Advances in Small Trials dEsign for Regulatory Innovation and eXcellence

Egbert  BiesheuvelNutricia  Research,  Utrecht

On  behalf  of    Kit  Roes,  Armin  Koch,  Martin  Posch,  Ferran Torres,Hanneke vd Lee,  Cor  Oosterwijk,  Caroline  van  Baal  and  

all   the  researchers  of  the  Asterix consortium

Outline

• Perspectives,  Patients and Evidence

• Concept  and  objectives  of  asterix

• Examples  of  progress  in  design  and  analysis

Perspectives, Patients and Evidence

Dependingon  definitions :  +/-­‐ 8000  rare  diseases

>  1000  new  therapiesdesignated as  orphan

90  orphanmedicineauthorised (7  in  2013,  14  in  2014)

Authorised does  not automatically lead  to  available for patients

Perspective  of  market  authorisation of  a  new  drugEvidence  based  decision of  allowing  physicians  to  add  a  new  drugto  their  treatment  options

EMA  Orphan medicinesfigures 2000-­‐2014

Perspectives, Patients and Evidence

In  the  meantime I  passed away

Perspectives, Patients and Evidence

The  European  legislation  on  orphan  medicinal  products  [Regulation  (EC)  No  141/2000]  emphasises that  patients  suffering  from  rare  conditions  should  be  

“entitled  to  the  same  quality  of  treatment  as  other patients.”

Current rationale  is  to present  evidence at  the  sameconfidence levels

Small  populations guidancedoes  stimulate alternatives for design  and  analyses

Careful case-­‐by-­‐case  decisions are  made,  that essentially may “relax”  level  of  evidence

Concept and objectives of asterix

• Unmet  need  for  drugs  to  treat  rare  diseases

• Difficulty  to  establish  efficient  and  reliable  evidence  from  clinical  trials  in  small  populations

• Absence  of  methods  to  include  patients  and  patient  perspectives to  generate  results  that  matter  to  patients

• Uncertainty  in  regulatory  decision  making  on  new  treatments

PatientThinkTank

Patient Think Tank

• Systematic involvement of  patients and their perspectives• PTT  comprises of  12  members  

Provide input  in  the  development of  methods to• optimize use of  info  in  patient registries to decide on  trial  design• include  patients  preferences  in  the  weighting  of  outcomes• includepatient opinionson  novel trial  designs

1st F-­‐2-­‐F  meeting  in  October  2014  A’dam on  adaptive  designs  andweighing  of  outcomes

2nd F-­‐2-­‐F  meeting  in  October  2015  Barcelona  on  Framework  

Example: Framework for guidance

• Guidance on  design  at  disease level  no  longer practical  (over  8000  rare  diseases)

• One general document  (at  present)  may not providesufficient guidance

• Framework  with intermediate approach,  driven by keycharacteristics of  disease and  treatment

Caridad Pontes,  Ferran Torres,   Josep Torrent et  al.

Proposed framework

Clinicalcourse

Acute

Chronic

Single  acute episode

Repeated acuteepisodes

Slow/  Non  progressive

Progressive led  byone system/organ

Progressivemultidimensional  

multiorgan

Staged disease

Life threateningFulminatingTime  to  event

If SOC,  back to  normal

Repeated eventsPredictable courseClear-­‐cut episodes

Numer of  events,  time

Single  organ drivenLife-­‐long disease

Predictable course,  surrogatesSOC  generally available

Adults,  disablingMultidimensional  single  organPatient reported outcomes,  QoLSurrogates requiring validationChildren,  life lasting,  registriesMultidimensional  multiorganPatient/caregiver otucomes

Poor SOC

Poor prognosisSubgroups requiredTime  to  end-­‐pointsSurrogates validated

Rareor  very

rareFrequency

Ultrarare(<1/10

5)

Example: Multiple endpoints

• New  Fall back procedures– Allow inference on  subsets of  endpoints– Control  the  Family  Wise Error  Rate (Type  I  error)– Can be designed to  bemore  powerful,

taking joint  distribution (correlations)  into account

• Simulation studies  show  benefit,also for small  sample  sizes

• Traditional  level  of  evidencelimits power  Robin  Ristl,  Martin  Posch

Example: Series of trials

• In  drug  development for rare  diseases,  synthesis throughmeta-­‐analysis  might improve robustness of  (regulatory)  assesment

• Sequential meta-­‐analysis  can improve efficiency  of  drug  development plans

• Facilitating adaptive licensing

Armin  Koch,  Kristina Weber,   Ingeborg  van  der  Tweel,  Konstantinos  Pateras

Adaptive licensing

Eichler  et  al.,  Clin Pharm &  Therapeutics 2012;  91:  42-­‐37

Example: Series of trials

Initial idea:• Bayesian  methods  for  extrapolation  in  rare  disease  • Reduce  the  burden  for  formal  proof  of  efficacy

Impact  on  decision  making  needs  to  be  fully  understood• Frequentist  and  Bayesian  strategies  are  compared  from  a  

decision  making  perspective• In  case  of  2  (completed)  trials:  (1)  full  meta-­‐analysis,  (2)  one  

trial  as  prior  for  the  other,  (3)  down-­‐weighing  the  first  trial

Concluding

• Innovations in  design  possible– but  more  complicated to  comprehend

• Serious reconsideration of  level  of  evidence needed,  in  concert  with new  models of  decision making

• Framework  will help  to  provide guidance in  theheterogenousworld of  rare  diseases

• Patient involvement is  essential

• We  have  just started…..