Identifying what we have learned when implementing risk-based monitoring

Transcription

1 Identifying what we have learned when implementing risk-based monitoring Outsourcing in Clinical Trials - West Coast 2017 Neil Vivian Senior Director of Business Solutions February, 2017 The EDC of choice for over 5,000 clinical trials around the world

2 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 2

3 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 3

4 Recap of Guidance FDA believes that risk-based monitoring could improve sponsor oversight of clinical investigations. Risk-based monitoring, including the appropriate use of centralized monitoring can meet statutory and regulatory requirements. There may be minimal benefit in comparing 100% of the source data for each subject to the CRFs for each study visit. Rather, it may be sufficient to compare the most critical data points for a sample of subjects and study visits as an indicator of data accuracy. Verify critical source data remotely in cases where such source data are accessible. 4

5 And the Survey says? BVMA Survey Overview Results of a survey that the BVMA (German Federation of CROs) initiated among the member companies of the EURCROF (the European Federation of CROs) following a Workshop on RBM. 180 respondents completed questionnaires between August 2014 and January all were analyzable Distribution across CROs - Large (71), Mid-Size (47), Small (48), Other (14) OmniComm is a member of the BVMA with active participation in their RBM working group. 5

6 Selected Survey Results 6

7 What is SDV s I pa t o Data Qualit? SDV Has Minimal Impact on Site-Entered Data 3.7% DATA THAT IS EVER CORRECTED 1.1% All Site-Entered Data Impacted By SDV Source: DIA Journal, Therapeutic Innovation & Regulatory Science, November

8 SDV More Rele a t for AE Reporti g 15% 11,8% 12% 9% 7,5% <= 1 day <= 7 days 6% 3,6% 3% 1,7% 0% AE SAE Source: DIA Journal, Therapeutic Innovation & Regulatory Science, November 2014 ~7% to 12% of AEs Entered Following SDV 8

9 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 9

10 Selecting data for SDV The guidance allows <100% SDV as part of a risk-based monitoring plan Should take into account Site Risk Rating Criticality of Data (e.g., Safety, Primary Endpoint) Triggers to change Monitoring Level (e.g., Serious Adverse Event) Balance of on-site and central monitoring Should be automatically managed in EDC If the system picks the data to SDV, it is more defensible than if people pick the data to SDV 10

11 System Functionality Sites Risk Rating Randomize Patients Monitoring Level All Forms None Critical All Critical All Critical All Critical 11

12 Notes on System Functionality Mo itori g Le el i itiall set usi g a stratified, lo k-based randomization Mo itori g Le el is a piece of clinical data May be hidden using standard security techniques May be the target of an edit check E.g., If patie t e perie es a SAE, ha ge alue to All If manually updated, requires a reason for change Only selected forms need to be monitored, others can be signed when complete (no open queries) 12

13 Additional DM Reports to Track Monitor Activity 13

14 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 14

15 Statistics (1) Item Number of Production Trials analyzed Number of Patients in each Trial Number of Monitoring Levels Average duration to date Value (except one with 20) ~1 year 15

16 Statistics (2) Risk Ratings For 11 of the 12 trials, either there is no site risk rating or the risk ratings are initially the same for all sites For one trial there is a distribution of sites across four risk ratings 16

17 Statistics (3) SDV Coverage Trial # Patients % Forms for SDV Trial A Trial B Trial C Trial D Trial E Trial F Trial G Trial H Trial I Trial J Trial K Trial L Weighted Average = 40% On average, an additional 5% of forms were monitored even though they didn t need to be 17

18 Setting Monitoring Level Many trials set first two patients at each site to 100% SDV Many trials have different monitoring level selections for Screen Failures Example reasons for changing the monitoring level: SAE for the patient Early termination Device implant performed Dose escalation 18

19 Summary of Risk-Based Monitoring Statistics Risk-Based Monitoring is now in mainstream, production use If ou re still o deri g a out it, ou re ehi d our peers The same techniques used for treatment randomization may be used for randomizing the monitoring levels Standard practice is to set all site risk ratings the same Can modify over time based on experience Target SDV coverage is around 40-50% 19

20 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 20

21 A little digression how do you submit expenses? 21

22 Scan the paper document 22

23 Attach to the expense record 23

24 If it orks for e pe se reporti g Wh a t e do the same for clinical data? You re ad! It s a iolatio of patient privacy The sites o t do it The sites do t have scanners 24

25 Example Informed Consent Form Patient name and Signature blacked out Date and Investigator information can be monitored remotely 25

26 Entry and Scanning Process 1 Scan 3 Attach PDF to CRF Confirm redaction 2 Enter data into EDC DEMOG DEMOG Demog IC DEMOG Informed Consent 4 Monitor centrally I/E 26

27 Program Agenda Guidance for risk-based monitoring a recap Dynamically selecting data for SDV overview of functionality Some statistics from production trials Unpicking the rationale and process in remote monitoring Sponsor experience and observations 27

28 Informed Consent Implementation - Statistics Item Study Phase Number of sites Number of patients Number of Informed Consent Versions Study duration to date Value IV ~ 2 years 28

29 Observations from the Sponsor I the egi i g the sites fou d it hard, ut it e a e eas after a fe ti es. It feels like atta hi g do u e ts to a e ail. It akes it er eas for the o itors to he k all poi ts fro e ept for the patie t s sig ature a d a e. the ICF The i e thi g here is that e a alread sig al issues ith the ICF efore e e er go ph si all to the site. Most sites e isit earl. Duri g o itori g e e er fou d the o plete ICF issi g. 29

30 Summary Attaching redacted source documents to the clinical data can significantly aid a o itor s ork Sites can be motivated to participate It s ot i possi le! 30

31 31