Sharing Data: Recovering Registry Addict

Transcription

1 Sharing Data: Observations From a Recovering Registry Addict Christopher Bredeson, MD, MSc., FRCPC Director, Hematologic Malignancies Professor of Medicine Medical College of Wisconsin DBV06_1.ppt

2 TODAY S TOPIC IS - REGISTRIES DBV06_2.ppt

3 TODAY S TOPIC IS - REGISTRIES DBV06_2.ppt

4 What I Am Supposed To Cover Overview Funding Privacy Security Governance Issues / challenges Results

5

6 What Is The Purpose Of The Registry? Research Natural history Rare events Tracking product Safety Lobbying Quality of care / standards

7 Definitions A patient registry is an organized system that uses observational study methods to collect uniform data (clinical and other) to evaluate specified outcomes for a population defined by a particular disease, condition, or exposure, and that serves one or more predetermined scientific, clinical, or policy purposes. The patient registry database describes a file (or files) derived from the registry.

8 We Need Data in Medicine Med/Scientific Community Payors Industry Patient Assess trends Determine efficacy Track product Allocate resources Monitor outcomes Multiple Stakeholders may use the same registry! DBV06_4.ppt

9 International BMT Registry Established in 1972 to monitor and study outcomes of bone marrow transplants Maintains a database of clinical information on recipients of autologous and allogeneic hematopoietic stem cell transplants in ~500 centers in >50 countries Collates basic data set on all patients in member centers (registration) and comprehensive data (research) on a subset Provides scientific and statistical support for analyzing those data Primary purpose outcomes research DBV06_23.ppt

10 IBMTR 1985 (1985 year of first major NIH funding) : 200 centers 1,000 transplants 35 publications Mortimer M. Bortin, MD Scientific Director Statistician 1 Data Management 3 Administrative Asst 1

11 Bortin Transplantation 1970 Vol. 9 p 571

12 Number of Transplants per Year Annual Numbers of Blood and Marrow Transplantations, Worldwide - 40,000 35,000 Allogeneic Autologous 30,000 25,000 20,000 15,000 10,000 5, Year

13 ACS/ NIH BMT reg (1970) IBMTR to Milwaukee (1972) SHORT HISTORY OF IBMTR / ABMTR <5 new teams/yr <60 cases/yr ( ) 1,000 th case (1981) IBMTR Clinical Trials Methodology Technology assessment Risk factors Descriptive analyses Major NIH funding (1985) ABMTR 10,000 th case (1989) >20,000 cases 1 st publications (1996) >60,000 cases Technology assessment Risk factors Descriptive analyses Study design Data collection forms Mmh01_2.ppt

14 Evolution Leads To Complexity

15

16 ACS/ NIH BMT IBMTR to <5 new teams/yr reg Milwaukee <60 cases/yr (1970) (1972) ( ) Someone Has To Own It! 1,000 th case (1981) IBMTR Clinical Trials Methodology Technology assessment Risk factors Descriptive analyses (Benign) Dictatorship Oligarchy Democracy Major NIH funding (1985) ABMTR 10,000 th case (1989) >20,000 cases 1 st publications (1996) >60,000 cases Technology assessment Risk factors Descriptive analyses Study design Data collection forms Mmh01_2.ppt

17 Centralized Operations

18 Structure and Governance

19 Oversight and Guidance Independent Elected and career people

20 Central leadership Career Content and methodologic expertise

21 Functional working unit for observational research Stat Center staff Elected Directors Open to all

22 Patient and donor perspective Assist in guiding research agenda Communicate with non-medical community

23 Types of research activities

24 Collaboration Between Different Entities/Organizations IBMTR MILWAUKEE Overall Coordination Scientific Leadership Medical Monitoring Statistical Design/ Analysis Protocol Development/ Implementation Electronic Communications Data Management Trial Oversight/ Monitoring Lab/ Repository Management EMMES WASHINGTON Patient Advocacy Contracting NMDP MINNEAPOLIS DCC02_3.ppt

25 Money Issues Cost ~ personnel and Data Infrastructure / operations Data acquisition Technology Paper is cheap up front, electronic data cheaper in the long-term Life cycle considerations Does the registry have a finite life? Data access in the afterlife? Open access? Real-time or later? Data = money

26 Funding Federal Contracts Hassle Factor Philanthropy Industry Unrestricted Industry Contract Peer Rev Grants Amount of Funding

27 Data and Privacy Which patients? All patients? All consenting patients? Selected or randomly selected patients? You want to be able to update data Need a link between center and registry database What are you sending? Data +/- tissue Who owns the data? Does the registry give the data back to the center?

28 Data and Consent Anonymous data De-identified data Center sends with unique ID# Center knows ID#=patient name Data with Identifiers included Patient consent Higher risk of disclosure of personal health information Usually still communicate with ID#

29 Ethical / Legal Issues Transparency Consent Incentives are dodgy Require review of IRB policies and Canadian regulations Should also pass the newspaper test If activity was described in the local paper would it be viewed favourably?

30 CIBMTR Data Available Registration Database Basic (essential) information on consecutive transplant recipients in participating institutions Age, sex, disease, disease stage and duration, graft type and treatment, conditioning regimen, posttransplant disease status, GVHD, survival, cause of death, new cancers Research Database Comprehensive patient, disease, treatment and outcome data Data elements selected to allow investigation of important issues in the field DBV06_30.ppt

31 CIBMTR Data Validity Data dictionary Data review / cleaning Computerized E.g. transplant date after birth date Manual Data up to date Calendar driven Carrots and sticks Data audits All centers every three years All critical data elements Sampling of others

32 Data Forms Data forms design is a difficult Data forms central to success or failure of registry Leave nothing to interpretation Minimal free text options Comprehensive data manual to accompany form help line

33 Technology Is Not Smart So much for computer dating!

34 There Are Still Barriers To Overcome

35 Data Dictionary: Speaking The Same Language 08/10/09 August 10, th of October 2009 October 9th, 2008

36 Getting The Data

37 Sharing The Rewards

38 WORKING COMMITTEES Participants Include Clinicians and Scientists from >300 institutions worldwide Acute Leukemia Chronic Leukemia Lymphoma Plasma Cell Disorders Solid Tumors Pediatric Cancer Non-Malignant Marrow Disorders Immune Deficiencies / Inborn Errors Autoimmune Diseases Graft Sources/Manipulation GVHD Late Effects & QOL Immunobiology Infection / Immune Reconstitution Regimen-related Toxicity Emerging Cellular Therapies Health Services & Psychosocial Issues Donor Health & Safety International Studies

39 Finding A Niche You call this a niche!

40 Accomplishments

41 PROBABILITY OF RELAPSE, % The Allogeneic Graft Vs. Leukemia Effect P< Identical Twin (N=34) HLA-Identical Sibling (N=340) YEARS ESH00_3a.ppt

42 PROBABILITY OF SURVIVAL, % Longterm Prognosis After BMT SAA (N = 1,029) AML (N = 2,058) CML (N = 2,146) ALL (N = 1,458) 20 P = YEARS NEJM 1999 EHA00_22.ppt

43 Cumulative Incidence, % Cumulative Incidence of PTLD and Invasive Solid Cancers Following Allogeneic BMT 8 6 Solid cancers (N = 80) 6.6% % PTLD (N = 78) 0 1.0% Years DBV06_54.ppt

44 Center For International Blood And Marrow Transplant Research (CIBMTR) Established July 2004 A research affiliation between the IBMTR and the NMDP to support clinical research in BMT & related fields Clinical Research includes Observational Studies (including immunobiologic correlates) Clinical Trials Health Services Research Statistical Methodology

45 CIBMTR Clinical Research Working Committees >200 Active Studies >7,600 samples distributed by the NMDP repository for immunobiology working committee studies (2006) 56 scientific papers submitted or accepted for publications in ASH meeting presentation: 2006: 10 selected studies (2 posters, 8 oral) 2007: 18 selected studies (12 posters, 6 oral) 2008: 24 selected studies (11 posters, 13 oral)

46 Summary A registry can be a powerful tool Someone must own it Fund with a balanced portfolio Form follows function Data elements: complete parsimony is better Play nicely in the sandbox it is their sand!