Empirical evidence for pediatric drug policy April 16, 2019 Front - PowerPoint PPT Presentation

Social values and child health priorities: Avram Denburg, MD MSc PhD FRCPC 2019 CADTH Symposium Empirical evidence for pediatric drug policy April 16, 2019

Front Matter Acknowle ledgments and Disc isclo losures • Research funding: CIHR, CCHCSP, CHEPA, Trudeau Foundation • Co-investigators: Julia Abelson, Shiyi Chen, Mita Giacomini, Jeremiah Hurley, Wendy Ungar • Conflicts of interests: None

Presentation Overview 1. 1. Ba Background • Health system priority setting • Pediatric drug policy in Canada 2. 2. St Stated preference su surv rvey • Aims • Methods • Results 3. 3. Dis iscussio ion • Limitations • Policy relevance

Resource Scarcity: Costs Outpacing Growth

Media ting Public Policy

Public Drug Policy for Children in Canada Th The Con ontext • Patchwork national coverage • Unique dimensions of child health relevant to drug policymaking • Lack of child-specific policy: Regulator  HTA  Payer • Opportunity to lead Denburg et al, CMAJ 2017 1

Pediatric Drug Access in Canada Hea ealt lth system ch chall llenges HTA ch challe llenges • Market dynamics • Biology • Epidemiology • Political economy Industry Evidence • Trial enrolment • Evidence of safety and efficacy • Life-course dynamics • Pediatric indication Regulator • Child and family utilities Economics • Externalities • HTA submission • Priority-setting for review HTA • Procedural • Substantive • Opportunity cost Ethics • Spatial • Political environment Payer Lack of coherent, consistent and equitable drug policy

Survey: Context and Justification Why so societal l valu alues? Why ch child ildren? • Growing recognition of the • Age recognized as a morally relevance of societal values for relevant variable, but little health system priority setting dedicated inquiry into allocative preferences regarding children • Two contrasting approaches to • Need for knowledge of the social the elicitation of societal values: values attached to health care • Population-based surveys priority setting affecting children • Deliberative engagement NESTED DEL DELIBERATI TION CHI CHILD-ADULT TRADE-OFFS

Survey: Objective and Methods • Objective: To test societal preferences • De Design: for health resource allocation between • ‘Clinical’ vignettes to ground choice children and adults and assess the scenarios impact of moral reasoning on • Randomization to moral reasoning preferences exercise • Numerical preference scores • Methodology: : Stated preference • Analysis is: survey • Descriptive statistics • Sa Sample: Population-based sample of • Univariate and multiple regression the Canadian public mixed models • Da • Multiple regression GEE model Data so sources: Marketing research firm panel, email invitation, online administration

Clinical scenario: Novel cancer therapy Dru Drug A Drug B Dru A new therapy is available for patients with A new therapy is available for patients with Chronic Ch Ch Child Malig alignancy. Chronic Adu Ch Adult Mal Malignancy. Patients are 10 10 years old old, on average. Patients are 40 years old old, on average. With this treatment, patients are cured of their With this treatment, patients are cured of their cancer and can expect to live to average life cancer and can expect to live to normal life expectancy (80 years). expectancy (80 years). Without the therapy, the disease causes death Without the therapy, the disease causes death within 6 months. within 6 months. • The health system can only afford to fund one of the two therapies at present. Which drug should it fund? Please slide the bar to any point on the scale to show your strength of support for funding one of the therapies. -5 0 5 Definitely Fund Definitely fund either fund Drug A drug Drug B

Randomization: Moral reasoning exercise Moral l Reasonin ing In Interv rventio ion: • Values drawn from antecedent conceptual phase, supplemented by review of broader literature on ethics and priority setting for health systems • List of principles, with associated descriptive statements • Participants to select ‘top 3’ (of 12) principles most relevant for each scenario Ge General: Chi hild ld-focused:  Efficacy  Potential  Equality  Fair innings  Suffering  Dependency  Rescue  Vulnerability  Personal responsibility  Rarity  Economic productivity  Distinction

Moral reasoning exercise: Sample Fund d treatm tment nt based on evidenc dence that t it works rks "Fund treatments best proven to be safe and effective."  "Since it is harder to study treatments in children, evidence is usually stronger for adult  treatments." Help everyone ne to live a f full life "Give the younger patients a chance for a full life."  "The older patients have had their turn."  Treat t people who will benefit longer "Giving the treatment to the younger group makes sense, since they will enjoy it longer."  "Lifelong potential should be factored into decisions about which health interventions to  fund." Treat t people with th fa family y or ot other er responsi ponsibilities "At 40, people may be raising families or have others who rely on them."  Treat t the most t vulnera erable "Resources should be directed to help those that cannot protect or advocate for  themselves." "Children are still developing, so can suffer lifelong consequences from untreated  disease." Treat t people who are productiv ductive "Helping people who are in the workforce has benefits for all." 

Specific Aims • Aim im 1: : Understand the direction and strength of societal preferences for health resource allocation between children and adults for disparate treatment scenarios • Aim im 2: : Assess the impact of a moral reasoning exercise on the expression of such preferences • Aim im 3: : Identify sociodemographic factors that significantly impact the expression of societal preferences on health resource allocation between children and adults • Aim im 4: : Test the divergence of participant preferences for children or adults from an assumption of neutrality • Aim im 5: : Characterize the principles that most influenced participants’ allocative decisions

Sample Enrollment • Nationally representative sample of Email invitations (n=12803) Canadian adult general public • Interlocking quotas for stratified Initiated survey (n=2777) sampling (age, gender and region), balanced against Statistics Canada Excluded (n=1221) norms ¨ Not meeting inclusion criteria (n= 32) ¨ Incomplete survey (n=516) ¨ Full quotas (n=500) • Primary comparison: difference in ¨ Poor quality (e.g. racing) (n=173) mean strength of preference between Randomized (n=1556) the intervention and control groups • Sample size: • 2-sided, 2-sample t-test with equal Allocation variance Allocated to intervention (n=773) Allocated to control (n=783) ¨ Received allocated intervention (n=773) ¨ Received allocated intervention (n=783) • Population mean difference = 0.3 • SD = 1.67 •  = 0.01 Analysis • Power = 80% Analysed (n=773) Analysed (n=783) ¨ Excluded from analysis (n=0) ¨ Excluded from analysis (n=0)  750 per r gr group (t (total l n =1 =1500)

Comparison of group mean preference scores across scenarios Circle/plus = mean; centre line = median; box = interquartile range (IQR: 1 st and 3 rd quartiles of the data); whisker (inner fences): lower = 1 st quartile - 1.5SD, upper = 3 rd quartile + 1.5SD; suspected outliers are noted with a circle (control group) or plus sign (intervention group) beyond the upper and lower inner fences