Developing attributes and attribute-levels for a discrete choice experiment on micro health insurance in rural Malawi

Would you zoom with your doctor? A discrete choice experiment to identify patient preferences for video and in-clinic consultations in German primary care

INTRODUCTION

The popularity of video consultations in healthcare has accelerated during the COVID-19 pandemic. Despite increased availability and obvious benefits, many patients remain hesitant to use video consultations. This study investigates the relative importance of the consultation mode compared to other attributes in patients' appointment choices in Germany.

METHODS

A discrete choice experiment was conducted to examine the influence of appointment attributes on preferences for video over in-clinic consultations. A total of 350 participants were included in the analysis.

RESULTS

The level of continuity of care (46%) and the waiting time until the next available appointment (22%) were shown to have higher relative importance than consultation mode (18%) and other attributes. Participants with fewer data privacy concerns, higher technology proficiency, and more fear of COVID-19 tended to prefer video over in-clinic consultations. The predicted choice probability of a video over a typical in-clinic consultation and opting out increased from <1% to 40% when the video consultation was improved from the worst-case to the best-case scenario.

CONCLUSION

This study provides insight into the effect of the consultation mode on appointment choice at a time when telemedicine gains momentum. The results suggest that participants preferred in-clinic over video consultations. Policymakers and service providers should focus on increasing the level of continuity of care and decreasing the time until the next available appointment to prompt the adoption of video consultations. Although participants preferred to talk to their physician in person over consulting via video per se, the demand for video consultations can be increased significantly by improving the other appointment attributes of video consultations such as the level of continuity of care.

Understanding how they really feel: Lesson learned from four approaches to soliciting user preferences for new contraceptive products in development

Background

An expanded range of contraceptive methods could reduce unintended pregnancies. User preferences research is important for successful development of products people want to use. This paper describes four approaches to preferences research soliciting user input in different ways: 1) perspectives on contraceptive method characteristics, 2) reactions to products in development, 3) trade-offs between contraceptive method characteristics, and 4) "blue-sky" ideas on novel contraceptive technologies.

Methods

We conducted two mixed-method studies: one implemented in Burkina Faso and Uganda combining three of these approaches, and the other implemented in India and Nigeria using two approaches. We share observations on the strengths and weaknesses of each approach and draw on our experience to highlight lessons learned for future user preferences studies.

Results

Each approach contributes to product development in different ways, and the usefulness of each methodology depends on the product development stage and corresponding informational needs.

Conclusions

Recommendations for future research include combining different methods, angles, and perspectives; using sequential designs whenever possible; tailoring product descriptions to user understanding for optimal feedback; and acknowledging the value and limitations of both quantitative results for modeling demand and idiosyncratic ideas to inspire development of new products.

Job preference of preventive medicine students during the COVID-19 pandemic: a discrete choice experiment survey in Shandong Province, China

BACKGROUND

Public health workers are a crucial part of the health workforce, particularly during the coronavirus disease (COVID-19) pandemic. They play an important role in achieving universal health coverage and sustainable development goals. Human resources in public health in China are in short supply, their distribution is unequal, and their turnover rate is high. A discrete choice experiment (DCE) was applied to investigate preventive medicine students' preferred job choice criteria and trends in trade-offs by calculating the marginal rate of substitution between these criteria. This study identified the properties of jobs primarily selected by preventive medicine students and estimated the monetary value of each attribute.

METHODS

Based on discussions and in-depth interviews with preventive medicine students and a literature review, we developed a DCE that assessed how students' stated preferences for a certain choice were influenced by several job attributes, including location, salary, bianzhi, career development opportunities, working environment, and workload. We applied this DCE to preventive medicine students in Shandong Province, China, using a brief, structured questionnaire. Conditional logit models were used to estimate the utility of each job's attributes. Willingness to pay (WTP) was estimated as the ratio of the value of the coefficient of interest to the negative value of the cost attribute.

RESULTS

A total of 307 respondents completed the questionnaire, and 261 passed the internal consistency test. All the attributes were statistically significant. Career development opportunities and work locations were the most important factors for the respondents. Preference heterogeneity existed among respondents, e.g., 3-year medical education college students placed a higher value on jobs with bianzhi compared to 5-year medical education college students. Furthermore, rural students' WTP for a job located in the county or city is much lower than that of urban students.

CONCLUSIONS

The heterogeneity of attributes indicates the complexity of job preferences. Monetary and nonmonetary job characteristics significantly influenced the job preferences of preventive medicine students in China. A more effective policy intervention to attract graduates to work in rural areas should consider both job incentives and the backgrounds of preventive medicine graduates.

Best practices for implementing biosafety inspections in a clinical laboratory: Evidence from a multi-site experimental study

OBJECTIVES

To explore the key components when designing best practice inspection interventions, so as to induce high compliance with safety guidelines for laboratory workers.

METHODS

Five key components of an inspection intervention, identified from a focus group discussion, were used as the attributes of a discrete choice experiment (DCE). In the DCE, participants were presented with two hypothetical scenarios and asked to choose the scenario in which they were more willing to comply with the laboratory safety guidelines. Data were collected from 35 clinical laboratories in seven healthcare institutes located in Chengdu, China. In total, 188 laboratory workers completed the DCE. The collected data were analyzed using conditional logit regression and latent class analysis.

RESULTS

Five key attributes were identified as the most important ones to best ensure laboratory safety: the inspector, the inspection frequency, the inspection timing, the communication of the inspection outcome, and a follow-up with either a reward or a punishment. By investigating the laboratory workers' responses to the attributes, properly implementing the five attributes could improve the workers' compliance from 25.86% (at the baseline case) to 74.54%. Compliance could be further improved with the consideration of the laboratory workers' heterogeneous reactions. In this study, two classes of workers, A and B, were identified. Compliance percentages for Classes A and B would be improved to 85.48% and 81.84%, respectively, when the key attributes were properly implemented for each class. The employment type and the size of the laboratory could be used to predict class membership.

CONCLUSION

The findings indicate the importance of an employee-centered approach in encouraging a worker's compliance. This approach also supports the design of tailored interventions by considering the laboratory workers' heterogeneous responses to the interventions.

Modelling the case-based learning preferences of undergraduate nursing students using a discrete choice experiment in China

PURPOSE

To investigate the preferences for case-based learning programmes among undergraduate nursing students.

METHOD

A questionnaire was designed based on a discrete choice experiment, and 227 undergraduate nursing students were investigated. In STATA 15.0 software, the data were statistically analysed using a mixed logit model.

RESULT

All attributes in our study were found to have a significant influence on undergraduate nursing students' preferences for case-based learning programmes. The students' preference for the CBL programme was influenced by the clinical internship experience and type of university. Furthermore, the most ideal scenario was found to be video case modality, unfolding delivery, provided by academic experts and clinical instructors, group size 9-11, adequate feedback, and fragmented case content.

CONCLUSION

The undergraduate nursing students' preferences for case-based learning programmes were affected by the provider, case modality, modality, group size, feedback, and case content. Our results can provide useful information for nursing educators to gain insight into student preferences and formulate case-based learning programs.

How can we elicit health workers' preferences for measures to reduce informal payments? A mixed methods approach to developing a discrete choice experiment in Tanzania

OBJECTIVE

While discrete choice experiments (DCEs) have been used in other fields as a means of eliciting respondent preferences, these remain relatively new in studying corrupt practices in the health sector. This study documents and discusses the process of developing a DCE to inform policy measures aimed at addressing informal payments for healthcare in Tanzania.

DESIGN

A mixed methods design was used to systematically develop attributes for the DCE. It involved five stages: a scoping literature review, qualitative interviews, a workshop with health providers and managers, expert review and a pilot study.

SETTING

Dar es Salaam and Pwani regions in Tanzania.

PARTICIPANTS

Health workers and health managers.

RESULTS

A large number of factors were identified as driving informal payments in Tanzania and thus represent potential areas for policy intervention. Through iterative process involving different methods (qualitative and quantitative) and seeking consensus views by diverse actors, we derived six attributes for a DCE: mode of payment, supervision at the facility level, opportunity for private practice, awareness and monitoring, disciplinary measures against informal payments and incentive payment for staff if a facility has less informal payments. 12 choice sets were generated and piloted with 15 health workers from 9 health facilities. The pilot study revealed that respondents could easily understand the attributes and levels, answered all the choice sets and appeared to be trading between the attributes. The results from the pilot study had expected signs for all attributes.

CONCLUSIONS

We elicited attributes and levels for a DCE to identify the acceptability and preferences of potential policy interventions to address informal payments in Tanzania through a mixed-methods approach. We argue that more attention is needed to the process of defining attributes for the DCE, which needs to be rigorous and transparent in order to derive reliable and policy-relevant findings.

Clinician Preferences for Oxybate Treatment for Narcolepsy: Survey and Discrete Choice Experiment

INTRODUCTION

Immediate-release sodium oxybate (SXB) has been Food and Drug Administration (FDA)-approved to treat narcolepsy since 2002; in 2020, a mixed-salt oxybates formulation was also approved. Both are taken at bedtime with a second dose taken 2.5-4 h later. A third oxybate option, an investigational extended-release SXB, may soon be available. This study was undertaken to understand clinicians' preferences between these 3 different oxybate treatments.

METHODS

Clinicians in active clinical practice for 3-35 years and experience treating patients with narcolepsy were recruited. A 30-min web-based survey quantified narcolepsy disease-state attitudes, treatment perceptions, and satisfaction with oxybates on 9-point scales. A discrete choice experiment (DCE) of 12 choice sets, with 2 hypothetical treatment profiles in each, was used to capture clinician preferences about overall oxybate therapy preference, impact on patient quality of life (QoL), and patient anxiety/stress. Attributes associated with current therapies and those expected to be available in the near future were included in the design.

RESULTS

The clinicians surveyed (n = 100) indicated that narcolepsy has a negative impact on patient QoL (mean rating, 7.7) and rated impact on QoL and treatment efficacy as the most important aspects of a narcolepsy treatment (mean rating, 7.3-7.7). Clinicians with experience prescribing oxybates had moderately high satisfaction with SXB and mixed-salt oxybates efficacy (mean ratings, 6.5-6.9) and safety (mean ratings, 6.1-6.7) and lower satisfaction with nightly dosing frequency (mean rating, 5.9 and 6.3, respectively). In the DCE, dosing frequency was the most important attribute driving overall product choice, patient QoL, and reducing patient anxiety/stress (relative attribute importance, 46.1, 41.7, and 44.0, respectively), with once nightly preferred over twice nightly.

CONCLUSION

Clinicians indicated a significantly higher preference for the once-at-bedtime dosing schedule versus twice nightly in selecting oxybate therapies overall and when aiming to improve patient QoL or reduce patient anxiety.

Stakeholder Engagement in Designing Attributes for a Discrete Choice Experiment With Policy Implications: An Example of 2 Swiss Studies on Healthcare Delivery

OBJECTIVES

The use of discrete choice experiments (DCEs) has become increasingly popular in health policy development by involving and analyzing the preferences of target groups for different aspects of the policy intervention. In this article, we aim to contribute to the standardization of the process of attribute and attribute level development for DCEs with policy relevance. To do so, we propose and empirically illustrate a framework tool for the development and reporting of attributes and attribute levels, with the systematic engagement of relevant stakeholders.

METHODS

The framework tool was kept general to allow its use as a standard reporting approach on the development process of a DCE, involving relevant stakeholders at each stage. The proposed 3-stage process consists of (1) collection, (2) selection, and (3) refinement of the DCE attributes and attribute levels. The application of the framework tool is illustrated by 2 projects, Health2040 and COCONUTS-both concerning the preferences of the Swiss population for the future organization of healthcare.

RESULTS

By engaging stakeholders, we identified and included attributes and attribute levels that would not have been identified solely through the existing literature. In addition, including multiple stakeholders from different professional backgrounds ensured that the selected attributes and attribute levels were policy relevant, were applicable in practice, and reflected the opinions of multiple actors in healthcare.

CONCLUSION

The proposed framework helps to strengthen the standardization of the reporting on attribute and attribute level development with stakeholder engagement and to guide future research teams designing a DCE with policy implications.

Development of attributes and levels of mental health insurance services using a discrete choice experiment

BACKGROUND

Despite the fact that mental illness is among the ten top diseases with the highest burden, the health services required by these patients do not have adequate insurance coverage. The purpose of this study is to develop the attributes and levels of mental health insurance services using a discrete choice experiment (DCE).

MATERIALS AND METHOD

This study involved a qualitative phase of the DCE that was conducted in Iran in 2020-2021 and included several stages. First, during a literature review, the attributes and levels were determined. Then, the attributes of health insurance were identified and weighed through virtual and in-person interviews with 16 mental health insurance professionals and policymakers in this field who were selected by purposive sampling. Finally, after a few sessions, through review studies, interviews, and a group of the expert panel, attributes and levels were finalized.

RESULTS

This study showed that coverage of inpatient services, outpatient services, place of receiving services, use of online internet services, limitation of services, and monthly premiums were the most important attributes of mental health insurance services.

CONCLUSION

To promote mental health insurance, policymakers and health insurance organizations should pay attention to premiums to be commensurate with the payment of people, packages of mental health services, and the ability of people to pay in appropriation with inflation. Identifying these attributes can determine people's willingness to pay and preferences for mental health insurance and lead to better planning for more comprehensive coverage for patients and increase the desirability of individuals in receiving services.

Adopting social health insurance in Nepal: A mixed study

Objective

The Social Health Insurance Program (SHIP) shares a major portion of social security, and is also key to Universal Health Coverage (UHC) and health equity. The Government of Nepal launched SHIP in the Fiscal Year 2015/16 for the first phase in three districts, on the principle of financial risk protection through prepayment and risk pooling in health care. Furthermore, the adoption of the program depends on the stakeholders' behaviors, mainly, the beneficiaries and the providers. Therefore, we aimed to explore and assess their perception and experiences regarding various factors acting on SHIP enrollment and adherence.

Methods

A cross-sectional, facility-based, concurrent mixed-methods study was carried out in seven health facilities in the Kailali, Baglung, and Ilam districts of Nepal. A total of 822 beneficiaries, sampled using probability proportional to size (PPS), attending health care institutions, were interviewed using a structured questionnaire for quantitative data. A total of seven focus group discussions (FGDs) and 12 in-depth interviews (IDIs), taken purposefully, were conducted with beneficiaries and service providers, using guidelines, respectively. Quantitative data were entered into Epi-data and analyzed with SPSS, MS-Excel, and Epitools, an online statistical calculator. Manual thematic analysis with predefined themes was carried out for qualitative data. Percentage, frequency, mean, and median were used to describe the variables, and the Chi-square test and binary logistic regression were used to infer the findings. We then combined the qualitative data from beneficiaries' and providers' perceptions, and experiences to explore different aspects of health insurance programs as well as to justify the quantitative findings.

Results and prospects

Of a total of 822 respondents (insured-404, uninsured-418), 370 (45%) were men. Families' median income was USD $65.96 (8.30-290.43). The perception of insurance premiums did not differ between the insured and uninsured groups (p = 0.53). Similarly, service utilization (OR = 220.4; 95% CI, 123.3-393.9) and accessibility (OR = 74.4; 95% CI, 42.5-130.6) were found to have high odds among the insured as compared to the uninsured respondents. Qualitative findings showed that the coverage and service quality were poor. Enrollment was gaining momentum despite nearly a one-tenth (9.1%) dropout rate. Moreover, different aspects, including provider-beneficiary communication, benefit packages, barriers, and ways to go, are discussed. Additionally, we also argue for some alternative health insurance schemes and strategies that may have possible implications in our contexts.

Conclusion

Although enrollment is encouraging, adherence is weak, with a considerable dropout rate and poor renewal. Patient management strategies and insurance education are recommended urgently. Furthermore, some alternate schemes and strategies may be considered.

Attributes development for a discrete choice experiment on preferences in sexual and reproductive health services for adolescents and young people in Nigeria

BACKGROUND

A major component of the validity of the discrete choice experiment (DCE) research design lies in the correct specification of attributes and levels relevant to the research focus. In this paper, we set out the validation steps we took in designing the tool for a DCE on preferences in sexual and reproductive health (SRH) services for adolescents and young people.

METHODOLOGY

This study was carried out among adolescents and young people (AYP) in Ogun State, Southwest Nigeria. We used a three-step mixed-methods process in developing the attributes and attribute-levels for our DCE tool. The first was to conduct a series of 16 focus group discussions (FGD) with AYP ensuring maximal variation (by age group, sex, marital status, and location). The FGD included a priority listing process in which participants were asked to list and rank the most important characteristics of optimal SRH services for AYP. The lists were harmonized and items were scored. The main (highest scoring) themes emerging from the harmonized priority list were converted into an initial set of attributes and the subthemes as level. These initial attributes and levels were presented to a panel of methods and content experts in a virtual modified Delphi process. This was for deciding on the importance of the attributes in providing optimum sexual and reproductive health services for young people, and the appropriateness of the levels. The same set of attributes was presented to another set of AYP in a series of four FGD to clarify meanings, and test whether the wordings were well understood. We applied some decision rules for including and excluding attributes and levels in the different phases of the development process.

RESULTS

We extracted an initial set of nine attributes with 2-4 levels each from the first FGD sessions. These were revised to a final set of seven attributes with 2-4 levels each based on findings from the expert review and final validation FGDs with AYP. The final attributes were: the type of staff, physical environment, health worker attitude, cost, waiting time, contraceptive availability, and opening hours.

CONCLUSION

The final set of attributes covered those relating to the services provided, the health workers providing the services, and the AYP. Our three-step process which included both quantitative and qualitative approaches ensured a rigorous process that produced a reliable combination of attributes and levels. Although we had to trade off some competing attributes to come to a final list, our decision rules helped us to conduct a transparent and reproducible process.

Eliciting relative preferences for the attributes of health insurance schemes among rural consumers in India

There is a limited understanding of the preferences of rural consumers in India for health insurance schemes. In this article, we investigate the preferences of the rural population for the attributes of a health insurance scheme by implementing a discrete choice experiment (DCE). We identified six attributes through qualitative and quantitative study: enrollment, management, benefit package, coverage, transportation facility, and monthly premium. A D-efficient design of 18 choices has been constructed, each comprising two health insurance choices. We collected the representative sample from 675 household heads of the rural population through personal interviews. The preferences for the attributes and attribute levels were estimated using the multinomial logit (MNL) and random-parameter logit (RPL) models. The analysis shows that all attribute levels significantly affect the choice behavior (P < 0.05). The relative order of preferences for attributes are; enrollment, benefit package, monthly premium, management, coverage, and transportation.

Attributes Underlying Patient Choice for Telerehabilitation Treatment: A Mixed-Methods Systematic Review to Support a Discrete Choice Experiment Study Design

BACKGROUND

Across most healthcare systems, patients are the primary focus. Patient involvements enhance their adherence to treatment, which in return, influences their health. The objective of this study was to determine the characteristics (ie, attributes) and associated levels (ie, values of the characteristics) that are the most important for patients regarding telerehabilitation (TR) healthcare to support a future discrete choice experiment (DCE) study design.

METHODS

A mixed-methods systematic review was conducted from January 2005 to the end of July 2020 and the search strategy was applied to five different databases. The initial selection of articles that met the eligibility criteria was independently made by one researcher, two researchers verified the accuracy of the extracted data, and all researchers discussed about relevant variables to include. Reporting of this systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and the Mixed Methods Appraisal Tool (MMAT) was used to assess the quality of the study. A qualitative synthesis was used to summarize findings.

RESULTS

From a total of 928 articles, 11 (qualitative [n = 5], quantitative [n = 3] and mixed-methods [n = 3] design) were included, and 25 attributes were identified and grouped into 13 categories: Accessibility, Distance, Interaction, Technology experience, Treatment mode, Treatment location, Physician contact mode, Physician contact frequency, Cost, Confidence, Ease of use, Feeling safer, and Training session. The attributes levels varied from two to five. The DCE studies identified showed the main stages to undertake these types of studies.

CONCLUSION

This study could guide the development of interview grid for individual interviews and focus groups to support a DCE study design in the TR field. By understanding the characteristics that enhance patients' preferences, healthcare providers can create or improve TR programs that provide high-quality and accessible care. Future research via a DCE is needed to determine the relative importance of the attributes.

Patient and Public Preferences for Coordinated Care in Switzerland: Development of a Discrete Choice Experiment

Objective

Our objective was to develop and test a discrete choice experiment (DCE) eliciting public and patient preferences for better-coordinated care in Switzerland.

Methods

We applied a multistage mixed-methods procedure using qualitative and quantitative approaches. First, to identify attributes, we performed a review of the DCE literature in healthcare with a focus on chronic care. Next, attribute selection involved stakeholders (N = 7) from various healthcare sectors to select the most relevant and actionable attributes, followed by three organized focus groups involving the general public and patients (N = 21) to verify the selection and the clarity of the DCE tasks and explanations. Finally, we conducted an online pilot in the target population to test the survey and obtain priors for a final six tested attributes to refine the final design of the experiment.

Results

After identifying an initial 33 attributes, a final list of six attributes was selected following stakeholder involvement and the three focus groups involving the target population. At the online pilot-testing stage with 301 participants, the majority of respondents found the DCE choice tasks socially relevant for Switzerland but challenging. The quality of the answers was relatively high. Most attributes had signs matching those in the literature and focus group discussions.

Conclusion

This article will be useful to researchers designing DCEs from a broad health policy perspective. The multistage approach involving a range of stakeholders was essential for the development of a DCE that is relevant for policy makers and well-accepted by the general public and patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40271-021-00568-2.

Qualitative Research Informing a Preference Study on Selecting Cannabis for Cancer Survivor Symptom Management: Design of a Discrete Choice Experiment

Introduction

The legalization of recreational cannabis use can enable cancer survivors to manage aspects of their care with cannabinoids without medical authorization or stigmatization. However, the absence of medical guidance—from the scientific literature or the healthcare system—makes it difficult for survivors to reach informed decisions about their care.

Objective

This article outlines the qualitative research undertaken to design a discrete choice experiment (DCE) aimed at understanding Canadian cancer survivors’ preferences for managing their cancer symptoms with cannabis in this complex socio-medical context.

Methods

In this study, we drew on previously published qualitative research (a literature review and interviews with cancer survivors) and the theory of planned behavior, holding weekly team meetings to review the qualitative data and identify initial attributes associated with medicinal cannabis consumption to inform the DCE design. The initial attributes were further assessed to determine whether they were sensitive to the Canadian context, modifiable to produce levels and trade-offs, and amenable to policy intervention, in order to form the DCE choice sets. The choice sets were tested via think-aloud exercises with members of the general population and included debriefing interviews. Think-aloud participants were recruited from patient groups and previous studies.

Results

Based on our review of the interview study, we identified the following attributes associated with selecting medicinal cannabis: effectiveness; chance of side effects; support from family, friends, and/or physicians; cost; and availability. Ability to perform everyday activities was added and monthly out-of-pocket cost was refined to render the DCE realistic to cancer survivors in the Canadian context. Revisions to the DCE instructions, terminology, and cost levels were made based on results from the think-aloud exercises (n = 10).

Conclusions

This qualitative study outlines the preference evidence collected regarding Canadian cancer survivors’ decisions to manage their symptoms with cannabis to inform a DCE quantitative survey. It contributes to transparent reporting of qualitative work in DCE development and to understanding cancer survivors’ preferences regarding medicinal cannabis consumption under legalization.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40271-021-00567-3.

Development of a Quantitative Instrument to Elicit Patient Preferences for Person-Centered Dementia Care Stage 1: A Formative Qualitative Study to Identify Patient Relevant Criteria for Experimental Design of an Analytic Hierarchy Process

Background: Person-centered care (PCC) requires knowledge about patient preferences. This formative qualitative study aimed to identify (sub)criteria of PCC for the design of a quantitative, choice-based instrument to elicit patient preferences for person-centered dementia care. Method: Interviews were conducted with n = 2 dementia care managers, n = 10 People living with Dementia (PlwD), and n = 3 caregivers (CGs), which followed a semi-structured interview guide including a card game with PCC criteria identified from the literature. Criteria cards were shown to explore the PlwD's conception. PlwD were asked to rank the cards to identify patient-relevant criteria of PCC. Audios were verbatim-transcribed and analyzed with qualitative content analysis. Card game results were coded on a 10-point-scale, and sums and means for criteria were calculated. Results: Six criteria with two sub-criteria emerged from the analysis; social relationships (indirect contact, direct contact), cognitive training (passive, active), organization of care (decentralized structures and no shared decision making, centralized structures and shared decision making), assistance with daily activities (professional, family member), characteristics of care professionals (empathy, education and work experience) and physical activities (alone, group). Dementia-sensitive wording and balance between comprehensibility vs. completeness of the (sub)criteria emerged as additional themes. Conclusions: Our formative study provides initial data about patient-relevant criteria of PCC to design a quantitative patient preference instrument. Future research may want to consider the balance between (sub)criteria comprehensibility vs. completeness.

Patient and clinician characteristics and preferences for increasing participation in placebo surgery trials: a scoping review of attributes to inform a discrete choice experiment

BACKGROUND

Orthopaedic surgeries include some of the highest volume surgical interventions globally; however, studies have shown that a significant proportion of patients report no clinically meaningful improvement in pain or function after certain procedures. As a result, there is increasing interest in conducting randomised placebo-controlled trials in orthopaedic surgery. However, these frequently fail to reach recruitment targets suggesting a need to improve trial design to encourage participation. The objective of this study was to systematically scope the available evidence on patient and clinician values and preferences which may influence the decision to participate in placebo surgery trial.

METHODS

A systematic review was conducted via a literature search in the MEDLINE, Embase, PsycInfo, CINAHL, and EconLit databases as of 19 July 2021, for studies of any design (except commentaries or opinion pieces) based on two key concepts: patient and clinician characteristics, values and preferences, and placebo surgery trials.

RESULTS

Of 3424 initial articles, we retained 18 eligible studies. Characteristics, preferences, values, and attitudes of patients (including levels of pain/function, risk/benefit perception, and altruism) and of clinicians (including concerns regarding patient deception associated with placebo, and experience/training in research) influenced their decisions to participate in placebo-controlled trials. Furthermore, some aspects of trial design, including randomisation procedures, availability of the procedure outside of the trial, and the information and consent procedures used, also influenced decisions to participate.

CONCLUSION

Participant recruitment is a significant challenge in placebo surgery trials, and individual decisions to participate appear to be sensitive to preferences around treatment. Understanding and quantifying the role patient and clinician preferences may play in surgical trials may contribute to the optimisation of the design and implementation of clinical trials in surgery.

Combining service design and discrete choice experiments for intervention design: An application to weather index insurance

In this paper we provide a detailed description of the methodological steps involved in conducting a Service Design study in combination with Discrete Choice Experiments (DCEs). It complements the conceptual and epistemological argument developed for this methodological combination in Osborne et al. (2021, World Development, in review WD-19535). Service Design for the co-creative development of policy interventions in complex adaptive systems involves an iterative process of moving between the six methodological stages of (1) problem co-definition, (2) actor-centred mapping, (3) experience-based problem diagnosis, (4) rapid prototyping, (5) design and testing and (6) upscaling. We suggest using DCEs as a quantitative method that is contextually adaptable and comparatively fast and cheap to implement, as part of stage (6) design and testing. Whilst both methods can operate independently with their own strengths and limitations, we find their combination to add value to the processes and outcomes of each. We illustrate the general methodological approach with a step-by-step description of its application to Weather Index Insurance in eastern Uganda.

Bullet points:

• Service Design co-creatively develops well-targeted solutions in complex adaptive systems.

• Discrete Choice Experiments quantitatively elicit actors’ preferences over the design of goods or services.

• Their combination can bring deeply contextualised, user-centred, operational and experimentally verified ideas for development interventions prior to their implementation.

Using discrete choice experiment to investigate patients' and parents' preferences for initiating ADHD medication

BACKGROUND

The decision to initiate medication is complex and is influenced by a variety of factors. There is limited information on the relative importance of factors that influence the initiation of ADHD medication.

AIMS

To investigate the factors, and their relative importance, that influence the decision to initiate medication in adults, and parents of children, with ADHD.

METHODS

A discrete choice experiment was conducted using eight choice tasks made up of five attributes that described the outcomes of initiating medication. A mixed multinomial logit model was used to estimate preferences for medication.

RESULTS

Respondents' overall preferences for initiating medication were negative (Mean (β)= -0.72705, p < 0.1), however, significant heterogeneity was noted in preferences (SD: 0.93604, p < 0.001). Side-effects were the most important factor for both adults (Relative importance (RI) = 40.39%) and parents (RI = 41.99%). Improvement in education had a greater weighting in adults' decision-making compared to parents (RI = 36.93% vs 30.47%) while improvement in aggressive (RI = 14.38% vs 11.84%) and social behaviour (RI = 12.59% vs 10.37%) was more important to parents.

CONCLUSIONS

Important differences in preferences of patients and parents were identified, highlighting that the decision to initiate medication is influenced differently in different individuals and groups.

Measuring Family Planning Provider Bias: A Discrete Choice Experiment among Burkinabé, Pakistani, and Tanzanian Providers

The unmet need for modern contraception remains high around the world, particularly for youth. While some of this unmet need is driven by limited health infrastructure and method mix availability, many adolescents who visit family planning providers still do not receive methods that fit their needs. This suggests that providers may be biased against youth and that interventions to change provider behavior could help close this gap. However, it is unclear if this bias is a result of age or other characteristics common among young women such as not being married and not having children. We use a discrete choice experiment in Burkina Faso, Pakistan, and Tanzania to disentangle the effects of age on providers' decisions to provide contraception from the effects of other potential confounding factors. We find that, although young women may experience the most bias, age is not the main driver. Rather, marital status and parity seem to influence provider decisions to offer services or counsel on modern methods. These findings suggest that interventions to reduce provider bias should focus on changing behavior towards unmarried and nulliparous women, regardless of their age.

Exploring willingness to pay for health insurance and preferences for a benefits package from the perspective of women from low-income households of Karachi, Pakistan

Background

Achieving universal health coverage (UHC) and reduction in out of pocket (OOP) expenditures on health, is a critical target of the Sustainable Development Goals (SDG). In low-middle income countries, micro-health insurance (MHI) schemes have emerged as a useful financing tool for laying grounds for Universal Health Coverage. The aim of this study was to provide evidence for designing a feasible health insurance scheme targeted at urban poor, by exploring preferences for an insurance benefits package and co-payments among women from low-income households in Karachi, Pakistan.

Methods

This was a descriptive cross-sectional study, conducted using household surveys between July–August 2015. A total of 167 female beneficiaries of Benazir Income Support Programme (BISP), a large-scale cash transfer scheme targeted at low-income households, were recruited in Karachi through a mix of convenience and snowball sampling. Hypothetical insurance benefits packages for a prospective health insurance scheme were formulated to capture respondents’ preferences for health insurance benefits package and co-payments. All data was analyzed using Stata (version 13).

Results

Respondents reporting expenditure on OPD and hospitalization in the last 2 weeks were 93.4 and 11.9% respectively. The highest median expenditure was incurred on medicines. Out of the proposed benefits package, a majority (53%) of the study participants opted for the comprehensive benefits package that provided coverage for emergency care, hospitalization, OPD consultation, diagnostic tests and transportation. For the co-payment plan, 38.9% participants preferred no co-payments that is 100% insurance coverage of medicines followed by hospitalization (25.9%). Nearly half of the respondents (49.4%) chose outpatient consultation for 50% co-payment. A majority of the participants (65.3%) agreed to 100% co-payment for the transportation cost.

Conclusion

Health insurance schemes can be introduced in urban areas, against collection of micro-payments, to prevent low-income households from facing financial catastrophe. A comprehensive benefits package covering emergency care, hospitalization, OPD consultation, diagnostic tests and transportation, is the most preferred among low-income beneficiaries.

Household preferences and willingness to pay for health insurance in Kampala City: a discrete choice experiment

INTRODUCTION

Uganda is in discussions to introduce a national health insurance scheme. However, there is a paucity of information on household preferences and willingness to pay for health insurance attributes that may guide the design of an acceptable health insurance scheme. Our study sought to assess household preferences and willingness to pay for health insurance in Kampala city using a discrete choice experiment.

METHODS

This study was conducted from 16th February 2020 to 10th April 2020 on 240 households in the Kawempe division of Kampala city stratified into slum and non-slum communities in order to get a representative sample of the area. We purposively selected the communities that represented slum and non-slum communities and thereafter applied systematic sampling in the selection of the households that participated in the study from each of the communities. Four household and policy-relevant attributes were used in the experimental design of the study. Each respondent attended to 9 binary choice sets of health insurance plans that included one fixed choice set. Data were analyzed using mixed logit models.

RESULTS

Households in both the non-slum and slum communities had a high preference for health insurance plans that included both private and public health care providers as compared to plans that included public health care providers only (non-slum coefficient β = 0.81, P < 0.05; slum β = 0.87, p < 0.05) and; health insurance plans that covered extended family members as compared to plans that had limitations on the number of family members allowed (non-slum β = 0.44, P < 0.05; slum β = 0.36, p < 0.05). Households in non-slum communities, in particular, had a high preference for health insurance plans that covered chronic illnesses and major surgeries to other plans (0.97 β, P < 0.05). Our findings suggest that location of the household influences willingness to pay with households from non-slum communities willing to pay more for the preferred attributes.

CONCLUSION

Potential health insurance schemes should consider including both private and public health care providers and allow more household members to be enrolled in both slum and non-slum communities. However, the inclusion of more HH members should be weighed against the possible depletion of resources and other attributes. Potential health insurance schemes should also prioritize coverage for chronic illnesses and major surgeries in non-slum communities, in particular, to make the scheme attractive and acceptable for these communities.

A Protocol for Developing a Discrete Choice Experiment to Elicit Preferences of General Practitioners for the Choice of Specialty

OBJECTIVES

Discrete choice experiment (DCE) has been increasingly used in studies investigating preferences of the health workforce. This pilot study aimed to provide a comprehensive protocol for design and quality assessment of a DCE to elicit preferences of general practitioners for the choice of specialty.

METHODS

We conducted 3 rigorous qualitative studies, including a review of literature, qualitative interviews, and focus group discussion to develop attributes and levels of the DCE. Then a D-efficient fractional factorial design was applied to generate choice tasks and to block them into 4 equal versions using SAS software. Two hundred and forty general practitioners participated in a pilot study conducted to evaluate the quality of the instrument using 7 tests.

RESULTS

The qualitative studies resulted in the development of 8 attributes and 24 levels. Experimental design led to the generation of 36 choice tasks divided into 4 blocks. The pilot study revealed that the instrument has validity and reliability and also indicated that the design is comprehensible and feasible.

CONCLUSIONS

This study provided an integrated, comprehensive framework for development and quality assessment of a DCE that could be used in other studies. The pilot study indicated that the instrument is of sufficient quality because it was developed using rigorous qualitative studies and accurate experimental design method.

A qualitative study to identify critical attributes and attribute-levels for a discrete choice experiment on oral pre-exposure prophylaxis (PrEP) delivery among young people in Cape Town and Johannesburg, South Africa

BACKGROUND

The uptake and adherence of daily oral PrEP has been poor in high-risk populations in South Africa including young people. We used qualitative research methods to explore user preferences for daily and on-demand oral PrEP use among young South Africans, and to inform the identification of critical attributes and attribute-levels for quantitative analysis of user preferences, i.e. a discrete choice experiment (DCE).

METHODS

Data were collected between September and November 2018 from eight group discussions and 20 in-depth interviews with young people 13 to 24 years in Cape Town and Johannesburg. Using a convenience sampling strategy, participants were stratified by sex and age. Interviewers used a semi-structured interview guide to discuss several attributes (dosing regimen, location, costs, side effects, and protection period) for PrEP access and use. Group discussions and in-depth interviews were audio-recorded, transcribed verbatim and translated to English. We used framework analysis to explore context-specific attributes and attribute-levels for delivering oral PrEP in South Africa. The adolescent community advisory board, expert and study team opinions were consulted for the final DCE attributes and levels.

RESULTS

We enrolled 74 participants who were 51% (n = 38/74) male, had a median age of 18.5 [Interquartile range = 16-21.25] years, 91% (n = 67/74) identified as heterosexual and 49% (n = 36/74) had not completed 12th grade education. Using the qualitative data, we identified five candidate attributes including (1) dosing regimen, (2) location to get PrEP, (3) cost, (4) route of administration and (5) frequency. After discussions with experts and the study team, we revised the DCE to include the following five attributes and levels: dosing regime: daily, and on-demand PrEP; location: private pharmacy, public clinic, mobile clinic, ATM); cost: free-of-charge, R50 (~2GBP), R265 (~12GBP); side effects: nausea, headache, rash; and duration of protection: fulltime protection versus when PrEP is used).

CONCLUSIONS

There is limited literature on qualitative research methods describing the step-by-step process of developing a DCE for PrEP in adolescents, especially in resource-constrained countries. We provide the process followed for the DCE technique to understand user preferences for daily and on-demand oral PrEP among young people in South Africa.

Patient Preferences in Surveillance: Findings From a Discrete Choice Experiment in the "My Follow-Up" Study

OBJECTIVES

Approximately 800 000 people die globally from colorectal cancer (CRC) every year. Prevention programs promote early detection, but for people with precancerous lesions, tailoring surveillance to include lifestyle-change programs could enhance prevention potential and improve outcomes.

METHODS

Those with intermediate or high-risk polyps removed during CRC screening colonoscopy within the Northern Ireland CRC Screening Programme were invited to complete 8 discrete choice questions about tailored surveillance, analyzed using random-parameters logit and a latent class modeling approach.

RESULTS

A total of 231 participants (77% male) self-reported comorbid hypertension (53%), high cholesterol (48%), and mean body mass index of 28.7 (overweight). Although 39% of participants were unaware of their CRC risk status, 30.9% indicated they were already making changes to reduce their risk. Although all respondents were significantly risk- and cost-averse, the latent class analysis identified 3 segments (classes): 1. Class 1 (26.8%) significantly favored phone or email support for a lifestyle change, a 17-month testing interval, and noninvasive testing. 2. Class 2 (48.4%) preferred the status quo. 3. Class 3 (24.7%) significantly favored further risk reduction and invasive testing.

CONCLUSIONS

This is the first documented preference study focusing on postpolypectomy surveillance offering lifestyle interventions. Although current care is strongly preferred, risk and cost aversion are important for participants. Latent class analysis shows that some respondents are willing to change diet and lifestyle behaviors, reflecting a teachable moment, with opportunities to personalize and optimize surveillance. Significant discordance between perceived and known risk of recurrence and limited recall of risk information provided within current practice suggest necessary improvements to surveillance programs.

Attributes Development for Pharmaceutical Subsidization: A Qualitative Study

Discrete choice experiments (DCES) as a stated preference method have used increasingly to determine preferences attached to some attributes associated to health. Although, the validity of this type of studies comprehensively depends on the appropriate determination of attributes and attribute-levels for DCES, there is little rigorous evidence regarding which factors or attributes and attribute- levels should be counted for eliciting public preferences in health resource allocation. This paper responds to such question by carefully doing a qualitative study. A qualitative study used semi-structured interviews, which were audio recorded, transcribed and subject to thematic analysis. Sixteen participants had been key informants and decision makers of pharmaceutical and health system. Initially, by conducting a meticulous literature review, an inclusive list of attributes associated with intended policy was identified. Qualitative data for the development of attributes and their levels were collected using 16 key informant interviews and were analyzed by software MAXQDA followed by a focus group discussion (FGD) with 7 people, well-familiar with the notion pharmaceutical policy and Pharmacoeconomics. The 311 codes in four main dimensions were initially identified by conducting interviews. However, for being manageable within a DCE, they were classified and limited to four attributes, including severity of disease without treatment, health gain after treatment, frequency of patients, and cost of treatment per patient. This qualitative study provides enough evidence for designing and doing a precise discrete choice experiment answering the question about public preferences in pharmaceutical subsidization and contributes empirical evidence to the limited methodological literature on attributes development for DCE, specifically within low and middle-income countries.

Preferences of healthcare providers for capitation payment in Kenya: a discrete choice experiment

Provider payment mechanisms (PPMs) are important to the universal health coverage (UHC) agenda as they can influence healthcare provider behaviour and create incentives for health service delivery, quality and efficiency. Therefore, when designing PPMs, it is important to consider providers' preferences for PPM characteristics. We set out to uncover senior health facility managers' preferences for the attributes of a capitation payment mechanism in Kenya. We use a discrete choice experiment and focus on four capitation attributes, namely, payment schedule, timeliness of payments, capitation rate per individual per year and services to be paid by the capitation rate. Using a Bayesian efficient experimental design, choice data were collected from 233 senior health facility managers across 98 health facilities in seven Kenyan counties. Panel mixed multinomial logit and latent class models were used in the analysis. We found that capitation arrangements with frequent payment schedules, timelier disbursements, higher payment rates per individual per year and those that paid for a limited set of health services were preferred. The capitation rate per individual per year was the most important attribute. Respondents were willing to accept an increase in the capitation rate to compensate for bundling a broader set of health services under the capitation payment. In addition, we found preference heterogeneity across respondents and latent classes. In conclusion, these attributes can be used as potential targets for interventions aimed at configuring capitation to achieve UHC.

Using discrete choice experiments to measure preferences for hard to observe choice attributes to inform health policy decisions

BACKGROUND

Models of preferences in health services research (HSR) and Health Economics are often defined by readily available information, such as that captured in claims data and electronic health records. Yet many important questions about patient choices cannot be easily studied because of a lack of critical data elements. The objective of this review is to outline the advantages of using stated preferences (SP) data in health services research, and to outline how these methods can be used to evaluate choices that have not yet been offered or studied.

MAIN BODY

This article focuses on the application of DCE's to relevant policy and health system delivery questions currently relevant, particularly in the United States. DCE's may be helpful to collect data from patient or consumer data that we currently do not have. The article provides examples of research questions that have been answered using SP data collected with a DCE. It outlines how to construct a DCE and how to analyze the data. It also discusses the methodological challenges and emphasizes important considerations regarding the design and estimation methods. SP data can be adopted in situations where we would like to have consumer choice data, but we currently do not. These are often hypothetical situations to analyze the decision-making process of individuals. With SP data it is possible to analyze trade-offs patients make when choosing between treatment options where these hard to measure attributes are important.

CONCLUSION

This paper emphasizes that a carefully designed DCE and appropriate estimation methods can open up a new world of data regarding trade-offs patients and providers in healthcare are willing to make. It updates previous "how to" guide for DCE's for health services researchers and health economists who are not familiar with these methods or have been unwilling to use them and updates previous description of these methods with timely examples.

How Do Type 2 Diabetes Patients Value Urban Integrated Primary Care in China? Results of a Discrete Choice Experiment

Objectives: Fragmented healthcare in China cannot meet the needs of the growing number of type 2 diabetes patients. The World Health Organization proposed an integrated primary care approach to address the needs of patients with chronic conditions. This study aims to measure type 2 diabetes patients' preferences for urban integrated primary care in China. Methods: A discrete choice experiment was designed to measure type 2 diabetes patient preferences for seven priority attributes of integrated care. A two-stage sampling survey of 307 type 2 diabetes mellitus (T2DM) patients in 16 community health stations was carried out. Interviews were conducted to explore the reasons underpinning the preferences. A logit regression model was used to estimate patients' willingness to pay and to analyze the expected impact of potential policy changes. Results: Travel time to care providers and experience of care providers are the most valued attributes for respondents rather than out-of-pocket cost. Attention to personal situation, the attentiveness of care providers, and the friendliness and helpfulness of staff were all related to interpersonal communication between patients and health care providers. Accurate health information and multidisciplinary care were less important attributes. Conclusions: The study provides an insight into type 2 diabetes patients' needs and preferences of integrated primary care. People-centered interventions, such as increasing coverage by family doctor and cultivating mutual continuous relationships appear to be key priorities of policy and practice in China.

Attribute development and level selection for a discrete choice experiment to elicit the preferences of health care providers for capitation payment mechanism in Kenya

BACKGROUND

Stated preference elicitation methods such as discrete choice experiments (DCEs) are now widely used in the health domain. However, the "quality" of health-related DCEs has come under criticism due to the lack of rigour in conducting and reporting some aspects of the design process such as attribute and level development. Superficially selecting attributes and levels and vaguely reporting the process might result in misspecification of attributes which may, in turn, bias the study and misinform policy. To address these concerns, we meticulously conducted and report our systematic attribute development and level selection process for a DCE to elicit the preferences of health care providers for the attributes of a capitation payment mechanism in Kenya.

METHODOLOGY

We used a four-stage process proposed by Helter and Boehler to conduct and report the attribute development and level selection process. The process entailed raw data collection, data reduction, removing inappropriate attributes, and wording of attributes. Raw data was collected through a literature review and a qualitative study. Data was reduced to a long list of attributes which were then screened for appropriateness by a panel of experts. The resulting attributes and levels were worded and pretested in a pilot study. Revisions were made and a final list of attributes and levels decided.

RESULTS

The literature review unearthed seven attributes of provider payment mechanisms while the qualitative study uncovered 10 capitation attributes. Then, inappropriate attributes were removed using criteria such as salience, correlation, plausibility, and capability of being traded. The resulting five attributes were worded appropriately and pretested in a pilot study with 31 respondents. The pilot study results were used to make revisions. Finally, four attributes were established for the DCE, namely, payment schedule, timeliness of payments, capitation rate per individual per year, and services to be paid by the capitation rate.

CONCLUSION

By rigorously conducting and reporting the process of attribute development and level selection of our DCE,we improved transparency and helped researchers judge the quality.

Treatment preferences of patients with relapsed and refractory multiple myeloma: a qualitative study

BACKGROUND

Multiple myeloma is a haematological malignancy characterized by significant morbidity and mortality. This study sought to develop an in-depth understanding of patients' lived experiences of relapsed or refractory multiple myeloma (RRMM) and its treatment, and to identify which features of treatment were most important to them.

METHODS

Qualitative interviews and focus groups (FGs) were conducted with 32 people living with RRMM across Canada. In Phase 1, interviews focused on participants' accounts of their experiences with the disease and its treatment and laid the groundwork for the FGs (Phase 2). The FGs developed a deeper understanding of patients' treatment priorities. Interview and FG transcripts were coded for emergent themes and patterns.

RESULTS

The interviews identified important side effects that had significant impacts on patients' lives, including physical, cognitive, and psychological/emotional side effects. Participants also identified specific treatment features (attributes) that were important to them. These were compiled into a list and used in the FGs to understand patients' priorities. Higher prioritized attributes were: life expectancy, physical and cognitive side effects, and financial impact. Mode of administration, treatment intervals, psychological side effects, and sleep/mood effects were identified as lower priorities.

CONCLUSIONS

RRMM and its treatments impact importantly on patients' quality-of-life across a range of domains. Patients prioritized treatment features that could enhance life expectancy, minimize side effects and offset financial burdens.

IMPLICATIONS FOR CANCER SURVIVORS

A clear articulation of patient priorities can contribute to efforts to design treatment with patients' concerns in mind, thereby promoting a more patient-centered approach to care.

The Preferred Attributes of a Trauma Team Leader: Evidence From a Discrete Choice Experiment

OBJECTIVE

Leaders of a pediatric trauma team are tasked with managing rapidly changing diagnostic and treatment challenges, while ensuring the entire team functions effectively to produce optimal patient outcomes. An effective trauma team leader is often thought to be self-evident, and there is little formal literature identifying the leadership characteristics and attributes associated with optimal trauma team performance. The purpose of this study was to elicit the trauma team leader traits and characteristics deemed of greatest utility by members of the pediatric trauma team.

DESIGN, SETTING, PARTICIPANTS

Members of the pediatric trauma team at British Columbia Children's Hospital were asked to participate in a semistructured interview to identify trauma team leader attributes associated with maximal team performance. Using the attributes, we constructed a discrete choice experiment (DCE). DCEs, developed in the economics and market research setting, allow participants to express preferences among finite alternatives, with subsequent statistical analysis that allows quantitative comparison of the utility of selected attributes.

RESULTS

After interviewing 21 trauma team practitioners, 6 themes were identified as being most important for trauma team leadership. The developed DCE was administered to 64 members of the trauma team. Analysis of the DCE revealed the most important attributes were collaboration, strong communication, and decisiveness. The attribute of least utility was experience. The specific leadership qualities that provided the most utility to the trauma team included "actively involves input for team" (mean utility [MU]: 0.70; standard error [SE]: 0.11) and "concise communication, at times closed-loop" (MU: 0.52; SE: 0.09). "Hesitant and unclear communication" (MU: -0.88; SE: 0.09) and "often indecisive" (MU: -0.68; SE: 0.10) were deemed most detrimental (negative utility) to the team's function.

CONCLUSIONS

This study is novel in applying a strategy to identify and quantify the relative value of trauma team leader attributes. When designing education initiatives for pediatric trauma care teams, defining trauma team quality metrics, and providing continuing medical education for the team leader, it is essential to incorporate preferred leadership characteristics. Crisis resource management skills benefit greatly from an understanding of the preferred attributes, as defined and evaluated by other trauma team members.

How important is social support in determining patients' suitability for transplantation? Results from a National Survey of Transplant Clinicians

BACKGROUND

National guidelines require programmes use subjective assessments of social support when determining transplant suitability, despite limited evidence linking it to outcomes. We examined how transplant providers weigh the importance of social support for kidney transplantation compared with other factors, and variation by clinical role and personal beliefs.

METHODS

The National survey of the American Society of Transplant Surgeons and the Society of Transplant Social Work in 2016. Using a discrete choice approach, respondents compared two hypothetical patient profiles and selected one for transplantation. Conditional logistic regression estimated the relative importance of each factor; results were stratified by clinical role (psychosocial vs medical/surgical providers) and beliefs (outcomes vs equity).

RESULTS

Five hundred and eighy-four transplant providers completed the survey. Social support was the second most influential factor among transplant providers. Providers were most likely to choose a candidate who had social support (OR=1.68, 95% CI 1.50 to 1.86), always adhered to a medical regimen (OR=1.64, 95% CI 1.46 to 1.88), and had a 15 years life expectancy with transplant (OR=1.61, 95% CI 1.42 to 1.85). Psychosocial providers were more influenced by adherence and quality of life compared with medical/surgical providers, who were more influenced by candidates' life expectancy with transplant (p<0.05). For providers concerned with avoiding organ waste, social support was the most influential factor, while it was the least influential for clinicians concerned with fairness (p<0.05).

CONCLUSIONS

Social support is highly influential in listing decisions and may exacerbate transplant disparities. Providers' beliefs and reliance on social support in determining suitability vary considerably, raising concerns about transparency and justice.

A novel design process for selection of attributes for inclusion in discrete choice experiments: case study exploring variation in clinical decision-making about thrombolysis in the treatment of acute ischaemic stroke

BACKGROUND

A discrete choice experiment (DCE) is a method used to elicit participants' preferences and the relative importance of different attributes and levels within a decision-making process. DCEs have become popular in healthcare; however, approaches to identify the attributes/levels influencing a decision of interest and to selection methods for their inclusion in a DCE are under-reported. Our objectives were: to explore the development process used to select/present attributes/levels from the identified range that may be influential; to describe a systematic and rigorous development process for design of a DCE in the context of thrombolytic therapy for acute stroke; and, to discuss the advantages of our five-stage approach to enhance current guidance for developing DCEs.

METHODS

A five-stage DCE development process was undertaken. Methods employed included literature review, qualitative analysis of interview and ethnographic data, expert panel discussions, a quantitative structured prioritisation (ranking) exercise and pilot testing of the DCE using a 'think aloud' approach.

RESULTS

The five-stage process reported helped to reduce the list of 22 initial patient-related factors to a final set of nine variable factors and six fixed factors for inclusion in a testable DCE using a vignette model of presentation.

CONCLUSIONS

In order for the data and conclusions generated by DCEs to be deemed valid, it is crucial that the methods of design and development are documented and reported. This paper has detailed a rigorous and systematic approach to DCE development which may be useful to researchers seeking to establish methods for reducing and prioritising attributes for inclusion in future DCEs.

Developing attributes and levels for a discrete choice experiment on basic health insurance in Iran

Background: Nonmarket stated preferences valuation, especially discrete choice experiments (DCEs), is one of the commonly used techniques in the health sector. The primary purpose of this approach is to help select attributes and attributes-levels that are able to properly describe health care products or services. This study aimed at developing attributes and attributes-levels for basic health insurance system in Iran.

Methods: This study was conducted in 3 phases. First, narrative review was performed to identify related attributes. Also, 9 experts were interviewed to identify relevant attributes of health insurance in context. Other 36 experts rated the attributes and levels. Then, the research team decided on the inclusion of attributes and levels in the final design. The design was constructed using generic and Defficient method with SAS 9.1. The design was divided into 3 blocks, each having 8 choice sets. Finally, the choice set was piloted with 45 participants.

Results: Public hospitals, and private hospitals benefits, dental insurance coverage, inpatient benefits, rehabilitation therapy, and paraclinical benefits, long-term care, medical devices benefits (Ortez, Protez, etc.), and monthly premium were identified and included in the final attribute design (D-efficiency = 98.16). The pilot study revealed that participants could easily understand and answer all the choice sets.

Conclusion: The results of our study indicated that health insurance service benefit packages and premium were among the most important attributes that need to be included in the final attribute design for Iranians. The policymakers and health insurance organizations should emphasize these attributes in the benefit packages to make improvements. The emphasis on these attributes can help elicit people’s preferences and willingness to pay for attributes.

The Best of Both Worlds: An Example Mixed Methods Approach to Understand Men's Preferences for the Treatment of Lower Urinary Tract Symptoms

BACKGROUND

Discrete choice experiments (DCEs) are widely used to quantify individuals' preferences for healthcare. Guidelines recommend the design of DCEs should be informed by qualitative research. However, only a few studies go beyond guidelines by fully presenting qualitative and quantitative research jointly together in a mixed methods approach (MMA).

OBJECTIVES

Using an example study about men's preferences for medical treatment of lower urinary tract symptoms (LUTS), we demonstrate how qualitative research can complement DCEs to gain a rich understanding of individuals' preferences.

METHODS

We were the first to combine online discussion groups (ODGs) with an online DCE. A thematic analysis of the ODGs and a conceptual map provided insights into men's quality of life (QoL) with LUTS and relevant treatment attitudes. This was used to design the DCE. Men's willingness to pay (WTP) for these attributes was estimated. Findings from ODGs and DCE were compared to understand WTP and preference heterogeneity.

KEY FINDINGS

Men mostly valued medicine that reduced urgency and night-time frequencies of urination but avoided sexual side effects. We find heterogeneity in the effect of sexual side effects on men's preferences. The ODGs suggest this is because several men may be sexually inactive due to their age, being widowed or having comorbidities. The ODGs also raised concern about men's awareness of LUTS.

CONCLUSION

We argue that the insights gained into men's preferences for treatment and how LUTS affects men's QoL could not have been obtained by either the qualitative research or the DCE alone.

Investigation of attributes which guide choice in cataract surgery services in urban Sydney, Australia

BACKGROUND

It is critical to consult patients to develop patient-centred cataract surgery care. We aimed to identify attributes patients consider when making decisions about cataract surgery in an Australian context, where both publicly and privately funded surgery are available. This is the first step in investigating how decisions are made about cataract surgery services.

METHODS

This observational qualitative study was undertaken in two public hospitals and one private practice in Sydney, Australia. The study involved 19 women and men with age-related cataracts and no previous cataract surgery, aged > 18 years, able to speak conversational English or Mandarin. A multi-stage attribute development process was followed, including: literature review, semi-structured interviews with surgery candidates in three eye clinics, and review by an expert panel. The main outcome measures were primary attributes for making choices about cataract surgery.

RESULTS

Wait time, cost, institutional reputation, surgeon experience and travel time were identified as principal attributes; lower value was placed on consultation length and accessibility. Non-English speaking participants indicated greater interest in pre-operative information than English speakers, but expressed trust in the Australian healthcare system.

CONCLUSIONS

Findings suggest individuals prioritise attributes which consume time or incur costs when accessing care (wait time, cost and travel time). They also consider factors associated with the outcome of their cataract surgery (surgeon experience and institutional reputation). Similar to other decision-making processes, patients are likely to trade between these different attributes depending on their personal preferences and circumstances.

A discrete choice experiment on preferences of patients with rheumatoid arthritis regarding disease-modifying antirheumatic drugs: the identification, refinement, and selection of attributes and levels

OBJECTIVE

To comprehensively describe the identification, refinement, and selection of attributes and levels for a discrete choice experiment (DCE) on preferences of patients with rheumatoid arthritis (RA) regarding disease-modifying antirheumatic drugs (DMARDs).

METHODS

A mixed-methods approach, consisting of three consecutive steps: a literature review, expert recommendations, and focus groups. Attributes and levels were identified by a scoping review and compiled into a list that was evaluated on its relevance by an expert panel. The list that resulted thereafter was used to inform three focus groups, including 23 patients with RA. New attributes and levels could be identified during the focus groups. Also, a ranking exercise was performed. The patients individually ranked the attributes (ie, the ones on the list and newly identified attributes) by relevance. The patients' individual rankings were summed to derive a ranking at group level and make an a priori selection of the most relevant attributes. The group discussions were transcribed for qualitative analysis.

RESULTS

Nineteen attributes, each specified by two to seven levels, were identified by the scoping review. The expert recommendations resulted in the removal of one attribute. Furthermore, two new attributes and levels were identified and two attributes were split into two. One new attribute was identified during the focus groups. The results of the ranking exercise and qualitative analysis led to the refinement and selection of the following attributes: route of administration, frequency of administration, chance of efficacy, onset of action, risk of serious infections, risk of liver injury, and risk of cancer. Each attribute was specified by three levels.

CONCLUSION

This study contributes to the limited literature on the development of attributes and levels. Future research should pay more attention to a comprehensive description of this process. It ensures transparency and thereby allows researchers to judge a DCE's quality and generalizability.

Valuing Attributes of Home Palliative Care With Service Users: A Pilot Discrete Choice Experiment

CONTEXT

Discrete choice experiment (DCE) is a quantitative method that helps determine which service attributes are most valued by people and consequently improve their well-being.

OBJECTIVES

The objective of this study was to test a new DCE on home palliative care (HPC).

METHODS

Cross-sectional survey using the DCE method with adult patients and their family caregivers, users of three HPC services in Portugal. Service attributes were based on a Cochrane review, a meta-ethnography, and the few existing DCEs on HPC: 1) team's availability, 2) support for family caregivers, 3) homecare support, 4) information and planning, and 5) waiting time. The experimental design consisted in three blocks of eight choice sets where participants chose between two service alternatives that combined different levels of each attribute. We piloted the DCE using cognitive interviewing. Interviews were analyzed for difficulties using Tourangeau's model of information processing.

RESULTS

The DCE was conducted with 21 participants of 37 eligible (10 patients with median Palliative Performance Scale score = 45, 11 caregivers). Most participants found the DCE easy (median 2 from 1 to 5), although two patients did not finish the exercise. Key difficulties related to comprehension (e.g., waiting time sometimes understood as response time for visit instead of time from referral to care start) and judgment (e.g., indecision due to similar service alternatives).

CONCLUSION

The DCE method is feasible and acceptable but not all patients are able to participate. In the main study phase, we will give more attention to the explanation of the waiting time attribute.

Applying stated-preference methods to improve health systems in sub-Saharan Africa: a systematic review

INTRODUCTION

Sub-Saharan African health systems must balance shifting disease burdens with desires for robust institutions. Stated-preference methods have been applied extensively to elicit health care workers' preferences and priorities for rural practice. This systematic review characterizes the range of their applications to African health systems problems. Areas covered: A PRISMA protocol was submitted to PROSPERO. Six databases were queried for peer-reviewed articles using quantitative stated-preference methods to evaluate a health systems-related trade-off. Quality was assessed using the PREFS checklist. Seventy-seven articles published between 1996 and 2017 met review criteria. Methods were primarily choice-based: discrete-choice experiments (n = 46), ranking/allocation techniques (n = 21), conjoint analyses (n = 7), and best-worst scaling (n = 3). Trade-offs fell into six 'building blocks': service features (n = 27), workforce incentives (n = 17), product features (n = 14), system priorities (n = 14), insurance features (n = 4), and research priorities (n = 1). Five countries dominated: South Africa (n = 11), Ghana (n = 9), Malawi (n = 9), Uganda (n = 9), and Tanzania (n = 8). Discrete-choice experiments were of highest quality (mean score: 3.36/5). Expert commentary: Stated-preference methods have been applied to many health systems contexts throughout sub-Saharan Africa. Studies examined established strategic areas, especially primary health care for women, prevention and treatment of infectious diseases, and workforce development. Studies have neglected the emerging areas of non-communicable diseases.

Factors that influence variation in clinical decision-making about thrombolysis in the treatment of acute ischaemic stroke: results of a discrete choice experiment

Background

Intravenous thrombolysis for patients with acute ischaemic stroke is underused (only 80% of eligible patients receive it) and there is variation in its use across the UK. Previously, variation might have been explained by structural differences; however, continuing variation may reflect differences in clinical decision-making regarding the eligibility of patients for treatment. This variation in decision-making could lead to the underuse, or result in inappropriate use, of thrombolysis.

Objectives

To identify the factors which contribute to variation in, and influence, clinicians’ decision-making about treating ischaemic stroke patients with intravenous thrombolysis.

Methods

A discrete choice experiment (DCE) using hypothetical patient vignettes framed around areas of clinical uncertainty was conducted to better understand the influence of patient-related and clinician-related factors on clinical decision-making. An online DCE was developed following an iterative five-stage design process. UK-based clinicians involved in final decision-making about thrombolysis were invited to take part via national professional bodies of relevant medical specialties. Mixed-logit regression analyses were conducted.

Results

A total of 138 clinicians responded and opted to offer thrombolysis in 31.4% of cases. Seven patient factors were individually predictive of the increased likelihood of offering thrombolysis (compared with reference levels in brackets): stroke onset time of 2 hours 30 minutes (50 minutes); pre-stroke dependency modified Rankin Scale score (mRS) of 3 (mRS4); systolic blood pressure (SBP) of 185 mmHg (140 mmHg); stroke severity scores of National Institutes of Health Stroke Scale (NIHSS) 5 without aphasia, NIHSS 14 and NIHSS 23 (NIHSS 2 without aphasia); age 85 years (65 years); and Afro-Caribbean (white). Factors predictive of not offering thrombolysis were age 95 years; stroke onset time of 4 hours 15 minutes; severe dementia (no memory problems); and SBP of 200 mmHg. Three clinician-related factors were predictive of an increased likelihood of offering thrombolysis (perceived robustness of the evidence for thrombolysis; thrombolysing more patients in the past 12 months; and high discomfort with uncertainty) and one factor was predictive of a decreased likelihood of offering treatment (clinicians’ being comfortable treating patients outside the licensing criteria).

Limitations

We anticipated a sample size of 150–200. Nonetheless, the final sample of 138 is good considering that the total population of eligible UK clinicians is relatively small. Furthermore, data from the Royal College of Physicians suggest that our sample is representative of clinicians involved in decision-making about thrombolysis.

Conclusions

There was considerable heterogeneity among respondents in thrombolysis decision-making, indicating that clinicians differ in their thresholds for treatment across a number of patient-related factors. Respondents were significantly more likely to treat 85-year-old patients than patients aged 68 years and this probably reflects acceptance of data from Third International Stroke Trial that report benefit for patients aged > 80 years. That respondents were more likely to offer thrombolysis to patients with severe stroke than to patients with mild stroke may indicate uncertainty/concern about the risk/benefit balance in treatment of minor stroke. Findings will be disseminated via peer-review publication and presentation at national/international conferences, and will be linked to training/continuing professional development (CPD) programmes.

Future work

The nature of DCE design means that only a subset of potentially influential factors could be explored. Factors not explored in this study warrant future research. Training/CPD should address the impact of non-medical influences on decision-making using evidence-based strategies.

Funding

The National Institute for Health Research Health Services and Delivery Research programme.

Qualitative Development of a Discrete Choice Experiment for Physical Activity Interventions to Improve Knee Osteoarthritis

OBJECTIVE

To describe the qualitative process used to develop attributes and attribute levels for inclusion in a discrete choice experiments (DCE) for older adult physical activity interventions.

DESIGN

Five focus groups (n=41) were conducted, grounded in the Health Action Process Approach framework. Discussion emphasized identification and prioritization attributes for a DCE on physical activity. Semi-structured interviews (n=6) investigated attribute levels and lay-language for the DCE. A focus group with physical activity researchers and health care providers was the final stakeholder group used to establish a comprehensive approach for the generation of attributes and levels. A DCE pilot test (n=8) was then conducted with individuals of the target patient population. All transcripts were analyzed using a constant comparative approach.

SETTING

General community and university-based research setting.

PARTICIPANTS

Volunteers (N=55) aged >45 years with knee pain, aches, or stiffness for at least 1 month over the previous 12 months.

INTERVENTIONS

Not applicable.

MAIN OUTCOME MEASURES

Interview guides, attributes, attribute levels, and discrete choice experiment.

RESULTS

The most influential identified attributes for physical activity were time, effort, cost, convenience, enjoyment, and health benefits. Each attribute had 3 levels that were understandable in the pilot test of the DCE.

CONCLUSIONS

The identification of 6 physical activity attributes that are most salient to adults with knee osteoarthritis resulted from a systematic qualitative process, including attribute-ranking exercises. A DCE will provide insight into the relative importance of these attributes for participating in physical activity, which can guide intervention development.

Focus Groups in Elderly Ophthalmologic Patients: Setting the Stage for Quantitative Preference Elicitation

OBJECTIVES

Patients suffering from age-related macular degeneration (AMD) are rarely actively involved in decision-making, despite facing preference-sensitive treatment decisions. This paper presents a qualitative study to prepare quantitative preference elicitation in AMD patients. The aims of this study were (1) to gain familiarity with and learn about the special requirements of the AMD patient population for quantitative data collection; and (2) to select/refine patient-relevant treatment attributes and levels, and gain insights into preference structures.

METHODS

Semi-structured focus group interviews were performed. An interview guide including preselected categories in the form of seven potentially patient-relevant treatment attributes was followed. To identify the most patient-relevant treatment attributes, a ranking exercise was performed. Deductive content analyses were done by two independent reviewers for each attribute to derive subcategories (potential levels of attributes) and depict preference trends.

RESULTS

The focus group interviews included 21 patients. The interviews revealed that quantitative preference surveys in this population will have to be interviewer assisted to make the survey feasible for patients. The five most patient-relevant attributes were the effect on visual function [ranking score (RS): 139], injection frequency (RS: 101), approval status (RS: 83), side effects (RS: 79), and monitoring frequency (RS: 76). Attribute and level refinement was based on patients' statements. Preference trends and dependencies between attributes informed the quantitative instrument design.

CONCLUSION

This study suggests that qualitative research is a very helpful step to prepare the design and administration of quantitative preference elicitation instruments. It especially facilitated familiarization with the target population and its preferences, and it supported attribute/level refinement.

Attribute Development Using Continuous Stakeholder Engagement to Prioritize Treatment Decisions: A Framework for Patient-Centered Research

OBJECTIVES

To develop a methodological approach for selecting, validating, and prioritizing attributes for health care decision making.

METHODS

Participants (n = 48) were recruited from community support groups if they had a child aged 26 years or younger diagnosed with a coexisting mental health condition and cognitive impairment. Six in-depth interviews eliciting care management experiences were transcribed and coded into themes following the principles of grounded theory and the constant comparative method. Six focus groups involving 42 participants assessed the relevance, priority, and meaning and inter-relationship among the themes. The positive predictive value and sensitivity assessed agreement on thematic meaning. A final list was selected from the top priorities with good agreement as candidate attributes. Attribute levels reflecting the range of experiences in care management decisions emerged from the verbatim passages within each coded theme.

RESULTS

Participants were the child's mother (73%), white (77%), married (69%), and on average 48 years old. The children were on average 14 years old; 44% had an intellectual disability, 25% had autism, and more than half had anxiety or attention-deficit/hyperactivity disorder. All 14 attributes identified from the in-depth interviews were deemed relevant. The positive predictive value exceeded 90%, and the sensitivity ranged from 64% to 89%. The final set of attributes formed the framework for care management decisions consisting of six attributes (medication, behavior, services, social, treatment effects, and school) each with three levels.

CONCLUSIONS

A systematic approach grounded in qualitative methods produced a framework of relevant, important, and actionable attributes representing competing alternatives in clinical decisions.

Developing a Patient-Centered Benefit-Risk Survey: A Community-Engaged Process

OBJECTIVES

To provide a community-engaged process to inform the design of a stated-preferences experiment. The process involved integrating patients and caregivers of people with Duchenne/Becker muscular dystrophy, advocates, clinicians, and the sponsor in conceptualizing and developing a benefit-risk survey on the basis of phase III trial results.

METHODS

Our community-engagement process for the development of a stated-preference survey included a set of five guiding principles with a foundation in the principles of community-engaged research. Engagement efforts were carried out through an informal network of three committees. Members of the leadership, stakeholder, and review committees comprised patients, caregivers, clinicians, advocacy leadership, and industry representatives.

RESULTS

Committee members participated in 15 hours of formal engagement including interviews and conference calls that ranged from 45 to 90 minutes, plus additional less-formal ad hoc communication. Committees comprised 20 individuals across three committees including adults with DMD (n = 6), parents of children with DMD (n = 6), clinicians (n = 3), members of research and advocacy organizations (n = 4), and an industry representative (n = 1). Community engagement informed attribute selection, survey length, word choice, and eligibility criteria. Challenges in the process included managing diverse stakeholder perspectives, time requirements, and the inherent tension between outcomes used in clinical trials versus attributes that correspond to patient- and family-relevant outcomes.

CONCLUSIONS

We demonstrated how community engagement can successfully influence study design to support the design of a relevant survey instrument that is ethical, acceptable, meaningful to the community, and enhances patient-centered benefit-risk assessment for regulatory decision making.

What health plans do people prefer? The trade-off between premium and provider choice

Within a healthcare system with managed competition, health insurers are expected to act as prudent buyers of care on behalf of their customers. To fulfil this role adequately, understanding consumer preferences for health plan characteristics is of vital importance. Little is known, however, about these preferences and how they vary across consumers. Using a discrete choice experiment (DCE) we quantified trade-offs between basic health plan characteristics and analysed whether there are differences in preferences according to age, health status and income. We selected four health plan characteristics to be included in the DCE: (i) the level of provider choice and associated level of reimbursement, (ii) the primary focus of provider contracting (price, quality, social responsibility), (iii) the level of service benefits, and (iv) the monthly premium. This selection was based on a literature study, expert interviews and focus group discussions. The DCE consisted of 17 choice sets, each comprising two hypothetical health plan alternatives. A representative sample (n = 533) of the Dutch adult population, based on age, gender and educational level, completed the online questionnaire during the annual open enrolment period for 2015. The final model with four latent classes showed that being able to choose a care provider freely was by far the most decisive characteristic for respondents aged over 45, those with chronic conditions, and those with a gross income over €3000/month. Monthly premium was the most important choice determinant for young, healthy, and lower income respondents. We conclude that it would be very unlikely for half of the sample to opt for health plans with restricted provider choice. However, a premium discount up to €15/month by restricted health plans might motivate especially younger, healthier, and less wealthy consumers to choose these plans.