Advertisement

A Community-Engaged Approach to Quantifying Caregiver Preferences for the Benefits and Risks of Emerging Therapies for Duchenne Muscular Dystrophy

      Abstract

      Background

      There is growing agreement that regulators performing benefit–risk evaluations should take patients’ and caregivers’ preferences into consideration. The Patient-Focused Drug Development Initiative at the US Food and Drug Administration offers patients and caregivers an enhanced opportunity to contribute to regulatory processes by offering direct testimonials. This process may be advanced by providing scientific evidence regarding treatment preferences through engagement of a broad community of patients and caregivers.

      Objective

      In this article, we demonstrate a community-engaged approach to measure caregiver preferences for potential benefits and risks of emerging therapies for Duchenne muscular dystrophy (DMD).

      Methods

      An advocacy oversight team led the community-engaged study. Caregivers’ treatment preferences were measured by using best–worst scaling (BWS). Six relevant and understandable attributes describing potential benefits and risks of emerging DMD therapies were identified through engagement with advocates (n = 5), clinicians (n = 9), drug developers from pharmaceutical companies and academic centers (n = 11), and other stakeholders (n = 5). The attributes, each defined across 3 levels, included muscle function, life span, knowledge about the drug, nausea, risk of bleeds, and risk of arrhythmia. Cognitive interviewing with caregivers (n = 7) was used to refine terminology and assess acceptability of the BWS instrument. The study was implemented through an online survey of DMD caregivers, who were recruited in the United States through an advocacy group and snowball sampling. Caregivers were presented with 18 treatment profiles, identified via a main-effect orthogonal experimental design, in which the dependent variable was the respondents’ judgment as to the best and worst feature in each profile. Preference weights were estimated by calculating the relative number of times a feature was chosen as best and as worst, which were then used to estimate relative attribute importance.

      Results

      A total of 119 DMD caregivers completed the BWS instrument; they were predominately biological mothers (67.2%), married (89.9%), and white (91.6%). Treatment effect on muscle function was the most important among experimental attributes (28.7%), followed by risk of heart arrhythmia (22.4%) and risk of bleeding (21.2%). Having additional postapproval data was relatively the least important attribute (2.3%).

      Conclusions

      We present a model process for advocacy organizations aiming to promote patient-centered drug development. The community-engaged approach was successfully used to develop and implement a survey to measure caregiver preferences. Caregivers were willing to accept a serious risk when balanced with a noncurative treatment, even absent improvement in life span. These preferences should inform the Food and Drug Administration’s benefit–risk assessment of emerging DMD therapies. This study highlights the synergistic integration of traditional advocacy methods and scientific approach to quantify benefit–risk preferences.

      Key words

      Introduction

      Duchenne muscular dystrophy (DMD) is a rare, life-threatening, inherited neuromuscular disorder that occurs in male subjects with an incidence of 1.3 to 2.9 per 10,000.

      Centers for Disease Control and Prevention (CDC). Prevalence of Duchenne/Becker muscular dystrophy among males aged 5-24 years—four states, 2007. MMWR Morb Wkly Rep. 2009;58:1119–1122

      • Emery A.E.
      Population frequencies of inherited neuromuscular diseases—a world survey.
      Diagnosis usually occurs around age 5 years, when differences in motor function become apparent, but symptoms may appear as early as infancy.
      • Bushby K.
      • Finkel R.
      • Birnkrant D.J.
      • et al.
      Diagnosis and management of Duchenne muscular dystrophy, part 1: diagnosis, and pharmacological and psychosocial management.

      Eagle M, Baudouin SV, Chandler C, et al. Survival in Duchenne muscular dystrophy: improvements in life expectancy since 1967 and the impact of home nocturnal ventilation. Neuromuscul Disord. 2002;12:926–929.

      The condition is associated with significant care-related
      • Pangalila R.F.
      • van den Bos G.A.
      • Stam H.J.
      • et al.
      Subjective caregiver burden of parents of adults with Duchenne muscular dystrophy.
      • Kenneson A.
      • Bobo J.K.
      The effect of caregiving on women in families with Duchenne/Becker muscular dystrophy.
      and financial burden.
      • Larkindale J.
      • Yang W.
      • Hogan P.F.
      • et al.
      Cost of illness for neuromuscular diseases in the US.
      • Ouyang L.
      • Grosse S.D.
      • Fox M.H.
      • Bolen J.
      A national profile of health care and family impacts of children with muscular dystrophy and special health care needs in the United States.
      • Ouyang L.
      • Grosse S.D.
      • Kenneson A.
      Health care utilization and expenditures for children and young adults with muscular dystrophy in a privately insured population.
      Affected individuals have progressive muscular weakness, loss of ambulation that typically occurs in the teen years, and premature death.
      • Bushby K.
      • Finkel R.
      • Birnkrant D.J.
      • et al.
      Diagnosis and management of Duchenne muscular dystrophy, part 1: diagnosis, and pharmacological and psychosocial management.
      The mean age of death is in the 20s and is commonly caused by respiratory failure or cardiac disease.
      • Bushby K.
      • Finkel R.
      • Birnkrant D.J.
      • et al.
      Diagnosis and management of Duchenne muscular dystrophy, part 1: diagnosis, and pharmacological and psychosocial management.

      Eagle M, Baudouin SV, Chandler C, et al. Survival in Duchenne muscular dystrophy: improvements in life expectancy since 1967 and the impact of home nocturnal ventilation. Neuromuscul Disord. 2002;12:926–929.

      Currently, there are no therapies approved by the US Food and Drug Administration (FDA) for DMD. The standard-of-care treatment is the off-label use of corticosteroids, which have been shown to stabilize muscle strength, delay loss of ambulation by 2 to 5 years, improve cardiopulmonary function, and enhance quality of life.
      • Bushby K.
      • Finkel R.
      • Birnkrant D.J.
      • et al.
      Diagnosis and management of Duchenne muscular dystrophy, part 1: diagnosis, and pharmacological and psychosocial management.
      • Biggar W.D.
      • Harris V.A.
      • Eliasoph L.
      • Alman B.
      Long-term benefits of Deflazacort treatment for boys with Duchenne muscular dystrophy in their second decade.
      • Angelini C.
      The role of corticosteroids in muscular dystrophy: a critical appraisal.
      • Bushby K.
      • Muntoni F.
      • Urtizberea A.
      • et al.
      Report on the 124th eNMC International workshop. Treatment of Duchenne muscular dystrophy; defining the gold standards of management in the use of corticosteroids.
      Several potential therapies are under clinical trial that target a variety of primary and secondary effect pathways.
      • Govoni A.
      • Magri F.
      • Brajkovic S.
      • et al.
      Ongoing therapeutic trials and outcome measures for Duchenne muscular dystrophy.
      Similar to other conditions (including other rare diseases
      • Milne C.P.
      Prospects for rapid advances in the development of new medicines for special medical needs.
      and early-on in the HIV epidemic),
      • Michaels M.
      • Collyar D.
      Research advocacy in traditional research settings: questions of influence and legitimacy.
      patients and caregivers managing DMD seek to accelerate approvals for drugs that may save lives.

      Parent Project Muscular Dystrophy. Benefit-risk assessments in rare disorders: The case for therapeutic development in Duchenne muscular dystrophy as the prototype for new approaches. http://www.parentprojectmd.org/BR. Accessed April 28, 2014

      In the context of serious, rare disorders with limited treatment options, patients and patient advocates want regulators to be more permissive.

      Genetic Alliance UK. New medicines for serious conditions: weighing the risks and benefits. The verdict of a jury of patients. http://www.geneticalliance.org.uk/docs/citizens-jury-report.pdf. Accessed March 1, 2014.

      Drugs under trial for DMD represent a significant opportunity for families to intervene, and in public forums, some parent advocates have demanded access to drugs, even absent conclusive data on efficacy and safety.

      Parent Project Muscular Dystrophy: Share Your Brave Story. http://community.parentprojectmd.org/profiles/blog/list?tag=My+BRAVE+Story. Accessed March 1, 2014.

      DMD provides an appealing model for assessing influences on treatment decision making for serious, progressive disorders. The natural history may lead to high-pressure decisions regarding the use of novel therapies. The pediatric onset provides additional complexity, as the majority of treatment decisions are made by parents/guardians who are also the primary caregivers. These decisions may later be re-evaluated by adolescent and adult DMD patients who could have different treatment preferences.
      • McNeil D.E.
      • Davis C.
      • Jillapalli D.
      • et al.
      Duchenne muscular dystrophy: drug development and regulatory considerations.
      There is growing agreement that regulatory benefit–risk evaluations should be informed by the perspectives of patients and caregivers who will ultimately make treatment decisions and bear the associated risks.
      • Hauber B.
      • Fairchild A.O.
      • Johnson F.R.
      Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
      To that end, in 2012, the FDA was congressionally mandated to engage patients to understand the impact of disease though the Patient-Focused Drug Development Initiative.

      US Food and Drug Administration (FDA). Enhancing benefit-risk assessment in regulatory decision-making. http://www.fda.gov/forindustry/userfees/prescriptiondruguserfee/ucm326192.htm. Accessed March 1, 2014.

      Although this program offers patients and caregivers an unprecedented opportunity to contribute to the regulatory process, the program is limited in scope and approach, with initially only 20 disease areas being targeted for public comment involving direct engagement with patients and caregivers.
      • Mullard A.
      Patient-focused drug development programme takes first steps.
      Advocates who do not represent 1 of the 20 chosen disorders are left with little guidance about how to provide input that is acceptable and useful to the FDA. Existing models for FDA engagement are largely limited to providing testimonial. Although such direct engagement is a primary strength and a mainstay of advocacy organizations’ efforts to inform decision makers about their community’s needs and perspectives,
      • Facey K.
      • Boivin A.
      • Gracia J.
      • et al.
      Patients’ perspectives in health technology assessment: a route to robust evidence and fair deliberation.
      there are limitations to focusing only on patient and caregiver testimonials, such as questions about how well those providing testimonial represent the views of the entire patient population.
      • Hauber B.
      • Fairchild A.O.
      • Johnson F.R.
      Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
      • Hansen H.P.
      • Draborg E.
      • Kristensen F.B.
      Exploring qualitative research synthesis: the role of patients’ perspectives in health policy design and decision making.
      • Vogt F.
      • Schwappach D.
      • Bridges J.
      Accounting for tastes: a German perspective on the inclusion of patient preferences in health care.
      Increasingly, decision makers are being asked to consider alternative methods to quantify treatment preferences and risk tolerance that take into account the views of large groups of stakeholders.
      • Coplan P.
      • Noel R.
      • Levitan B.
      • et al.
      Development of a framework for enhancing the transparency, reproducibility and communication of the benefit–risk balance of medicines.
      Draft FDA guidance has indicated a willingness to incorporate such evidence into the regulatory process.
      US Food and Drug Administration (FDA)
      Factors to Consider When Making Benefit-Risk Determinations in Medical Device Premarket Approval and De Novo Classifications. FDA Guidance for Industry and Food and Drug Administration Staff. Silver Spring, Md: Center for Devices and Radiological Health, Center for Biologics Evaluation and Research.
      Quantitative preference elicitation methods allow stakeholders to introduce formal evidence-based decision making into the regulatory process and have been used to explore decision making and preferences among a variety of patient populations.
      • Hauber B.
      • Fairchild A.O.
      • Johnson F.R.
      Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
      The purpose of the present study was to demonstrate a process by which a patient advocacy organization might develop scientific evidence on treatment preferences. We aimed to model a replicable, community-engaged approach to exploring preferences in a large sample of decision makers. Specifically, our goal was to explore caregiver preferences for emerging treatments for DMD. This study is not only informative to those seeking to understand the treatment preferences and risk tolerance of DMD caregivers, but it serves to highlight principles of patient-centered outcomes research
      • Fleurence R.
      • Selby J.V.
      • Odom-Walker K.
      • et al.
      How the Patient-Centered Outcomes Research Institute is engaging patients and others in shaping its research agenda.
      by illustrating how an advocacy organization can take leadership in generating policy-relevant evidence.

      Materials and Methods

      Parent Project Muscular Dystrophy (PPMD), an advocacy organization focused on finding a cure for DMD, led the study. The advocacy oversight team comprising PPMD staff members (a clinician, a scientist experienced in drug development, and 2 caregivers of individuals with DMD) collaborated with the research team to design and implement the study. The oversight team made study decisions through a consensus process. Consistent with the preferences of PPMD, the authoring research team was a smaller team of PPMD staff and academic collaborators.
      The teams began by defining a research question about treatment preferences based on the stated needs of the Duchenne community; they then identified the study population (caregivers of individuals with DMD) and a recruitment strategy. The teams choose a stated preference method (described under the heading “Methods”) that fit the study needs. In the development of the treatment experiment (described under the heading “Identifying Attributes and Levels”), the team used a community-engaged approach involving multidisciplinary stakeholder informants. The survey was piloted by a small group of end-users (described under the heading “Survey Pilot: Cognitive Interview”), and it was modified based on their input. Preliminary and final analyses were reported to regulators, industry, and the Duchenne community in an accessible and timely manner.

       Methods

      Methods to measure the preferences of patients, caregivers, and other stakeholders are now well established
      • Bridges J.
      • Hauber A.B.
      • Marshall D.
      • et al.
      A checklist for conjoint analysis applications in health: report of the ISPOR Conjoint Analysis Good Research Practices Taskforce.

      Marshall D, Bridges J, Hauber AB, et al. Conjoint analysis applications in health—how are studies being designed and reported? An update on current practice in the published literature between 2005 and 2008. Patient. 2010;3:249–256.

      and are increasingly being applied to study benefit–risk preferences.
      • Hauber B.
      • Fairchild A.O.
      • Johnson F.R.
      Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
      Although good research practices have been created to aid in the development of stated preference applications in medicine,
      • Bridges J.
      • Kinter E.
      • Schmeding A.
      • et al.
      Can patients diagnosed with schizophrenia complete choice-based conjoint analysis tasks.
      approaches such as conjoint analysis and discrete choice experiments remain complex. They require qualitative skills to appropriately identify attributes and levels and develop supporting survey text,
      • Coast J.
      • Flynn T.N.
      • Salisbury C.
      • et al.
      Maximising responses to discrete choice experiments: a randomised trial.
      as well as quantitative skills to design the experiment
      • Reed Johnson F.
      • Lancsar E.
      • Marshall D.
      • et al.
      Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force.
      and analyze results.

      International Society for Pharmacoeconomics and Outcomes Research. ISPOR Statistical Methods for Conjoint Analysis Task Force. http://www.ispor.org/TaskForces/Analyzing_Data_Conjoint_Analysis.asp. Accessed March 1, 2014.

      The study used best–worst scaling (BWS) case 2, an emerging stated-preference method that can be used to scientifically assess preferences.
      • Erdem S.
      • Rigby D.
      • Wossink A.
      Using best-worst scaling to explore perceptions of relative responsibility for ensuring food safety.
      Referred to as the “profile case,” BWS case 2 presents profiles one at a time to elicit preferences; Flynn et al
      • Flynn T.N.
      • Louvier J.J.
      • Peters T.J.
      • Coast J.
      Best-worst scaling: what it can do for health care research and how to do it.
      fully described the method and provide use guidance. BWS has been recognized as an approach that is easier to design and analyze than conjoint analysis and discrete choice; however, this method is relatively novel in the context of measuring benefit–risk preferences.
      • Hauber B.
      • Fairchild A.O.
      • Johnson F.R.
      Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
      The study reported here presents a novel use of the BWS case 2.
      • Bridges J.
      Stated preference methods in health care evaluation: an emerging methodological paradigm in health economics.
      • Flynn T.N.
      Valuing citizen and patient preferences in health: recent developments in three types of best-worst scaling.
      BWS is thought to be less cognitively demanding on participants than discrete choice or conjoint experiments.
      • Flynn T.N.
      • Louvier J.J.
      • Peters T.J.
      • Coast J.
      Best-worst scaling: what it can do for health care research and how to do it.
      In addition, relevant to our aim of demonstrating a replicable community-engaged model, BWS benefits from a straightforward analytic approach, the results of which are consistent with more complex approaches
      • Louviere J.J.
      • Flynn T.N.
      Using best-worst scaling choice experiments to measure public perceptions and preferences for healthcare reform in Australia.
      • Gallego G.
      • Bridges J.
      • Flynn T.
      • et al.
      Using best-worst scaling in horizon scanning for hepatocellular carcinoma technologies.
      that may be unfamiliar to many researchers in the clinical domain. To guide the development, implementation, and analysis of our preference elicitation instrument, we used the standards outlined in the International Society for Pharmacoeconomics and Outcomes Research checklist for conjoint analysis
      US Food and Drug Administration (FDA)
      Factors to Consider When Making Benefit-Risk Determinations in Medical Device Premarket Approval and De Novo Classifications. FDA Guidance for Industry and Food and Drug Administration Staff. Silver Spring, Md: Center for Devices and Radiological Health, Center for Biologics Evaluation and Research.
      and specific guidance on the use of BWS.

      International Society for Pharmacoeconomics and Outcomes Research. ISPOR Statistical Methods for Conjoint Analysis Task Force. http://www.ispor.org/TaskForces/Analyzing_Data_Conjoint_Analysis.asp. Accessed March 1, 2014.

       Identifying Attributes and Levels

      BWS case 2 experiments use attributes (representing topic areas) and levels (representing attribute variables, such as amount of or impact on the attributes); Figure 1 presents an example of a BWS case 2 task. Identification of relevant and comprehensible attributes and levels is required for a meaningful study outcome. We used a stakeholder-informed approach to identify attributes and levels that were clinically relevant, meaningful, and understandable to caregivers. The development of attributes and levels was informed by PPMD’s 20 years of experience with patients and families; extensive history consulting on, reviewing, and funding clinical research; and an ongoing interview study of clinical trial experiences.

      Peay HL, Sheffer H, Tibben A. Expectations and decision making in clinical trials for Duchenne and Becker muscular dystrophy. Abstract presented at: 18th International Congress of The World Muscle Society; October 1–5, 2013; Asilomar, Calif.

      Figure thumbnail gr1
      Figure 1Example of best–worst scaling task.
      PPMD identified and invited stakeholder informants to participate (October–December 2012) through an existing advocacy-facilitated industry roundtable, PPMD’s grassroots family networks, PPMD’s clinician database, or after self-nomination following community notification of the program launch. Stakeholder informants, including patient/disease advocates (n = 5), clinicians (n = 9), drug developers from pharmaceutical companies and academic centers (n = 11), and other stakeholders (n = 5), participated in group or individual sessions. Attributes and levels for the emerging therapies were proposed and refined through iterative rounds of stakeholder engagement. Industry informants were important to successfully identify appropriate attributes because of their ability to forecast benefits and risks of premarket drugs. The oversight committee incorporated their informative input while protecting against potential bias.
      This approach yielded >20 potential benefit, adverse effect, and risk attributes. Items were grouped under themes and refined. Several attributes were rejected by the oversight team for relevance, similarity to other items, or concerns about the ability of the target population to understand the attribute. Examples of attributes that were not chosen are the ability to participate in day-to-day family activities and risks to renal and hepatic function. The participation benefit was considered to be less concrete and treatment-associated than muscle function, and thus was not chosen. The clinical implications of renal and hepatic damage were difficult to describe in a brief and accessible format. In addition, we chose not to use quantitative risks in this survey because the average US adult has only a basic level of quantitative literacy.

      National Center for Education Statistics. National assessment of adult literacy (NAAL). A first look at the literacy of America’s adults in the 21st century. https://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006470. Accessed March 1, 2014.

      Through a consensus process, the advocacy oversight team ultimately selected 6 treatment attributes with 3 levels each. The attributes and associated levels were chosen to be reasonable based on drugs under trial, with the notable exception that the highest risk levels represent considerably more risk than has been associated with drugs under trial, to date. The proposed attribute list was again shared with stakeholder informants. Based on their input, the final items (effect on muscle function, life span, knowledge about the drug, nausea, risk of bleeding, and risk of heart arrhythmia) were chosen by the oversight group and study team.
      The BWS experiment attributes and levels are presented in Table I. The functional benefits chosen were “stops the progression of weakness” and “slows the progression of weakness” because drugs under trial are unlikely to result in a cure or significant improvement in strength for patients with DMD. The life span attribute was presented independently of the weakness attribute because drugs that affect skeletal muscle may not improve cardiac outcomes in DMD
      • Townsend D.
      • Yasuda S.
      • Li S.
      • et al.
      Emergent dilated cardiomyopathy caused by targeted repair of dystrophic skeletal muscle.
      and thus may not improve longevity. Caregiver participants were prompted to separate muscle function from life span by use of this cardiac example and a sample task. Given that the quality of evidence and associated uncertainty may affect preferences, we included an attribute relating to knowledge about the drug, described as the number of years of postapproval data. Nausea represents a realistic, easily understood adverse effect that may result in increasing burden as patients lose mobility. The choice of bleeding was prompted by a Phase II DMD trial that was terminated in 2011 (unpublished data). Arrhythmia was chosen as an attribute that is salient to caregivers because it is part of the DMD natural history.
      • Shirokova N.
      • Niggli E.
      Cardiac phenotype of Duchenne muscular dystrophy: insights from cellular studies.
      Table IAttribute, levels, and descriptions resulting from stakeholder engagement.
      Attributes and Attribute LvelsAdditional Description/Explanation in the Survey
      Effect on muscle function“How the treatment affects muscle function”
      iStops the progression of weakness“Most people who take this treatment don’t get any weaker over time.
      iiSlows the progression of weakness“Most people who take this treatment continue to get weaker over time, but more slowly than they would without treatment.”
      iiiDoes not change progression of weakness
      Lifespan“By this we mean how many extra years of life are expected because of the treatment.”
      i5 year gain in expected lifespan“5 extra years of life”
      ii2 year gain in expected lifespan“2 extra years of life”
      iiiNo extra gain to expected lifespan“0 extra years, meaning that the treatment may not change the person’s lifespan at all.”
      Knowledge about the drug“Important information about treatments comes after FDA approval, from tracking people who take the treatment over time. Tracking helps us better understand benefits, risks, potential drug interactions, and how the treatment affects people of different ages and stages of progression. Imagine that everyone who takes the treatment is tracked.”
      i2 years of post-approval drug information available“The treatment has been on the market for 2 years and we have data from 900 people with Duchenne.”
      ii1 year of post-approval drug information available“The treatment has been on the market for one year and we have data from 200 people with Duchenne.”
      iiiNo post-approval drug information available“The treatment has just been approved and no post-approval data is available.”
      Nausea
      iNo increased chance of nausea
      iiCauses loss of appetite“A person taking the treatment loses his/her appetite”
      iiiCauses loss of appetite with occasional vomiting“A person taking the treatment loses his/her appetite and has occasional vomiting”
      Risk of bleeds“Risk of bleeding”
      iNo increased risk of bleeds
      iiIncreased risk of bleeding gums and increased bruising“Bleeding gums and increased bruising, without increased risk of more dangerous bleeding”
      iiiIncreased risk of hemorrhagic stroke and lifelong disability“Hemorrhagic (bleeding) stroke, which could lead to lifelong disability in memory and reasoning. People found to have this risk would have to stop taking the treatment.”
      Risk of heart arrhythmia“Risk of heart rhythm problems”
      iNo increased risk of heart arrhythmia
      iiIncreased risk of harmless heart arrhythmia“Occasional, harmless heart arrhythmia”
      iiiIncreased risk of dangerous heart arrhythmia and sudden death“Dangerous arrhythmia, which could lead to surgery to put in a defibrillator and risk of sudden death. People found to have this risk would have to stop taking the treatment.”

       Experimental Design

      Following good research practices, we developed the BWS case 2 experiment to accommodate the 6 attributes with 3 levels each.
      • Bridges J.
      • Kinter E.
      • Schmeding A.
      • et al.
      Can patients diagnosed with schizophrenia complete choice-based conjoint analysis tasks.
      We applied a 3^6 main-effects orthogonal design, identified from the SAS database of orthogonal arrays.

      Kuhfeld W. Orthogonal arrays [TS-723]. Cary (NC): SAS http://support.sas.com/techsup/technote/ts723.html. Accessed March 1, 2014

      This array consisted of 18 full-profile combinations of the attributes and levels, the minimum such number necessary to ensure no structural relationships (ie, correlations) between the attributes.
      • Reed Johnson F.
      • Lancsar E.
      • Marshall D.
      • et al.
      Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force.
      As illustrated in Figure 1, in each BWS task, caregivers were presented with one of the treatment profiles and asked to judge which aspects they thought were the best and the worst. Before completing the tasks, caregivers were presented with a detailed description of all the attributes and levels to be considered in the task (including warm-up questions where appropriate), detailed instructions, and an explained example task. Furthermore, we confirmed that these “treatments do not currently exist” and that we were “interested in knowing what [the caregiver] would choose if they did.” As a matter of context, we informed the caregivers that “we are imagining that these are approved treatments provided by the doctor, and not treatments given during a clinical trial” and asked them to “assume that all your child’s medical bills, including the costs of the treatment, are covered by health insurance.” We also assured respondents that this “was not a test” and that there were “no right or wrong answers.” Each task incorporated a full profile (ie, all 6 attributes were shown) consisting of a specific level for each attribute. Preferences were elicited via caregivers making a judgment as to what aspect constituted the best and then the worst of the treatment.

       Survey Instrument

      The BWS instrument was included in a broader survey. In addition to basic demographic questions, to ensure that the study sample did not represent individuals with unusually high risk-taking personality traits, the participant section included the 6-item risk-taking measure comprising items from the Jackson Personality Inventory.
      • Pearson S.D.
      • Goldman L.
      • Orav E.J.
      • et al.
      Triage decisions for emergency department patients with chest pain: do physicians’ risk attitudes make the difference?.
      The previously published mean (SD) score for a physician group was 19 (4), with a range of 11 to 30. We also sought to describe caregivers’ numeracy by using the 3-item short form (SNS-3) of the Subjective Numeracy Scale (SNS).

      Wallston KA, McNaughton C, Storrow A, et al. Validation of a short, 3-item version of the Subjective Numeracy Scale (SNS-3). Presented at: Health Literacy Research Conference; October 17-18, 2011; Chicago, IL

      Poor numeracy has been associated with ability to evaluate risks and benefits of health options
      • Peters E.
      • Hibbard J.
      • Slovic P.
      • Dieckmann N.
      Numeracy skill and the communication, comprehension, and use of risk-benefit information.
      and to negatively affect utility assessment.
      • Zikmund-Fisher B.
      • Smith D.M.
      • Ubel P.A.
      • Fagerlin A.
      Validation of the Subjective Numeracy Scale: effects of low numeracy on comprehension of risk communications and utility elicitations.
      Previous research has determined a norm mean score for the SNS of 4.03 (1.04), with a range of 1 to 6.
      • Fagerlin A.
      • Zikmund-Fisher B.J.
      • Ubel P.A.
      • et al.
      Measuring numeracy without a math test: development of the Subjective Numeracy Scale.
      Information collected about the affected children included the number of children with DMD, age, where the children lived, a mobility item (with 11 responses ranging from independent walking outdoors to remaining in bed) that represents disease progression, insurance type (private vs government), and whether the children ever experienced a life-threatening emergency.

       Survey Pilot: Cognitive Interviewing

      Cognitive interviews with 7 parents of individuals with DMD of varying ages and disease stages were used to assess comprehension, refine terminology, and explore the acceptability of the instrument. The investigator (H.L.P.) observed the survey experience of the interviewees by using videoconferencing. Interviewees were asked to “think aloud” as they completed the survey, and the investigator used verbal probes to explore anticipated problem areas, assess understanding of the survey elements, and evaluate willingness to trade among the attributes and levels. The instrument was then modified to the final version based on pilot test feedback, again using a consensus process involving the study team and the advisory group.

       Participants and Recruitment

      Participants were caregivers (parents or guardians) of at least 1 living child with DMD. The caregivers lived in the United States, were at least 18 years of age, and were able to complete an online survey in English. Their affected child could be any age or at any stage of the disease. The survey was administered online by using the Qualtrics survey system (Qualtrics, LLC, Provo, Utah) from January 17, 2013, to February 21, 2013. In the study advertisements, PPMD committed to sharing the information learned from the survey back to the DMD community. Recruitment occurred with the use of newsletter notices, social media, recruitment e-mails from PPMD, the DuchenneConnect self-report registry, and through word-of-mouth recruiting. The anonymous survey was determined to be exempt by the Western Institutional Review Board (no. 1-756840-1).

       Statistical Analysis

      In BWS, the dependent variable is the participants’ judgment about the best and worst feature in each profile presented to them.
      • Molassiotis A.
      • Emsley R.
      • Ashcroft D.
      • et al.
      Applying best-worst scaling methodology to establish delivery preferences of a symptom supportive care intervention in patients with lung cancer.
      Although the results from a BWS can be estimated by using complicated techniques such as conditional logit
      • McFadden D.
      Conditional logit analysis of qualitative choice behavior.
      or hierarchical Bayes,
      • Orme B.K.
      Getting Started With Conjoint Analysis: Strategies for Product Design and Pricing Research.
      one of the benefits of the method is that it can be analyzed very simply. The simplest techniques focus on the number of times a particular level of an attribute was chosen as best and as worst when it was available in the choice task (unpublished data). A relative best-minus-worst (BW) score can be calculated by subtracting the number of times a feature was chosen as worst from the number of times it was chosen as best, then dividing by the total number of times it was available to be chosen. Early applications of this method have demonstrated a very high level of correlation between such simple techniques and more complicated regression-based techniques.
      • Louviere J.J.
      • Flynn T.N.
      Using best-worst scaling choice experiments to measure public perceptions and preferences for healthcare reform in Australia.
      • Gallego G.
      • Bridges J.
      • Flynn T.
      • et al.
      Using best-worst scaling in horizon scanning for hepatocellular carcinoma technologies.
      As with all techniques used to estimate ordinal, multinomial outcomes, our scoring approach assumes equal spacing between things that were chosen as best (BW score, 1), those that were not chosen (BW score, 0), and those chosen as worst (BW score, –1).
      We chose to estimate the importance weights for each level by using the relative BW score because it could be easily understood by the broadest readership (including the community of patients and caregivers that we engaged in developing this survey). In addition to the simplicity of the BW score, this approach has several advantages. First, regression-based techniques require the use of the omitted category
      • Louviere J.J.
      • Flynn T.N.
      Using best-worst scaling choice experiments to measure public perceptions and preferences for healthcare reform in Australia.
      or the use of complex effects coding procedures
      • Gallego G.
      • Bridges J.
      • Flynn T.
      • et al.
      Using best-worst scaling in horizon scanning for hepatocellular carcinoma technologies.
      to estimate choice models. Second, by using this simple approach, we have ensured that all estimated parameters remain on the same ratio scale. This method allows comparisons to be made across the attributes, as well as identification of global best and worst attribute levels across all the attributes. Because the BW score is estimated as a mean across the sample, we also report the SEs for these means. This process allows us to conduct t tests to determine whether the scores were significantly different from zero. We did this for each attribute level and have reported the P values for each test.
      Finally, we used the relative BW score for each level within each attribute to assess the overall importance of each attribute, conditioned on the levels chosen.
      • Bridges J.
      • Mohamed A.F.
      • Finnern H.W.
      • et al.
      Patients’ preferences for treatment outcomes for advanced non-small cell lung cancer: a conjoint analysis.
      With this technique, relative attribute importance was estimated by subtracting the lowest relative BW score associated with a level of that attribute from the highest relative BW score associated with a level of that attribute. We then divided each difference by the sum of all differences across the 6 attributes and reported the result as a percentage.

      Results

      A total of 124 caregivers who self-identified as being a parent or guardian of an individual with DMD began the treatment experiment. Two individuals dropped out after the first treatment task; 1 dropped out after the third treatment task; 1 dropped out after the fifth treatment task; and 1 dropped out after the 15th treatment task. The remaining 119 caregivers completed the entire survey.
      Table II summarizes the characteristics of the sample. Caregivers were predominantly white, married, biological mothers, and had 1 affected child. Education level ranged from high-school or General Educational Development diploma to graduate or professional degree; the median response was "4-year college degree.” Annual household income ranged from “<$25,000” to “>$100,000”; the median response was “$75,000 to $100,000.” Caregivers’ ages ranged from 28 to 66 years (mean, 43.7 years), and the age of the affected children ranged from 2 to 38 years (mean, 12.1 years). Slightly more than one half of the children were reported as having participated in clinical research and more than one third in a clinical trial. Almost all of the affected children lived in the home of the caregiver (n = 117 [98%]). The majority of the affected children used private insurance for their medical care (n = 101 [85%]), although 34% (n = 40) endorsed that their child used a state/government program. Caregivers reported that 19% (n = 22) of their children had experienced a life-threatening emergency.
      Table IICharacteristics of participants and affected children (N=119)
      CharacteristicValue
      Participant characteristics
       Mean (SD) caregiver age, y43.7 (7.7)
       Mean (SD) child age, y21.1 (6.4)
      Caregiver characteristics
       Relationship to children
        Biological mother67.2%
        Biological father28.6%
        Adoptive mother3.4%
        Adoptive father0.8%
       Marital status
        Married/long-term relationship89.9%
        Divorced/separated9.2%
        Widowed0.8%
       Race
        White91.6%
       Education
        High school/GED4.2%
        Some college14.3%
        Technical school5.0%
        Associated degree7.6%
        4-year college degree42.9%
        Graduate/professional degree25.2%
       Income
        <$25,0005.9%
        $25,000–$50,0008.4%
        $50,000–$75,00018.5%
        $75,000–$ 100,00018.5%
        >$100,00047.1%
       Child characteristics
       No. of affected children
        1 child92.4%
        ≥2 children7.6%
       Living arrangements
        In caregiver’s home98.3%
        Independent0.8%
        Other0.8%
       Ambulation status
        Ambulatory
      Ability to walk independently outside for at least short distances.
      63.9%
        Nonambulatory36.0%
       Research participation
        Clinical research58.0%
        Clinical trial34.0%
       Had life-threatening emergency
        Yes18.5%
        No81.5%
      Note: In some cases, percents do not add to 100% because of missing values.
      GED = General Educational Development diploma.
      low asterisk Ability to walk independently outside for at least short distances.
      The affected children represented a range of disease progression. When these children were dichotomized into an “ambulatory” group, defined as those who could walk independently outdoors for at least short distances, and a “nonambulatory” group, defined as those who could not walk outdoors without help, 75 (64%) were parents with children in the ambulatory group and 43 (36%) were parents with children in the nonambulatory group.
      The mean (SD) risk-taking score was 17 (4), with a range of 7 to 30 (higher scores indicate more risk-taking endorsement). The caregivers in this study scored significantly lower on the risk-taking score than the physician reference group

      Peay HL, Sheffer H, Tibben A. Expectations and decision making in clinical trials for Duchenne and Becker muscular dystrophy. Abstract presented at: 18th International Congress of The World Muscle Society; October 1–5, 2013; Asilomar, Calif.

      (mean, 19 [4]; P < 0.005), indicating lower risk-taking personality traits. The mean SNS-3 score was 4.90 (1.1), with a range of 2.33 to 4.87, which was higher than the reference population.

      National Center for Education Statistics. National assessment of adult literacy (NAAL). A first look at the literacy of America’s adults in the 21st century. https://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006470. Accessed March 1, 2014.

      This result was consistent with the high educational levels reported in our study (68% with at least a college degree).
      Table III presents relative BW scores, SEs, and P values, and Figure 2 diagrams the relative utility of each level as measured by using relative BW scores. All of the BW scores were significant at P < 0.001 except for “no increased chance of nausea,” “no postapproval drug information available,” and “no increased risk of bleeding.” By a large margin, the highest utilities as measured by using relative BW scores were for “stops progression of weakness” (0.877) and “slows progression of weakness” (0.800). These scores had almost twice the utility of the next-highest score, “5-year gain in expected life span” (0.464). The “2-year gain in expected life span” had a similar priority to the “5-year gain in expected life span” (0.408).
      Table IIIBest-worst results.
      Attribute descriptionBestWorstBest-worstRelative b-w Score5.E.T-testP-Value
      Effect on muscle function
       Stops the progression of weakness62826268770.01369.441<0.001
       Slows the progression of weakness57105710.8000.01553.357<0.001
       Does not change progression of weakness68125−57−0.0800.019−4.149<0.001
      Lifespan
       5 veargain in expected lifespan348173310.4640.02022.741<0.001
      2 veargain in expected lifespan29932910.4080.01921.186<0.001
       No extra gain to expected lifespan1293−81−0.1130.014−8.269<0.001
      Knowledge about the drug
      2 years of post-approval drug info available10969400.0560.0193.0150.001
       1 year of post-approval drug info available204160.0220.0073.2880.001
       No post-approval drug info available4156−IS−0.0210.014−1.5240.064
      Nausea
       No increased chance of nausea1926−7−0.0100.009−1.0440.148
       Causes loss of appetite195−94−0.1320.013−10.272<0.001
       Causes loss of appetite with occasional vomiting17217−200−0.2800.019−14.981< 0.001
      Risk of bleeds
       No increased risk of bleeds311-8−0.0110.005−2.1430.016
       Increased risk of bleedinggums and increased bruising0190−190−0.2660.017−16.079<0.001
       Increased risk of hemorrhagic stroke and lifelong disability0514−514−0.7200.017−42.807< 0.001
      Risk of heart arrhythmia
       No increased risk of heart arrhythmia532−27−0.0380.008−4.498<0.001
       Increased risk of harmless heart arrhythmia1122−121−0.1690.014−11.943<0.001
       Increased risk of dangerous arrhythmia and sudden death0551−561−0.7860.015−51.131<0.001
      Figure thumbnail gr2
      Figure 2Relative best–worst scores for attribute levels. Mod = moderate; Sev = severe.
      Caregivers attributed the highest negative BW scores to “increased risk of dangerous heart arrhythmia and sudden death” (–0.786), followed by “increased risk of hemorrhagic stroke and lifelong disability” (–0.720). This was followed by “causes loss of appetite with occasional vomiting” (–0.280). Although the 2 most serious risks had high negative scores, either (but not both) could be offset by a treatment that stopped the progression of weakness. The amount of knowledge about the drug was not given high relative BW scores at any level, with mean scores ranging from 0.056 to –0.021.
      Table IV includes the relative attribute importance for the entire group of caregivers. At the attribute level, effects on muscle function accounted for the largest proportion of the variance (28.7%), followed by arrhythmia (22.4%), bleeding (21.2%), life span (17.3%), nausea (8.1%), and knowledge about the drug (2.3%).
      Table IVRelative attribute importance.
      AttributeMaximumMinimumDifferencePercent
      Percent relative importance calculated as the difference between the maximum and minimum utility for each attribute divided by the sum of all such differences.
      Effect on muscle function0.877–0.0800.95728.7
      Life span0.464–0.1130.57717.3
      Knowledge about the drug0.056–0.0210.0772.3
      Nausea–0.010–0.2800.2708.1
      Risk of bleeding–0.011–0.7200.70921.2
      Risk of heart arrhythmia–0.038–0.7860.74822.4
      Sum3.338100
      low asterisk Percent relative importance calculated as the difference between the maximum and minimum utility for each attribute divided by the sum of all such differences.

      Discussion

      Although the FDA is committed to patient-centered drug development, the agency has limited resources. Representing a disease community that was not selected for the congressionally mandated community engagement program, PPMD led a study to proactively inform the FDA’s benefit–risk assessments. The process we used can be a model for facilitating patient-centered drug development through an exploration of the priorities and preferences of patients, families, and other stakeholders. Although the individual stories of highly motivated advocates are powerful and influential, it is difficult to know whether these testimonials represent the perspectives of the majority of patients and families. We describe a successful community-engaged process to understand treatment preferences in a large group of decision makers supported by the use of best/worst scaling. To the best of our knowledge, this study represents the first time a patient advocacy organization has led a quantitative preferences study of this complexity, highlighting a successful advocacy/academic collaboration that integrates traditional advocacy methods, family-centered outcomes research, and a scientific approach to quantifying preferences.
      Within the context of our experiment, caregivers attributed very high scores to stopping or slowing the progression of muscle weakness. Change in life span was not scored as highly. Feedback during cognitive interviewing suggested that parents associated better muscle function with higher quality of life, indicating that parents value quality more than length of life. This finding is consistent with both anecdotal reports and an interview study of parents of children involved in clinical trials,

      Peay HL, Sheffer H, Tibben A. Expectations and decision making in clinical trials for Duchenne and Becker muscular dystrophy. Abstract presented at: 18th International Congress of The World Muscle Society; October 1–5, 2013; Asilomar, Calif.

      in which parents expressed a preference for better quality of life for their child over a longer life span.
      We found that the presence of a serious risk could be compensated for by a treatment that stops or slows progression of weakness, even absent any other benefits. The burden of DMD may be associated with parents’ willingness to accept a serious risk for a noncurative treatment. The data support a limit to parents’ risk tolerance, however: for the levels of benefit provided in the experiment, they would not accept a treatment with 2 serious risks.
      Our community-engaged process contributed to successful recruitment of sufficient numbers of caregivers for a complex, time-intensive survey in only 5 weeks, notwithstanding the fact that the study focuses on a rare disease. The caregivers’ children represented a range of ages and disease stages, and thus our outcomes reflect the preferences of parents with children across the disease course. Although the development of an appropriate experimental design is a complex task, it is one that is well suited to be led by advocacy organizations with expert input and collaboration.

       Limitations

      There are several limitations to the study. First, the study sample, although likely to be representative of caregivers whose children are enrolled in clinical trials, may not be generalizable to the broader DMD community. However, we have demonstrated that this population was not unusually high in risk-taking personality traits and had adequate numeracy to reduce concern about numeracy bias in survey responses. Although the recruitment of caregivers (or patients) through advocacy groups has a risk of bias, it also has real benefits over qualitative approaches. Using the model process, we plan to refine the experiments and conduct a larger study with a more representative parent group and a neuromuscular clinician group. Especially important in our next study is to elicit treatment preferences from affected teenagers and adults, anticipating that DMD patients and caregivers may not assess benefit and risk in the same way.
      • McNeil D.E.
      • Davis C.
      • Jillapalli D.
      • et al.
      Duchenne muscular dystrophy: drug development and regulatory considerations.
      Second, although the study used a rigorous approach to attribute identification, the simulated treatments described in the experiment may not represent the benefit and risk profile of therapies that are ultimately approved for DMD. As with all stated-preference experiments, it remains unknown whether the presence or absence of additional attributes would influence the results. On the spectrum of patient centered to clinically centered specification of attributes, we favored the former to be consistent with the goals of patient-centered outcomes research and explore attributes meaningful to our caregiver participants. In our future studies, and when more is known about the benefits and risks of treatment, we aim to incorporate more clinically centered attributes while continuing to maintain a priority on utilizing meaningful attributes.
      Third, we conducted an aggregate analysis, and important structures in preference heterogeneity may have been overlooked. We have previously reported the differences in treatment preferences by stratifying data according to child’s ambulation status.

      Peay HL, Hollin I, Sheffer H, Bridges J. Measuring treatment preferences of children with Duchenne muscular dystrophy using best-worst scaling. Presented at the 63rd Annual Meeting of the American Society of Human Genetics; October 23, 2013; Boston, Mass.

      Although there was a small but significant difference when completing such a stratification, this could have been explained by scale differences between the 2 groups. In follow-up studies, we will aim to have a larger sample size to allow for both stratification and segmentation analysis,
      • Deal K.
      Segmenting patients and physicians using preferences from discrete choice experiments.
      which will enable being able to adequately describe preference heterogeneity.
      Fourth, we used a simple technique for estimating preferences, compared with more advanced regression techniques. As a supplemental analysis (not reported here), we reanalyzed our data by using a conditional logit. One obvious difference between the methods is that conditional logit requires using effects coding for each attribute, making each attribute have the same mean. As such, although each attribute remains on a ratio scale, the translocation of the origin inherent in effects coding implies that level importance cannot be compared across attributes. The advantage of our simple approach is that all preference weights can be estimated directly (ie, without using effects coding), and hence they all sit on the same ratio scale. We modified these results to make them comparable to the conditional logit (ie, we subtracted the attribute mean from each attribute level), and they produced nearly identical results to the conditional logit, with both methods having identical ordering (Spearman’s ρ = 1.0) and near perfect correlation (Pearson’s ρ = 0.997).
      Finally, because BWS is a relatively new stated-preference method, there is the possibility that it may present a distorted version of preference. However, there is growing interest in the method given its simplicity compared with more traditional conjoint analysis methods, which may affect respondent efficiency (ie, do responses to choice tasks reflect respondents true preferences?). We plan to validate these results against a simple conjoint analysis that was conducted as part of this study, but more research is needed to compare BWS and conjoint analysis methods.

       Implications

      The study findings are highly relevant to industry and regulators who are conducting benefit–risk assessments for potential DMD therapies. Emerging results from clinical trials suggest a slowing of motor decline, as measured by using the 6-minute talk test, and no known effect on life span.
      • Govoni A.
      • Magri F.
      • Brajkovic S.
      • et al.
      Ongoing therapeutic trials and outcome measures for Duchenne muscular dystrophy.
      Caregivers’ significant and yet finite risk tolerance has regulatory implications as well; however, given the modest risk profile emerging from many DMD clinical trials, our finding of high tolerance for adverse effects and drug-related uncertainty is also relevant.
      This study intended to leverage the FDA’s ongoing commitment to identifying methods of systematic patient engagement and, more specifically, their commitment to the use of statistical methods exploring and comparing benefits and risks to systematically quantify patients’ anecdotal reports.

      Johnson FR, Hauber AB, Poulos CM. Research Report: A Brief Introduction to the Use of Stated-Choice Methods to Measure Preferences for Treatment Benefits and Risks. RTI Press publication No. RR-0009-0909. Research Triangle Park, NC: RTI International. http://www.rti.org/rtipress. Accessed December 20, 2013

      PPMD was able to report the outcome of this study to FDA representatives in both private and public meetings. Equally important, PPMD reported the results back to the DMD community through social media, a webinar, and in-person meetings and conferences. As the FDA evaluates new drug applications for DMD therapies, they should be mindful of the value that parent decision makers place on even moderate benefits to function, their tolerance for considerable risk, and their tolerance for uncertainty.

      Conflicts of Interest

      This study was funded by Parent Project Muscular Dystrophy, which had a role in the study design; in the collection, analysis, and interpretation of data; in writing the manuscript; and in the decision to submit the manuscript for publication. Ms. Peay and Mr. Fischer are employees of PPMD, and Dr. Bridges received compensation from PPMD for this project. The authors indicated that they have no conflicts of interest regarding the content of this article.

      Acknowledgments

      We appreciate the leadership and commitment of the PPMD oversight committee: Pat Furlong, Brian Denger, Sharon Hesterlee, and Kathi Kinnett. We are indebted to the stakeholder informants, parents who participated in the cognitive interviews, and caregivers who completed the survey.
      Ms. Peay was responsible for community engagement, survey development, and writing and contributed to data analysis and data interpretation. Ms. Hollin was responsible for figure creation and contributed to data analysis, data interpretation, and writing. Mr. Fischer contributed to community engagement and survey development. Dr. Bridges was responsible for study design, data analysis, and data interpretation and contributed to survey development and writing.

      References

      1. Centers for Disease Control and Prevention (CDC). Prevalence of Duchenne/Becker muscular dystrophy among males aged 5-24 years—four states, 2007. MMWR Morb Wkly Rep. 2009;58:1119–1122

        • Emery A.E.
        Population frequencies of inherited neuromuscular diseases—a world survey.
        Neuromuscul Disord. 1991; 1: 19-29
        • Bushby K.
        • Finkel R.
        • Birnkrant D.J.
        • et al.
        Diagnosis and management of Duchenne muscular dystrophy, part 1: diagnosis, and pharmacological and psychosocial management.
        Lancet Neurol. 2010; 9: 77-93
      2. Eagle M, Baudouin SV, Chandler C, et al. Survival in Duchenne muscular dystrophy: improvements in life expectancy since 1967 and the impact of home nocturnal ventilation. Neuromuscul Disord. 2002;12:926–929.

        • Pangalila R.F.
        • van den Bos G.A.
        • Stam H.J.
        • et al.
        Subjective caregiver burden of parents of adults with Duchenne muscular dystrophy.
        Disabil Rehabil. 2012; 34: 988-996
        • Kenneson A.
        • Bobo J.K.
        The effect of caregiving on women in families with Duchenne/Becker muscular dystrophy.
        Health Soc Care Community. 2010; 18: 520-528
        • Larkindale J.
        • Yang W.
        • Hogan P.F.
        • et al.
        Cost of illness for neuromuscular diseases in the US.
        Muscle Nerve. 2014; 49: 431-438
        • Ouyang L.
        • Grosse S.D.
        • Fox M.H.
        • Bolen J.
        A national profile of health care and family impacts of children with muscular dystrophy and special health care needs in the United States.
        J Child Neurol. 2012; 27: 569-576
        • Ouyang L.
        • Grosse S.D.
        • Kenneson A.
        Health care utilization and expenditures for children and young adults with muscular dystrophy in a privately insured population.
        J Child Neurol. 2008; 23: 883-888
        • Biggar W.D.
        • Harris V.A.
        • Eliasoph L.
        • Alman B.
        Long-term benefits of Deflazacort treatment for boys with Duchenne muscular dystrophy in their second decade.
        Neuromuscul Disord. 2006; 16: 249-255
        • Angelini C.
        The role of corticosteroids in muscular dystrophy: a critical appraisal.
        Muscle Nerve. 2007; 36: 424-435
        • Bushby K.
        • Muntoni F.
        • Urtizberea A.
        • et al.
        Report on the 124th eNMC International workshop. Treatment of Duchenne muscular dystrophy; defining the gold standards of management in the use of corticosteroids.
        Neuro Muscul Disord. 2004; 14: 526-534
        • Govoni A.
        • Magri F.
        • Brajkovic S.
        • et al.
        Ongoing therapeutic trials and outcome measures for Duchenne muscular dystrophy.
        Cell Mol Life Sci. 2013; 70: 4585-4602
        • Milne C.P.
        Prospects for rapid advances in the development of new medicines for special medical needs.
        Clin Pharmacol Ther. 2014; 95: 98-109
        • Michaels M.
        • Collyar D.
        Research advocacy in traditional research settings: questions of influence and legitimacy.
        in: Earp J.L. French E.A. Gilkey M.B. Patient Advocacy for Healthcare Quality: Strategies for Achieving Patient-Centered Care. Jones and Bartlett, Boston, Mass2007: 445-473
      3. Parent Project Muscular Dystrophy. Benefit-risk assessments in rare disorders: The case for therapeutic development in Duchenne muscular dystrophy as the prototype for new approaches. http://www.parentprojectmd.org/BR. Accessed April 28, 2014

      4. Genetic Alliance UK. New medicines for serious conditions: weighing the risks and benefits. The verdict of a jury of patients. http://www.geneticalliance.org.uk/docs/citizens-jury-report.pdf. Accessed March 1, 2014.

      5. Parent Project Muscular Dystrophy: Share Your Brave Story. http://community.parentprojectmd.org/profiles/blog/list?tag=My+BRAVE+Story. Accessed March 1, 2014.

        • McNeil D.E.
        • Davis C.
        • Jillapalli D.
        • et al.
        Duchenne muscular dystrophy: drug development and regulatory considerations.
        Muscle Nerve. 2010; 41: 740-745
        • Hauber B.
        • Fairchild A.O.
        • Johnson F.R.
        Quantifying benefit-risk preferences for medical interventions: an overview of a growing empirical literature.
        Appl Health Econ Health Policy. 2013; 11: 319-329
      6. US Food and Drug Administration (FDA). Enhancing benefit-risk assessment in regulatory decision-making. http://www.fda.gov/forindustry/userfees/prescriptiondruguserfee/ucm326192.htm. Accessed March 1, 2014.

        • Mullard A.
        Patient-focused drug development programme takes first steps.
        Nat Rev Drug Discov. 2013; 12: 651-652
        • Facey K.
        • Boivin A.
        • Gracia J.
        • et al.
        Patients’ perspectives in health technology assessment: a route to robust evidence and fair deliberation.
        Int J Technol Assess Health Care. 2010; 26: 334-340
        • Hansen H.P.
        • Draborg E.
        • Kristensen F.B.
        Exploring qualitative research synthesis: the role of patients’ perspectives in health policy design and decision making.
        Patient. 2011; 4: 143-152
        • Vogt F.
        • Schwappach D.
        • Bridges J.
        Accounting for tastes: a German perspective on the inclusion of patient preferences in health care.
        Pharmacoeconomics. 2006; 24: 419-423
        • Coplan P.
        • Noel R.
        • Levitan B.
        • et al.
        Development of a framework for enhancing the transparency, reproducibility and communication of the benefit–risk balance of medicines.
        Clin Pharmacol Ther. 2010; 89: 312-315
        • US Food and Drug Administration (FDA)
        Factors to Consider When Making Benefit-Risk Determinations in Medical Device Premarket Approval and De Novo Classifications. FDA Guidance for Industry and Food and Drug Administration Staff. Silver Spring, Md: Center for Devices and Radiological Health, Center for Biologics Evaluation and Research.
        Food and Drug Administration. March 28, 2012;
        • Fleurence R.
        • Selby J.V.
        • Odom-Walker K.
        • et al.
        How the Patient-Centered Outcomes Research Institute is engaging patients and others in shaping its research agenda.
        Health Affairs. 2013; 32: 393-400
        • Bridges J.
        • Hauber A.B.
        • Marshall D.
        • et al.
        A checklist for conjoint analysis applications in health: report of the ISPOR Conjoint Analysis Good Research Practices Taskforce.
        Value Health. 2011; 14: 403-413
      7. Marshall D, Bridges J, Hauber AB, et al. Conjoint analysis applications in health—how are studies being designed and reported? An update on current practice in the published literature between 2005 and 2008. Patient. 2010;3:249–256.

        • Bridges J.
        • Kinter E.
        • Schmeding A.
        • et al.
        Can patients diagnosed with schizophrenia complete choice-based conjoint analysis tasks.
        Patient. 2011; 4: 267-275
        • Coast J.
        • Flynn T.N.
        • Salisbury C.
        • et al.
        Maximising responses to discrete choice experiments: a randomised trial.
        Appl Health Econ Health Policy. 2006; 5: 249-260
        • Reed Johnson F.
        • Lancsar E.
        • Marshall D.
        • et al.
        Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force.
        Value Health. 2013; 16: 3-13
      8. International Society for Pharmacoeconomics and Outcomes Research. ISPOR Statistical Methods for Conjoint Analysis Task Force. http://www.ispor.org/TaskForces/Analyzing_Data_Conjoint_Analysis.asp. Accessed March 1, 2014.

        • Erdem S.
        • Rigby D.
        • Wossink A.
        Using best-worst scaling to explore perceptions of relative responsibility for ensuring food safety.
        Food Policy. 2012; 37: 661-670
        • Flynn T.N.
        • Louvier J.J.
        • Peters T.J.
        • Coast J.
        Best-worst scaling: what it can do for health care research and how to do it.
        J Health Econ. 2007; 26: 171-189
        • Bridges J.
        Stated preference methods in health care evaluation: an emerging methodological paradigm in health economics.
        Appl Health Econ Health Policy. 2003; 2: 213-224
        • Flynn T.N.
        Valuing citizen and patient preferences in health: recent developments in three types of best-worst scaling.
        Expert Rev Pharmacoecon Outcomes Res. 2010; 10: 259-267
        • Louviere J.J.
        • Flynn T.N.
        Using best-worst scaling choice experiments to measure public perceptions and preferences for healthcare reform in Australia.
        Patient. 2010; 3: 275-283
        • Gallego G.
        • Bridges J.
        • Flynn T.
        • et al.
        Using best-worst scaling in horizon scanning for hepatocellular carcinoma technologies.
        Int J Technol Assess Health Care. 2012; 28: 339-346
      9. Peay HL, Sheffer H, Tibben A. Expectations and decision making in clinical trials for Duchenne and Becker muscular dystrophy. Abstract presented at: 18th International Congress of The World Muscle Society; October 1–5, 2013; Asilomar, Calif.

      10. National Center for Education Statistics. National assessment of adult literacy (NAAL). A first look at the literacy of America’s adults in the 21st century. https://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006470. Accessed March 1, 2014.

        • Townsend D.
        • Yasuda S.
        • Li S.
        • et al.
        Emergent dilated cardiomyopathy caused by targeted repair of dystrophic skeletal muscle.
        Mol Ther. 2008; 16: 832-835
        • Shirokova N.
        • Niggli E.
        Cardiac phenotype of Duchenne muscular dystrophy: insights from cellular studies.
        J Mol Cell Cardiol. 2013; 58: 217-224
      11. Kuhfeld W. Orthogonal arrays [TS-723]. Cary (NC): SAS http://support.sas.com/techsup/technote/ts723.html. Accessed March 1, 2014

        • Pearson S.D.
        • Goldman L.
        • Orav E.J.
        • et al.
        Triage decisions for emergency department patients with chest pain: do physicians’ risk attitudes make the difference?.
        J Gen Intern Med. 1995; 10: 557-564
      12. Wallston KA, McNaughton C, Storrow A, et al. Validation of a short, 3-item version of the Subjective Numeracy Scale (SNS-3). Presented at: Health Literacy Research Conference; October 17-18, 2011; Chicago, IL

        • Peters E.
        • Hibbard J.
        • Slovic P.
        • Dieckmann N.
        Numeracy skill and the communication, comprehension, and use of risk-benefit information.
        Health Affairs. 2007; 3: 741-748
        • Zikmund-Fisher B.
        • Smith D.M.
        • Ubel P.A.
        • Fagerlin A.
        Validation of the Subjective Numeracy Scale: effects of low numeracy on comprehension of risk communications and utility elicitations.
        Med Decis Making. 2007; 27: 663-671
        • Fagerlin A.
        • Zikmund-Fisher B.J.
        • Ubel P.A.
        • et al.
        Measuring numeracy without a math test: development of the Subjective Numeracy Scale.
        Med Decis Making. 2007; 27: 672-680
        • Molassiotis A.
        • Emsley R.
        • Ashcroft D.
        • et al.
        Applying best-worst scaling methodology to establish delivery preferences of a symptom supportive care intervention in patients with lung cancer.
        Lung Cancer. 2012; 77: 199-204
        • McFadden D.
        Conditional logit analysis of qualitative choice behavior.
        in: Zarembka P. Frontiers in Econometrics. Academic Press, New York, NY1974: 105-142
        • Orme B.K.
        Getting Started With Conjoint Analysis: Strategies for Product Design and Pricing Research.
        Research Publishers. 2006
        • Bridges J.
        • Mohamed A.F.
        • Finnern H.W.
        • et al.
        Patients’ preferences for treatment outcomes for advanced non-small cell lung cancer: a conjoint analysis.
        Lung Cancer. 2012; 77: 224-231
      13. Peay HL, Hollin I, Sheffer H, Bridges J. Measuring treatment preferences of children with Duchenne muscular dystrophy using best-worst scaling. Presented at the 63rd Annual Meeting of the American Society of Human Genetics; October 23, 2013; Boston, Mass.

        • Deal K.
        Segmenting patients and physicians using preferences from discrete choice experiments.
        Patient. 2014; 7: 5-21
      14. Johnson FR, Hauber AB, Poulos CM. Research Report: A Brief Introduction to the Use of Stated-Choice Methods to Measure Preferences for Treatment Benefits and Risks. RTI Press publication No. RR-0009-0909. Research Triangle Park, NC: RTI International. http://www.rti.org/rtipress. Accessed December 20, 2013

      Linked Article

      • Erratum
        Clinical TherapeuticsVol. 36Issue 8
        • Preview
          In the article by H.L. Peay, et al titled “A Community-Engaged Approach to Quantifying Caregiver Preferences for the Benefits and Risks of Emerging Therapies for Duchenne Muscular Dystrophy” (Clin Ther. 2014;36;624–637), on page 631 in Table II there was a typo in the mean age of children row. It should have read 12.1, not 21.1. The correct Table is shown below.
        • Full-Text
        • PDF