Individuals’ Preferences for Esophageal Cancer Screening: A Discrete Choice Experiment

ObjectivesThe increasing incidence of esophageal adenocarcinoma (EAC) and the dismal prognosis has stimulated interest in the early detection of EAC. Our objective was to determine individuals’ preferences for EAC screening and to assess to what extent procedural characteristics of EAC screening tests predict willingness for screening participation.MethodsA discrete choice experiment questionnaire was sent by postal mail to 1000 subjects aged 50 to 75 years who were randomly selected from the municipal registry in the Netherlands. Each subject answered 12 discrete choice questions of 2 hypothetical screening tests comprising 5 attributes: EAC-related mortality risk reduction, procedure-related pain and discomfort, screening location, test specificity, and costs. A multinomial logit model was used to estimate individuals’ preferences for each attribute level and to calculate expected rates of uptake.ResultsIn total, 375 individuals (37.5%) completed the questionnaire. Test specificity, pain and discomfort, mortality reduction, and out-of-pocket costs all had a significant impact on respondents’ preferences. The average expected uptake of EAC screening was 62.8% (95% confidence interval [CI] 61.1-64.5). Severe pain and discomfort had the largest impact on screening uptake (–22.8%; 95% CI –26.8 to –18.7). Male gender (β 2.81; P < .001), cancer worries (β 1.96; P = .01), endoscopy experience (β 1.46; P = .05), and upper gastrointestinal symptoms (β 1.50; P = .05) were significantly associated with screening participation.ConclusionsEAC screening implementation should consider patient preferences to maximize screening attendance uptake. Based on our results, an optimal screening test should have high specificity, cause no or mild to moderate pain or discomfort, and result in a decrease in EAC-related mortality.     

US Residents’ Preferences for Sharing of Electronic Health Record and Genetic Information: A Discrete Choice Experiment

ObjectivesThe aim to this study was to assess preferences for sharing of electronic health record (EHR) and genetic information separately and to examine whether there are different preferences for sharing these 2 types of information.MethodsUsing a population-based, nationally representative survey of the United States, we conducted a discrete choice experiment in which half of the subjects (N = 790) responded to questions about sharing of genetic information and the other half (N = 751) to questions about sharing of EHR information. Conditional logistic regression models assessed relative preferences across attribute levels of where patients learn about health information sharing, whether shared data are deidentified, whether data are commercialized, how long biospecimens are kept, and what the purpose of sharing the information is.ResultsIndividuals had strong preferences to share deidentified (vs identified) data (odds ratio [OR] 3.26, 95% confidence interval 2.68-3.96) and to be able to opt out of sharing information with commercial companies (OR 4.26, 95% confidence interval 3.42-5.30). There were no significant differences regarding how long biospecimens are kept or why the data are being shared. Individuals had a stronger preference for opting out of sharing genetic (OR 4.26) versus EHR information (OR 2.64) (P = .002).ConclusionsHospital systems and regulatory bodies should consider patient preferences for sharing of personal medical records or genetic information. For both genetic and EHR information, patients strongly prefer their data to be deidentified and to have the choice to opt out of sharing information with commercial companies.     

Assessing Payers’ Preferences for Real-World Evidence in the United States: A Discrete Choice Experiment

ObjectivesTo rank the US payers’ preferences for attributes of real-world evidence (RWE) studies in the context of chronic disease and to quantify trade-offs among them.MethodsWe conducted a discrete choice experiment in which 180 employees from payer organizations were tasked to choose between 2 RWE studies assuming they were assessing evidence to inform formulary decisions for chronic disease treatment. Each RWE study was characterized by 7 attributes with 3 levels each: very informative, moderately informative, and not measured. We used a D-optimal main-effects design. Survey data were fitted to a conditional logit model to obtain a relative measure of the ranking of importance for each attribute.ResultsClinical outcomes were the most preferred attribute. It was 4.68 times as important as productivity outcomes—the least preferred attribute. It was followed by health-related quality of life (2.78), methodologic rigor (2.09), resource utilization (1.71), and external validity (1.56).ConclusionsThis study provides a quantification of the value payers place on key RWE attributes. Across attributes, payers have higher preferences for clinical and health-related quality of life outcomes than the other attributes. Between attributes’ levels, payers prefer high levels of information in clinical outcomes and methodologic rigor but are indifferent in other attributes. Our results bridge the gap between the information that payers seek and the attributes that RWE studies prioritize and effectively guide future research design.     

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

ObjectivesApproximately 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.MethodsThose 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.ResultsA 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.ConclusionsThis 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.     

Mothers’ Preferences and Willingness to Pay for Human Papillomavirus Vaccination for Their Daughters: A Discrete Choice Experiment in Hong Kong

Objectives

To determine the preference of mothers in Hong Kong and their willingness to pay (WTP) for human papillomavirus (HPV) vaccination for their daughters.

How Do Treatment Preferences of Patients With Cancer Compare With Those of Oncologists and Family Members? Evidence From a Discrete Choice Experiment in China

ObjectivesThis study aimed to compare the treatment preference among oncologists, patients with lung cancer, and their family members in China.MethodsA face-to-face discrete choice experiment survey was conducted among oncologists, patients, and their family members. Participants completed 13 choice sets describing 6 key attributes, namely, overall survival time, risk of severe adverse effect, severity of pain, appetite, physical functioning status, and monthly cost. Mixed logit model and latent class analysis were used to estimate attribute level preference weights and the relative importance (RI) for attributes. The willingness to pay (WTP) and maximum acceptable risk (MAR) were also estimated. The RI, WTP, and MAR of oncologists, patients, and family members were compared.ResultsA total of 121 oncologists and 161 dyads of patients and family members completed the survey. Overall survival time, physical functioning status, and pain were the 3 most important attributes across all 3 groups. Oncologists and family members had higher RI on overall survival time than patients (48% and 51% vs 38%). Patients had higher RI on physical functioning status and pain (23% and 14%) than oncologists (13% and 12%) and family members (16% and 11%). For extending survival, patients had the least WTP, and family members had the highest MAR. The latent class analysis identified 2 classes in the patient group and 3 classes in oncologist and family member groups.ConclusionsThere were differences in preferences for survival, risk, quality of life, and costs associated with cancer treatments among patients, oncologists, and family members. This finding highlights the need of involving patients in treatment decision making in China.     

Discrete Choice Experiment to Estimate Breast Cancer Patients’ Preferences and Willingness to Pay for Prophylactic Granulocyte Colony-Stimulating Factors

ObjectivesRising out-of-pocket costs for cancer patients have increased shared decision making. Clinical guidelines recommend prophylactic granulocyte colony-stimulating factor (G-CSF) for patients receiving chemotherapy with a 20% or greater risk of febrile neutropenia. A discrete choice experiment was conducted to explore breast cancer patients’ preferences and willingness to pay (WTP) for prophylactic G-CSF to decrease the risk of chemotherapy-induced febrile neutropenia.MethodsAn online discrete choice experiment questionnaire survey of a national US convenience sample of self-reported breast cancer patients with prior chemotherapy treatment was conducted. Sixteen paired G-CSF treatment scenarios, each with four attributes (risk of disruption to chemotherapy schedule due to low white blood cell counts, risk of developing an infection requiring hospitalization, frequency of administration, and total out-of-pocket cost) were presented with a follow-up “no treatment” option. Participant preferences and WTP out of pocket were estimated by logistic regression.ResultsParticipants (n = 296) preferred G-CSF regimens with lower out-of-pocket costs, lower risk of chemotherapy disruption, lower risk of infection, and greater convenience (one G-CSF injection per chemotherapy cycle). Participants’ WTP was $1076 out of pocket per cycle to reduce the risk (high to low) of disrupting their chemotherapy schedule, $884 per cycle to reduce the risk (24% [high] to 7% [low]) of infection, and $851 per cycle to decrease the number of G-CSF injections (11 to 1) per cycle.ConclusionsParticipants highly valued specific features of prophylactic G-CSF treatment including maintaining their chemotherapy schedule, lowering their risk of infection, and reducing the number of injections. Physicians should consider patient preferences to inform the best treatment choices for individual patients.     

Does Attribute Framing in Discrete Choice Experiments Influence Willingness to Pay? Results from a Discrete Choice Experiment in Screening for Colorectal Cancer

Objective:  Recent reviews of discrete choice methodology identified methodological issues warranting further exploration, including the issue of "framing." The objective of this study was to conduct a methodological exploration of the effect of attribute framing on marginal rates of substitution (MRS), including willingness to pay (WTP) from a discrete choice experiment (DCE), within the context of colorectal cancer screening preferences.
Methods:  The survey, a fractional factorial design of a two-alternative, unlabeled experiment, was mailed to a sample of 1920 subjects in NSW, Australia. Participants were randomized to one of four alternative "frames" of information. Attributes included: accuracy of the test for finding cancers, accuracy of the test for finding large polyps, how good the test is at saying you don't have cancer, cost, dietary and medication restrictions and sample collection. A mixed logit model was used to estimate preferences; MRS between attributes, including WTP, was calculated.
Results:  A total of 1157 surveys from 1920 (60.2%) were returned. Accuracy of the test for finding cancer was most likely to influence choice of test, followed by accuracy of the test for finding large polyps. Under some circumstances, framing of the attributes (e.g., cancers found vs. cancers missed) influenced the relative importance of attributes. Attribute framing significantly influenced estimates of WTP, and benefit: harm trade-offs that were calculated from MRS.
Conclusions:  Attribute framing can influence willingness to pay and benefit: harm trade-offs from DCEs. Appropriate design and analysis methods should be explored to further characterize the influence and extent of framing in discrete choice studies.     

Patient Preferences in the Treatment of Hemophilia A: A Best–Worst Scaling Case 3 Analysis

ObjectiveTo assess patient preferences for benefits and risks in hemophilia A treatment.MethodsA systematic literature search and pretest interviews were conducted to determine the most patient-relevant endpoints in terms of effects, risks, and administration of hemophilia A treatments. A Best–Worst Scaling (BWS; Case 3 or multiprofile case) approach was applied in a structured questionnaire. Patients were surveyed by interviewers in a computer-assisted personal interview. Treatments in the choice scenarios comprised bleeding frequency per year, application type, risk of thromboembolic event risk, and inhibitor development. Each respondent answered 13 choice tasks, including 1 dominant task, comparing 3 treatment profiles. Data were analyzed using a mixed logit model (random-parameters logit).ResultsData from 57 patients were used. The attributes “bleeding frequency per year” and “inhibitor development” had the greatest impact on respondents’ choice decisions. Patients disliked being at risk of inhibitor development more than being at risk of thromboembolic events. The type of application, whether intravenous or subcutaneous, was of less importance for patients. There was a significant preference variation for all attributes.ConclusionsPatients value low frequency of bleeding per year and low risk of development of inhibitors the most. An increase of risk and frequency would significantly decrease the impact on choice decisions. The type of application does not seem to influence the choice decision very much compared with the other attributes. Regarding preference heterogeneity, further analysis is needed to identify subgroups among patients and their characteristics. This may help to adapt individually patient-tailored treatment alternatives for hemophilia A patients.     

Do Profit Margins of Pharmaceuticals Influence Reimbursement Decisions? A Discrete Choice Experiment Among Dutch Healthcare Decision Makers

ObjectivesThis study aimed to investigate whether the profit margins of pharmaceuticals would influence the outcome of reimbursement decisions within the Dutch policy context.MethodsWe conducted a discrete choice experiment among 58 Dutch decision makers. In 20 choice sets, we asked respondents to indicate which of 2 pharmaceutical treatment options they would select for reimbursement. Options were described using 5 attributes (disease severity, incremental costs per quality-adjusted life-year, health gain, budget impact, and profit margin) with 3 levels each. Additionally, cognitive debriefing questions were presented, and for validation debriefing, interviews were conducted. Choice data were analyzed using mixed logit models, also to calculate marginal effects and choice probabilities.ResultsResults indicated that the specified levels of profit margins significantly influenced choices made. Decision makers were less likely to reimburse a product with a higher profit margin. The relative importance of profit margins was lower than that of the included traditional health technology assessment criteria, but not negligible. When asked directly, 61% of respondents indicated that profit margin should play a role in reimbursement decision making, although concerns about feasibility and the connection to price negotiations were voiced.ConclusionsOur results suggest that if available to decision makers the profit margin of pharmaceutical products would influence reimbursement decisions within the Dutch policy context. Higher profit margins would reduce the likelihood of reimbursement. Whether adding profit margin as an additional, explicit criterion to the health technology assessment decision framework would be feasible and desirable is open to further exploration.     

Distressed, Immobilized, or Lacking Employer Support? A Sub-classification of Acute Work-Related Low Back Pain

Introduction One possibility for reducing the disabling effects of low back pain (LBP) is to identify subgroups of patients who might benefit from different disability prevention strategies. The aim of this study was to test the ability to discern meaningful patient clusters for early intervention based on self-reported concerns and expectations at the time of an initial medical evaluation. Methods Workers seeking an initial evaluation for acute, work-related LBP (N?=?496; 58?% male) completed self-report measures comprising of 11 possible risk factors for chronicity of pain and disability. Outcomes of pain, function, and return-to-work were assessed at 3-month follow-up. A K-means cluster analysis was used to derive patient subgroups based on risk factor patterns, and then these subgroups were compared with respect to 3-month outcomes. Results Eight of the 11 measures showed significant associations with functional recovery and return-to-work, and these were entered into the cluster analysis. A 4-cluster solution met criteria for cluster separation and interpretability, and the four clusters were labeled: (a) minimal risk (29?%), (b) workplace concerns (26?%); (c) activity limitations (27?%); and (d) emotional distress (19?%). Functional outcomes were best in the minimal risk group, poorest in the emotional distress group, and intermediate in the other two groups. A global severity index at baseline also showed highest overall risk in the emotional distressed group. Conclusions Patterns of early disability risk factors from this study suggest patients have differential needs with respect to overcoming emotional distress, resuming normal activity, and obtaining workplace support. Classifying patients in this manner may improve the cost-benefit of early intervention strategies to prevent long-term sickness absence and disability due to LBP.     

Calculating Preference Weights for the Labor and Delivery Index: A Discrete Choice Experiment on Women’s Birth Experiences

ObjectiveThe aim of this study was to calculate preference weights for the Labor and Delivery Index (LADY-X) to make it suitable as a utility measure for perinatal care studies.MethodsIn an online discrete choice experiment, 18 pairs of hypothetical scenarios were presented to respondents, from which they had to choose a preferred option. The scenarios describe the birth experience in terms of the seven LADY-X attributes. A D-efficient discrete choice experiment design with priors based on a small sample (N = 110) was applied. Two samples were gathered, women who had recently given birth and subjects from the general population. Both samples were analyzed separately using a panel mixed logit (MMNL) model. Using the panel mixed multinomial logit (MMNL) model results and accounting for preference heterogeneity, we calculated the average preference weights for LADY-X attribute levels. These were transformed to represent a utility score between 0 and 1, with 0 representing the worst and 1 representing the best birth experience.ResultsIn total, 1097 women who had recently given birth and 367 subjects from the general population participated. Greater value was placed on differences between bottom and middle attribute levels than on differences between middle and top levels. The attributes that resulted in larger utility increases than the other attributes were “feeling of safety” in the sample of women who had recently given birth and “feeling of safety” and “availability of professionals” in the general population sample.ConclusionsBy using the derived preference weights, LADY-X has the potential to be used as a utility measure for perinatal (cost-) effectiveness studies.     

Are Pain Intensity and Pain Related Fear Related to Functional Capacity Evaluation Performances of Patients with Chronic Low Back Pain?

Introduction: Pain related fear and pain intensity have been identified as factors negatively influencing Functional Capacity Evaluation (FCE) performances in patients with CLBP. Conflicting results have been reported in the literature. The objective of this study was to analyze the relationships between pain intensity and pain-related fear on the one hand, and performances during an FCE on the other hand in two samples of patients with chronic low back pain (CLBP). Methods: Two cross sectional observation studies were performed with two samples of patients with CLBP (study 1: n = 79; study 2: n = 58). Pain related fears were operationally defined as the score on the Tampa Scale of Kinesiophobia in study 1, and the Fear Avoidance Beliefs Questionnaire (FABQ) in study 2. Pain intensity was measured with a Numeric Rating Scale in both studies. Avoidance behavior observed during FCE was in both studies operationally defined as the unwillingness to engage in high intensity performance levels of three different functional activities: high intensity lifting, prolonged standing in a forward bend position, and fast repetitive bending at the waist. Results: A total of 25 correlations between pain and pain related fear, and performance variables were calculated, out of which 7 were significant (p < 0.05). The strength of these significant correlations ranged from r = −0.23 to r = −0.50. Multivariate linear regression analyses revealed non-significant relationships in most instances. Pain and pain related fear contributed little if any to these models. Conclusions: The relation between pain and pain related fear and FCE performance is weak or non-existent in patients with CLBP.     

What Characteristics of Nursing Homes Are Most Valued by Consumers? A Discrete Choice Experiment with Residents and Family Members

Objectives

To generate a scoring algorithm weighted on the preferences of consumers for assessing the quality of care in nursing homes (i.e., aged care homes or institutions) in six key domains.

Does Changing the Age of a Child to be Considered in 3-Level Version of EQ-5D-Y Discrete Choice Experiment–Based Valuation Studies Affect Health Preferences?

ObjectivesThere has been some debate about the choice of perspective and the age of the child considered when completing preference elicitation tasks in the 3-level version of EQ-5D-Y (EQ-5D-Y-3L) valuation protocol. This study aimed to clarify the impact on latent scale EQ-5D-Y-3L values of varying the age of the child experiencing the health state considered by respondents completing the discrete choice experiment (DCE) tasks of the protocol.MethodsWe conducted an online DCE with a representative sample of 1000 adults in the United Kingdom and 1000 adults in the United States. Respondents selected the health state they prefer from a series of DCE paired EQ-5D-Y-3L health state comparisons using their own perspective and that of a hypothetical child from the following age groups: “5-7 years old,” “8-10 years old,” “11-13 years old,” and “14-15 years old.” Data analysis was conducted using separate multinomial logit models for each perspective and country. We also estimated combined models including data from each possible pair of perspectives and used interactions between EQ-5D-Y-3L levels and perspective to determine whether any differences were statistically significant.ResultsNo statistically significant differences in coefficients between perspectives were found in the United States. In the United Kingdom, there were differences between the own perspective and the 5 to 7 years old perspective (looking after myself level 3) and between the 5 to 7 years old perspective and the 8 to 10 years old perspective (usual activities level 3).ConclusionsOur results suggest that there is minimal impact on latent scale values when using different ages of the hypothetical child in the current EQ-5D-Y-3L valuation protocol.