Applying Risk Based Approaches to Portfolio Management

With the recent recommendations and formal guidelines from regulatory bodies and a matured KRI library from Transclerate for Risk based approaches to Trial Monitoring, there has been a rise in development and application of Risk based approach to Trial Monitoring.

I present an evidence-based application of how we have broadened a risk-based approach to manage a portfolio of studies and in some cases proactively identify risk at a Portfolio, Study, Region, Investigator and Site levels

Some of the nuances we will go through the paper will include:

While discussing specifics on technology and how one could build and implement such a system, we will go through some business process one must go through to enable a system of this capability at scale:

1. Can we build a one size fits all risk model for any sponsor or therapeutic area?
2. Will this model need to be configurable? (If so, at which level?)
3. As a bonus, we will also look at how one could setup Risk based Performance alerts without having to rely on hard-coded thresholds

Transcript

1. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 1 ©2019 Comprehend. Confidential. All Rights Reserved. | Page 1 Applying Risk Based Approaches to Portfolio Management Chandi Kodthiwada Sr. Product Manager
2. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 2 How is everybody doing?
3. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 3 Emotional Overview Dashboard Source: Beehance
4. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 4 Why Change Status Quo processes? Lack of Timely Data Increases Trial Timeline Enrollment data fragmented across multiple vendors and systems Reporting provided from the CRO monthly No insight into enrollment issues Inability to take timely action when enrollment went off track Unnecessary 5 month Trial Delay Why: Lacked timely visibility that issue was due to strict inclusion criteria Disjointed Deficient Delayed
5. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 5 Event Driven Analytics meets Portfolio Management
6. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 6 Event Driven Analytics * All brand names are trademarks of their respective organizations – used for illustrative purposes only
7. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 7 Event Driven Analytics - What next?
8. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 8 ©2019 Comprehend. Confidential. All Rights Reserved. | Page 8 How do we identify “Events of Interest” ? By Applying Risk Based Approaches
9. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 9 Events of Interest
10. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 10 Risk Model – application across Entities
11. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 11 ©2019 Comprehend. Confidential. All Rights Reserved. | Page 11 Risk Model - Configurability
12. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 12 Risk Model Configurability at KPI
13. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 13 Quality Agreement Risk Model Configurability at Study A Study with above threshold settings for Screen Failures will classify the performance of Site as below: • A Site with <=10% Screen Failures shows as Green/low(No)-Risk • A site with >10% but <=20% Screen Failures classified as Yellow/Moderate-Risk • Any Site with >20% Screen Failure appears as Red/Risky
14. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 14 Risk Model Configurability – Overall Risk Scores Overall Risk at Study = 1 234 5 (Weight i ∗ Risk Element i )/n Model supports any number of Key Risk Indicators and can be tailored to Study-level or deployed as a standard enterprise model • Risk Element = Key Performance Indicator • Weights can be used to influence Overall risk calculation and make a specific KRI more risk-averse compared to others
15. ©2019 Comprehend. Confidential. All Rights Reserved. | Page 15 In Summary… Chandi Kodthiwada Sr. Product Manager As Analytics Leaders, it is no longer acceptable to stop at Dashboards: Figure out how to “inform” your business and facilitate workflows Look at Industry Standards and Thresholds to identify which performance indicators to monitor automatically & at what thresholds Adopt platforms that let you seamlessly centralize data, generate insights and fold them into your business’ daily workflows One Risk Model can’t “fit for all” – Whatever model you adopt, make room for customization at various levels (ex: TA, Study, Site, CRO, Region, Country)