness to pay (WTP) for seeds with different characteristics. We engaged farmers growing cassava (Manihot esculenta Crantz) across a wide geographic scope in Cambodia (n = 321, 4 provinces) and Lao PDR (n = 391, 5 provinces) in modified second-price experimental auctions to determine WTP for three cassava seed classes: farmer-produced planting stems of an undescribed variety, farmer-produced planting stems of an elite variety with lower susceptibility to Cassava Mosaic Disease (CMD), and virus-free tested planting stems of the same elite variety. In Lao PDR, mean auction bids for a bundle of 20 planting stems were 0.54, 0.92, and 1.40 USD for farmer, elite, and elite tested stems, while in Cambodia the analogous mean bids were 0.58, 0.77, and 1.16 USD, respectively. A significantly higher WTP associated with elite variety and tested stems in both countries was influenced by both geographic location (province) and individual socioeconomic farmer characteristics. In Lao PDR this included lower WTP of women, higher WTP of ethnic minorities, and, for improved stems, lower WTP of farmers with increased cassava growing experience. In Cambodia fewer socioeconomic variables affected WTP. Our results indicate significant increases in WTP for clean planting materials and new varieties, supporting market-based clean seed approaches as a component of regional control strategies for the CMD epidemic currently threatening a global agri-food value chain.

“It’s hard to say anything definitive about what severity really is”: lay conceptualisations of severity in a healthcare context

BackgroundDemand for healthcare outweighs available resources, making priority setting a critical issue. ‘Severity’ is a priority-setting criterion in many healthcare systems, including in Norway, Sweden, the Netherlands, and the United Kingdom. However, there is a lack of consensus on what severity means in a healthcare context, both in the academic literature and in policy. Further, while public preference elicitation studies demonstrate support for severity as a relevant concern in priority setting, there is a paucity of research on what severity is taken to mean for the public. The purpose of this study is to explore how severity is conceptualised by members of the general public.MethodsSemi-structured group interviews were conducted from February to July 2021 with members of the Norwegian adult public (n = 59). These were transcribed verbatim and subjected to thematic analysis, incorporating inductive and deductive elements.ResultsThrough the analysis we arrived at three interrelated main themes. Severity as subjective experience included perceptions of severity as inherently subjective and personal. Emphasis was on the individual’s unique insight into their illness, and there was a concern that the assessment of severity should be fair for the individual. The second theme, Severity as objective fact, included perceptions of severity as something determined by objective criteria, so that a severe condition is equally severe for any person. Here, there was a concern for determining severity fairly within and across patient groups. The third theme, Severity as situation dependent, included perceptions of severity centered on second-order effects of illness. These included effects on the individual, such as their ability to work and enjoy their hobbies, effects on those surrounding the patient, such as next of kin, and effects at a societal level, such as production loss. We also identified a concern for determining severity fairly at a societal level.ConclusionsOur findings suggest that severity is a polyvalent notion with different meanings attached to it. There seems to be a dissonance between lay conceptualisations of severity and policy operationalisations of the term, which may lead to miscommunications between members of the public and policymakers.

Generating Usage-related Questions for Preference Elicitation in Conversational Recommender Systems

A key distinguishing feature of conversational recommender systems over traditional recommender systems is theirability to elicit user preferences using natural language. Currently, the predominant approach to preference elicitation is to ask questions directly about items or item attributes. Users searching for recommendations may not have deep knowledge of the available options in a given domain. As such, they might not be aware of key attributes or desirable values for them. However, in many settings, talking about the planned use of items does not present any difficulties, even for those that are new to a domain. In this article, we propose a novel approach to preference elicitation by asking implicit questions based on item usage. As one of the main contributions of this work, we develop a multi-stage data annotation protocol using crowdsourcing, to create a high-quality labeled training dataset. Another main contribution is the development of four models for the question generation task: two template-based baseline models and two neural text-to-text models. The template-based models use heuristically extracted common patterns found in the training data, while the neural models use the training data to learn to generate questions automatically. Using common metrics from machine translation for automatic evaluation, we show that our approaches are effective in generating elicitation questions, even with limited training data. We further employ human evaluation for comparing the generated questions using both pointwise and pairwise evaluation designs. We find that the human evaluation results are consistent with the automatic ones, allowing us to draw conclusions about the quality of the generated questions with certainty. Finally, we provide a detailed analysis of cases where the models show their limitations.

Task sensitivity and noise: How mechanical properties of preference elicitation tasks account for differences in preferences across tasks.

Task sensitivity and noise: How mechanical properties of preference elicitation tasks account for differences in preferences across tasks.