Exploring a dataset for pragmatic clinical studies in ...

1 Dr Andrew Cook Dr Eleanor Woodford-Guegan Exploring a dataset for pragmatic clinical studies in development Trastuzumab Trials Trial Duration Outcome Measures Persephone 12 months vs 6 months Quality of Life Health Economics Cardiotoxicity PHARE 12 months vs 6 months Disease Free Survival Cardiotoxicity SOLD 9 weeks vs 12 months Disease Free Survival Cardiotoxicity How do we get to know what other funders are planning? We need to bring all those plans together in an easily readable and accessible format. Agree Minimum dataset NETSCC M1 M12 Test dataset NETSCC M13 M18 Implement Database HAS M19 M 36 DatasetEpidemiologyStudyOutcomePICOF underContactTrial TitleTrial ReferenceResearch QuestionSource of QuestionQuestion DetailBrown et al, BMJ 2006 333 p804-6 Finding Participants EUnetHTA INAHTA Personal Contacts Snowballing NIHR HTANIHR ProgrammeGrantsRfPBCombined Nor wegian ResponseZonMWCVZHASI nstitute du CancerMoHASERNIP-SUnited StatesInstitute of He alth EconomicsBlue Cross / Blue Shield Technology Evaluation CenterTHE dataset Item Definition Mandatory Example Funder The name of the funder considering the study Y NIHR HTA Programme Contact The name of the funder considering the study Y Title The study title.

2 In English Y Multi-centre randomised controlled trial of the cost-effectiveness of infra-inguinal percutaneous transluminal angioplasty (PTA) versus reconstructive surgery for severe limb ischaemia (BASIL) Unique ID The funder s reference for the study Y 96/05/01 Country The country(s) where the study is planned to take place Y United Kingdom Item Definition Mandatory Example Native Language Title The study title, in the funder s preferred language N Research Question The primary research question of the study, expressed in English language Y What is the clinical and cost effectiveness of angioplasty compared to surgery in the management of severe limb ischaemia? Native Language Research Question The primary research question of the study, expressed in the funder s preferred language N Source of Question Where the question came from eg was it specified by a national policy making or research coordinating body?

3 N URL of NICE guidance specifying the question Page 2 of 5 Item Definition Mandatory Example P Natural language description of the population eligible for study inclusion N Patients with severe lower limb ischaemia I Natural language description of the new intervention(s) considered in the study N Angioplasty C Natural language description of the control intervention(s) considered in the study N Surgery O Natural language description of the key outcomes of interest N Mortality, amputation, quality of life, cost effectiveness Page 3 of 5 Item Definition Mandatory Example P-MESH MESH terms for the population eligible for study inclusion Y Femoral Artery [ ] I-MESH MESH terms for the new intervention(s) considered in the study Y Angioplasty [ ] C-MESH MESH terms for the control intervention(s) considered in the study Y Blood Vessel Prosthesis Implantation [ ] O-MESH MESH terms for the key outcomes of interest Y Amputation [ ] Quality Adjusted Life Years [ ] Cost-Benefit Analysis [ ] Fatal Outcome [ ] Page 4 of 5 Item Definition Mandatory Example Outcome The eventual outcome of the planned study Y Open Summary A summary of the study N This study Page 5 of 5 IS IT ANY USE?

4 Two things to Efficacy Is it able to identify matching studies ? Effectiveness How useful is it in real life? Efficacy Does it work at all? Indication Intervention Comparator 1 Breast Cancer Trazuzumab (Herceptin) Other Monoclonal 2 Aortic Stenosis TAVI Surgery, Conservative Treatment 3 Osteoporosis Vertebroplasty Conservative Treatment Physio/OT 4 Macular Degeneration Bevacizumab (Avastin) Bevacizumab for alternative indication WILL IT WORK IN THE REAL WORLD? The Matching Algorithm Delphi participants suggested this should prefer sensitivity over specificity ie if in doubt, indicate a match Therefore expect false positives Consequences False Positives False Negatives Contact matched organisation. Discover identified study is not sufficiently similar Two studies go ahead from different funders without benefit of knowing about each other.

5 (or systematic review is undertaken without knowledge of major new information immanent) A few hours work /$/ 1000 Duplicated clinical Trial Insufficient Systematic Review /$/ 100,000 - /$/ 1,000,000+ NEXT We re on target to deliver the final dataset to HAS by the end of June HAS will then implement a registry using this dataset on their Eiffel platform For any pragmatic research funders, or similar organisations, it s not too late to get involved. Thanks