Steps to PRO/ObsRO development | Challenges | Solutions |
---|---|---|
Literature Review/Desk Research | • Literature often limited • Broad age ranges covered | • Consider grey literature (blogs, dissertations) • Conduct interviews with expert clinicians, nurses and patient advocacy groups |
Concept Elicitation | • Limitations in memory, cognitive ability, language by age/condition • Children can be shy • Rarity of condition makes recruitment/saturation hard to achieve • Parents unable to report some symptoms/domains not known to them | • Carefully guided interview guides and well trained interviewers • Creative interview techniques, toys and drawings • Collapse age groups where appropriate • Must achieve saturation within each narrow age range – can this be relaxed for orphan indications? Get FDA feedback early • Consider other respondents (teachers, nanny etc) |
Selection/ development of a measure | • Few disease specific measures exist in paediatrics and orphan diseases • Existing instruments don’t meet FDA/EMA guidance • Who is the best respondent? • How should you administer the questionnaire? • How should the questionnaire be worded? • Child can’t remember without a concrete event to recall to • Parent items must be observable….but they may not be with the child all day | • Think about PRO selection early • Talk to patient advocacy groups • Engage FDA early • Consider EPRO vs pen/paper vs IVRS in context of condition and age of child • Questions/responses should be clear and simply worded • Short recall period required • Consider all types of respondents (parent, teacher, nurse, child, dr) • Consider ‘child told me’ questions |
Cognitive debriefing/ content validity testing | • Hypothetical situations don’t work with children • Children give you answers they think you want to hear • Small sample sizes in rare conditions | • Allow child to complete diaries at home for a few days prior • Use carefully worded interview guides and well trained interviewers • Questioning should not be too repetitive nor lengthy • When analysing, check for consistency between behaviour and responses • Collapse age groups as appropriate |
Psychometric validation | • Sample should be stratified by age group, but small samples in orphan indications | • Consider validating as part of trial and/or include data from cognitive debriefing (move forward at risk) • Consider collapsing across age groups • Consult regulatory early • Utilize psychometrics done in other diseases if adapting a measure |