Caroline Blackwell - Collection, Management, and Sharing of Data in Clinical Trials

Transcript

1. Collection, Management, and Sharing of Data in Clinical Trials July 17, 2014
2. Objectives • Understand the basic data management needs of clinical trials • Explore the different methods for data sharing available • Examine experiences from ongoing or recently completed trials
3. What is a Clinical Trial? • Research studies that explore whether a medical strategy, treatment, or device is safe and effective for humans. • May show which medical approaches work best for certain illnesses or groups of people. • Produce the best data available for health care decision-making. National Institutes of Health, 2014
4. Examples of Clinical Trials • Drug studies (monitored by FDA, classified in Phases 0-IV) • Device trials • Lifestyle interventions (exercise, diet) • Comparative effectiveness studies
5. Types of Data Collected • Physical examinations • Participant interviews • Questionnaires (self or interviewer- administered) • Laboratory measurements • Imaging • Existing medical records
6. How do we determine what data to collect? • Things to consider: • Requirements for treatment administration/adherence assessment • Rate of occurrence of outcomes of interest • Participant health care needs • Cost of participant visit • Participant convenience/fatigue
7. Ensuring High Quality Data • Standardized measurement techniques used when possible (national guidelines, validated procedures from other studies) • Technology can improve quality: • Teleforms, eForms, vioFFQ • Digital Monitoring Devices • Systems for data management
8. Data Management Systems • Many studies use web-based data management • Data management often a consideration in scientific review for federally submitted grants • Data entry screens designed, when possible, to mirror paper data collection forms to minimize errors in data entry • Allows for use of validation checks (including determination of eligibility for trial) and data entry rules
9. Web-Based Data Management • Track adherence to protocol, scheduling of visits, participant compliance • Can be customized or off-the-shelf • Enables centralized data management, especially valuable in large multi-center trials • Also allows for easy export of complete data sets to existing data warehouses or repositories and sharing of data between groups
10. NIH Data Sharing Repositories • Numerous repositories exist for different NIH Institutes and Centers • Contain clinical data, bio-samples (blood, urine, tissue), images • Individual institutes have varying policies for accessing and submitting data • http://www.nlm.nih.gov/NIHbmic/nih_data_shari ng_repositories.html
11. Ancillary Studies • Many large NIH trials also offer opportunities for new collaborators to propose ancillary studies using samples, data, etc. already collected • Data, samples usually controlled by the study group not a data repository • ACCORD (diabetes) • Women’s Health Initiative (HRT) • ARIC (atherosclerosis) • NIH funding mechanisms may be available for ancillary studies
12. Wake Forest Baptist Medical Center Community-Based Translational Studies The Healthy Living Partnerships to Prevent Diabetes (HELP PD) study
13. Background on Diabetes Prevention • Sedentary Lifestyle + Increased Calorie Consumption • 34% of adults over 20 years of age are overweight; additional 34% considered obese1 • Over 20 million people with diabetes, 50 million with pre-diabetes in the US2 • Estimated healthcare burden of diabetes: more than $197 billion in US in 20103 1NHANES 2007-2008 2CDC. National diabetes fact sheet:2007. 3Zhang P et al. Global healthcare expenditure on diabetes for 2010 and 2030. Diabetes Res Clin Pract 2010.
14. Diabetes Prevention Program (DPP) • Landmark, multi-center clinical trial funded by NIH • 3,234 participants at 27 sites across the US • Participants randomly assigned to one of four interventions: • Intensive lifestyle intervention • Metformin • Placebo • Troglitazone (stopped early due to potential for liver toxicity)
15. DPP Lifestyle Intervention • 16 lesson intensive curriculum with emphasis on behavior modification, caloric restriction, and increased physical activity • Goal of 7% weight loss • Individual intervention delivered by professional case managers (RDs, RNs, and exercise physiologists) in a clinical setting • 4 mg/dL reduction in fasting glucose and 4.9% weight loss in lifestyle intervention group over 2.8 mean years of follow-up; 58% reduction in incidence of diabetes
16. Origins of HELP PD Based on evidence from DPP, NIH requested proposals for translational studies HELP PD developed in response to this call for proposals
17. What is Translational Research? • Translational research includes two areas: • the process of applying discoveries generated during research in the laboratory, and in preclinical studies, to the development of trials and studies in humans • the study and facilitation of the application of research findings to the community, aimed at enhancing the adoption of best practices Bench research Clinical Research Community application
18. Other Translations of DPP Author Year Study N Treatment condition and setting Comparison Condition Follow-up Weight loss Effect Glucose Effect Boltri 2008 Diabetes Prevention in a Faith-based setting 26 Group-based, churches NA 6, 12 months -5.6 lb, -1.0 lb -6.4 mg/dl Ackerman 2008 DEPLOY 92 Group-based, YMCA Brief counseling 6, 12 months -6%, -2% -0.1% HbA1c, NS McBride 2008 ALL 37 Group-based, cardiac rehabilitation NA 3, 12 months -5.0 kg, -4.3 kg NR Kramer 2009 Group Lifestyle Balance 93 Group-based, primary care NA 3, 12 months -4.9%, -4.5% -1.5 mg/dl NS Whittemore 2009 DPP NP 58 Group-based, nurse practitioners Enhanced standard care 9.3 months -5%, -11% NR McTigue 2009 WiLLoW 166 Group-based, primary care NA 10-14 months -5.19 kg NR Amundson 2009 MT Diabetes Control Program 355 Group-based, health- care facilities NA 4 months -6.7 kg (6.7%) NR
19. HELP PD Key Innovations • Expanded the DPP lifestyle model to be community-located, group-based, and peer-led • Comparison group • Number of participants • Measurement of weight AND glucose as outcomes • Length of follow-up
20. Study Hypothesis • A lifestyle intervention (addressing healthy eating, physical activity, and weight loss) administered through a community-based diabetes prevention program model will have a beneficial and clinically meaningful impact on glucose and insulin metabolism, and markers of the metabolic syndrome.
21. HELP PD Study Design • Randomized trial of a community based translation of DPP • Testing a group-based behavioral lifestyle change strategy versus usual care • Fasting glucose is the primary outcome; assessment visits every six months • Intervention delivered through a local Diabetes Care Center with RDs and Community Health Workers (CHWs) Katula J. et al. Contemp Clin Trials. 2009 Sep 13.
22. Changes in Body Weight (kg) 93.02 91.55 90.93 90.93 92.23 94.38 87.14 87.44 88.33 88.81 82 84 86 88 90 92 94 96 Baseline 6 Months 12 Months 18 Months 24 Months Weight (kilograms) Length of Follow-Up UC LWL Katula J. et al. Am J Prev Med. 2013 April. 44 (Suppl 4).
23. Changes in Body Weight (%) 0 -1.18 -1.33 -0.93 -0.57 0 -7.52 -7.21 -5.82 -5.39 -8 -7 -6 -5 -4 -3 -2 -1 0 1 Baseline 6 Months 12 Months 18 Months 24 Months Weight Loss (%) Length of Follow-Up UC LWL Katula J. et al. Am J Prev Med. 2013 April. 44 (Suppl 4).
24. Changes in Fasting Glucose (mg/dL) 105.71 106.97 104.16 107.42 107.6 105.37 101.69 101.11 102.7 103.28 98 100 102 104 106 108 110 Baseline 6 Months 12 Months 18 Months 24 Months Fasting Plasma Glucose (mg/dL) Length of Follow-Up UC LWL Katula J. et al. Am J Prev Med. 2013 April. 44 (Suppl 4).
25. Summary of Results • Clinically significant reductions in body weight, BMI, waist circumference, and fasting blood glucose achieved during the first year • These results largely maintained in the second year as compared to the enhanced usual care condition • All achieved with lay community health workers and community based systems with high potential for dissemination
26. Comparing HELP PD to DPP • Were the two studies conducted in populations that were similar enough that we can even compare the results? • Results demonstrate that HELP PD was effective, but was it as effective as DPP? • Could HELP PD be a less expensive/more sustainable model than DPP for large-scale dissemination efforts?
27. Data Sharing in HELP PD • Requested limited DPP data-set from NIDDK Central Repository • Goals: • compare the baseline characteristics of HELP PD and DPP participants • contrast the magnitude of relative effects observed in the two trials • compare the cost effectiveness of the two interventions
28. NIDDK Central Repository • Established in 2003 • Access is granted for one-year time periods (can be renewed) • Three divisions: • Biosample Repository (Rockville, MD) • Genetics Repository (Piscataway, NJ) • Data Repository (Calverton, MD)
29. Baseline Characteristics Characteristic HELP PD N=301 DPP N=3665 Age, mean + SD 57.9 ± 9.5 50.6 ± 10.4 Sex, n (%) Male 128 (42.5) 1228 (33.5) Female 173 (57.5) 2437 (66.5) Race/Ethnicity, n(%) Caucasian/White 220 (73.3) 2117 (57.8) African American/Black 74 (24.7) 751 (20.5) Hispanic (of any race) 4 (1.3) 609 (16.6) Mixed/Other/Unknown 2 (0.7) 188 (5.1) Education, n (%) < High School 8 (2.7) 1902 (51.9) High School or Equivalent 53 (17.6) 760 (20.7) > High School 217 (72.1) 1003 (27.4) Other 23 (7.6) 0 Fasting Glucose (mg/dL), mean + SD 105.5 ± 11.3 107.33 ± 7.72 Fasting Insulin (uIu/mL), mean + SD 16.7 ± 9.8 26.4 ± 14.8 (n=2662) BMI (kg/m2), mean + SD 32.7 ± 4.0 33.5 ± 5.8 Blood Pressure (mmHg), mean + SD Systolic 127.2 + 14.1 124.2 ± 14.6 Diastolic 73.2 + 9.4 78.6 ± 9.3 Waist Circumference (cm), mean + SD 104.7 ± 10.0 104.9 ± 14.5 (n=2662) Blackwell C. et al. Contemp Clin Trials. 2011; 32(1): 40-9.
30. Cost of HELP PD • Per capita direct medical cost of the HELP PD Lifestyle Intervention for 2 years: $850 • Per capita direct medical cost of the DPP Lifestyle Intervention for 2 years: $2,631 • No other translations of DPP to date have done a comparable economic analysis Lawlor M. et al. Am J Prev Med. 2013 April. 44 (Suppl 4).
31. Intervention Comparison: Weight Loss 0 10 20 30 40 50 60 70 80 90 100 gain lost > 0% lost > 5% lost > 7% lost > 10% lost > 15% HELP DPP
32. Benefits of Data Sharing • Enabled head-to-head comparison • Provides compelling evidence that could impact future large scale programming/implementation and policy making (National Diabetes Prevention Program)
33. Wake Forest Baptist Medical Center Large Multi-Center Clinical Trials The Action to Control Cardiovascular Risk in Diabetes (ACCORD) Trial
34. Background on Diabetes and CVD • Type 2 diabetics die of cardiovascular disease (CVD) events at rates 2 to 4 times higher than people who do not have diabetes • Risk factors for CVD events: • Glycemia (blood sugar control) • Lipids (HDL and TG) • Systolic blood pressure
35. Wake Forest Baptist Medical Center 35 Overall Goal for ACCORD To test three complementary medical treatment strategies for type 2 diabetes to enhance options for reducing the very high rate of major CVD morbidity and mortality
36. Study Hypotheses 36 In middle-aged or older people with type 2 diabetes who are at high risk for having a CVD event because of existing clinical or subclinical CVD or CVD risk factors: 1. Does a therapeutic strategy that targets a HbA1c of < 6.0% reduce the rate of CVD events more than a strategy that targets a HbA1c of 7.0% to 7.9% (with the expectation of achieving a median level of 7.5%) ? 2. In the context of good glycemic control, does a therapeutic strategy that uses a fibrate to raise HDL- C/lower triglyceride levels and uses a statin for treatment of LDL-C reduce the rate of CVD events compared to a strategy that only uses a statin for treatment of LDL-C? 3. In the context of good glycemic control, does a therapeutic strategy that targets a systolic blood pressure (SBP) of < 120 mm Hg reduce the rate of CVD events compared to a strategy that targets a SBP of < 140 mm Hg?
37. 37 ACCORD Design • Multi-center design- 77 sites in US and Canada • 10,251 participants • Double 2 X 2 factorial design: participants randomized to either intensive or standard glycemia and: – Intensive or standard blood pressure – Blinded lipid trial
38. ACCORD Outcome Measures • Primary outcome measure was first occurrence of a major cardiovascular disease event (nonfatal MI or stroke or cardiovascular death) • MIs, strokes, and deaths adjudicated by a committee masked to treatment assignment • Other outcomes include total mortality, microvascular outcomes, HRQL, and cost- effectiveness • Sub-studies on diabetic retinopathy (Eye) and cognitive decline (MIND)
39. ACCORD: Results • The glycemia trial was terminated early due to higher mortality rates in the intensive compared with the standard glycemia treatment strategies • Treatment for all participants was stopped on June 30, 2009 and transitioned to their own physicians • Participants still being followed in a non- treatment observational study (ACCORDION)
40. ACCORD: Results • Higher rate of all deaths and CVD deaths in the intensive group – this was unexpected. There was a lower rate of non-fatal heart attacks in the intensive group. Overall fewer people died in both glycemia groups than expected. • There was more hypoglycemia, more weight gain & more adverse events in the intensive group – but these do not explain the higher risk of death. • Bottom line – the standard treatment was safer than the intensive treatment in ACCORD.
41. Data Sharing in ACCORD • Requests for ancillaries studies ongoing • Pooled analyses that have included data from the ACCORD datasets have provided important information on the safety of available diabetes and cardiovascular drugs • Ancillary studies have also provided significant information on the relationships between the management of diabetes and eye disease, cognitive decline, and bone health
42. Wake Forest Baptist Medical Center Industry Trials The Almased Multi-Center Diabetes Intervention Trial (AMDIT)
43. Weight-Loss, Meal Replacements, and Diabetes • Weight-loss can reduce hemoglobin A1c (blood sugar) in adults with diabetes • Meal replacements can be effective for weight loss in diabetics • Soy protein has demonstrated positive influences on body composition, metabolic risk factors, and lipids
44. AMDIT Study Overview • Funding provided by makers of Almased, an all natural soy-based meal replacement product • 5 international sites • USA • Germany (Coordinating Center) • Brazil • United Kingdom • India
45. AMDIT Study Objectives • Evaluate the effectiveness of a soy-based meal replacement on glycemic control and metabolic effects in patients with type 2 diabetes over a period of 12 months in different populations • Primary outcome variable is HbA1c • Other outcomes include fasting blood glucose, fasting insulin level and insulin resistance (HOMA), risk factors for cardiovascular disease, and diabetes-related metabolic factors and complications.
46. AMDIT Study Design • 48 participants at each site- randomized to receive either group-based lifestyle intervention or meal replacement regimen • 2:1 randomization scheme • Participants followed for one year; seen every 2 months
47. Opportunities for Data Sharing • Current trends in diabetes management that favor use of Rx interventions to reduce blood sugar levels and manage disease- paradigm shift to look at non-pharmacological treatments • Comparison with other trials that have targeted weight loss (LookAhead)
48. Summary: Why Data Sharing Matters • Comparison of results between trials- can impact policy and future funding • Validate results across populations • Determine trends or patterns that may be undetectable in smaller samples- safety of drugs or interventions
49. What can you do? • Become familiar with resources available- both to access data and contribute to repositories • Serve as a contact point for faculty and staff as they access these resources