Rajneesh Patil, Senior Director, RBM and Analytics, QuintilesIMS01.23.17
Over the past few years, Risk-Based Monitoring (RBM) – in which sponsors allocate monitoring resources based on the level of risk identified for a specific trial or site – has emerged as a more efficient way to facilitate trial delivery. RBM strategies optimize data flows, enhance data quality and improve oversight and control of clinical trial sites. And the benefits of RBM strategies have only just begun.
At the annual SCOPE Summit for Clinical Ops Executives, Rajneesh Patil, senior director, RBM and Analytics at QuintilesIMS is giving a presentation on Implementing Risk-Based Monitoring that will explore the future in this area, and how predicative analytics and machine learning are enabling a more holistic and proactive approach to identify and mitigate risk.
Rajneesh Patil of QuintilesIMS
Learning from the past
RBM relies on data and analytics to assess site performance, identify potential outliers and determine an appropriate course of action. However, the early generations of these tools focus on single parameters to identify issues, which, out of context, can often Lead to false positives. For example, a site may show that 80 percent of patients have missed a dosing event, but if the site only has five patients with varying recruitment dates, the ‘outlier status’ is not as alarming as a site with 50 patients and the same 80 percent of missed rate. The advanced analytics models can differentiate this relative risk between the two sites and indicate that the latter site in this example is at higher risk.
As a result, Central Monitors (CMs) and Clinical Research Associates (CRAs) spend a Lot of their time responding to triggers that may not require much action. It’s still an improvement over pre-RBM days, when CMs/CRAs split their time evenly between every site looking for anomalies that often didn’t exist -- but we can do better. Indeed, we are doing better.
Advanced and predictive analytics tools can help normalize these observations by taking a holistic view of the site data. These tools are able to evaluate multiple factors to identify whether an anomaly is a true risk or simply an outcome of other factors, such as exposure time or enrollment numbers.
This is achieved through machine learning approach, which uses algorithms and decision support models to learn from past searches, making the system ‘smarter’ with every review. This learning isn’t just limited to the current trial. These platforms will be able to learn from historic clinical data sets, making them increasingly ‘knowledgeable’ with every application and resource. That translates to greater oversight, higher quality data and a safer trial environment. It also means sponsors can utilize clinical research staff more efficiently and allow them to focus on more value-added tasks such as ensuring sites are operating within trial parameters and Compliance measures.
No more false positives
We are already implementing these tools with positive results. In one review of a cardiovascular outcome study, our predictive analytics identified two sites at risk for compliance problems due to eligibility concerns. We drilled down into the data set, and found that while patients were meeting the criteria, several subjects were borderline eligible and there was too little variation among all of the patients enrolled. Based on these results we went back to the medical monitor who reviewed inclusion criteria and re-educated site leaders on eligibility requirements and recruiting strategies. By identifying this anomaly early, they were able to address the problem before it impacted study results. If this anomaly had gone unchecked, those patients might have completed the trial only to be excluded from submission data, and/or thrown all of the study results into question.
In another evaluation of four large studies, we found that using advanced and predictive analytics as part of an RBM program cut false positives by a substantial amount. These positive results suggest that in the very near future, these models will become a baseline best practice for trial monitoring.
The use of advanced and predictive analytics in RBM is the start of a major change in our industry, helping get ahead of safety issues and operational challenges. RBM adoption will only increase as data and technology continue to evolve in the coming years. The sooner we begin implementing this technology, the more knowledge we can harness for the good of sponsors, sites and ultimately, patients.
Rajneesh Pail is senior director, Risk-Based Monitoring (RBM) and Analytics at QuintilesIMS. In this role, he leads the optimization of the Centralized Monitoring processes and analytics capabilities. Patil has deep expertise spanning RBM, program management, continuous process improvement, clinical data management and predictive analytics for Clinical Trials . Patil holds a Bachelor’s degree in Dental Surgery from India and Foundations in Public Health from Australia. He is a certified Six Sigma Greenbelt with an MBA in Operations Management and is a certified Project Management Professional ( PMP ).
At the annual SCOPE Summit for Clinical Ops Executives, Rajneesh Patil, senior director, RBM and Analytics at QuintilesIMS is giving a presentation on Implementing Risk-Based Monitoring that will explore the future in this area, and how predicative analytics and machine learning are enabling a more holistic and proactive approach to identify and mitigate risk.
Rajneesh Patil of QuintilesIMS
RBM relies on data and analytics to assess site performance, identify potential outliers and determine an appropriate course of action. However, the early generations of these tools focus on single parameters to identify issues, which, out of context, can often Lead to false positives. For example, a site may show that 80 percent of patients have missed a dosing event, but if the site only has five patients with varying recruitment dates, the ‘outlier status’ is not as alarming as a site with 50 patients and the same 80 percent of missed rate. The advanced analytics models can differentiate this relative risk between the two sites and indicate that the latter site in this example is at higher risk.
As a result, Central Monitors (CMs) and Clinical Research Associates (CRAs) spend a Lot of their time responding to triggers that may not require much action. It’s still an improvement over pre-RBM days, when CMs/CRAs split their time evenly between every site looking for anomalies that often didn’t exist -- but we can do better. Indeed, we are doing better.
Advanced and predictive analytics tools can help normalize these observations by taking a holistic view of the site data. These tools are able to evaluate multiple factors to identify whether an anomaly is a true risk or simply an outcome of other factors, such as exposure time or enrollment numbers.
This is achieved through machine learning approach, which uses algorithms and decision support models to learn from past searches, making the system ‘smarter’ with every review. This learning isn’t just limited to the current trial. These platforms will be able to learn from historic clinical data sets, making them increasingly ‘knowledgeable’ with every application and resource. That translates to greater oversight, higher quality data and a safer trial environment. It also means sponsors can utilize clinical research staff more efficiently and allow them to focus on more value-added tasks such as ensuring sites are operating within trial parameters and Compliance measures.
No more false positives
We are already implementing these tools with positive results. In one review of a cardiovascular outcome study, our predictive analytics identified two sites at risk for compliance problems due to eligibility concerns. We drilled down into the data set, and found that while patients were meeting the criteria, several subjects were borderline eligible and there was too little variation among all of the patients enrolled. Based on these results we went back to the medical monitor who reviewed inclusion criteria and re-educated site leaders on eligibility requirements and recruiting strategies. By identifying this anomaly early, they were able to address the problem before it impacted study results. If this anomaly had gone unchecked, those patients might have completed the trial only to be excluded from submission data, and/or thrown all of the study results into question.
In another evaluation of four large studies, we found that using advanced and predictive analytics as part of an RBM program cut false positives by a substantial amount. These positive results suggest that in the very near future, these models will become a baseline best practice for trial monitoring.
The use of advanced and predictive analytics in RBM is the start of a major change in our industry, helping get ahead of safety issues and operational challenges. RBM adoption will only increase as data and technology continue to evolve in the coming years. The sooner we begin implementing this technology, the more knowledge we can harness for the good of sponsors, sites and ultimately, patients.
Rajneesh Pail is senior director, Risk-Based Monitoring (RBM) and Analytics at QuintilesIMS. In this role, he leads the optimization of the Centralized Monitoring processes and analytics capabilities. Patil has deep expertise spanning RBM, program management, continuous process improvement, clinical data management and predictive analytics for Clinical Trials . Patil holds a Bachelor’s degree in Dental Surgery from India and Foundations in Public Health from Australia. He is a certified Six Sigma Greenbelt with an MBA in Operations Management and is a certified Project Management Professional ( PMP ).
