Moving from the bedside to the desktop: Dr. Joshua Grill on employing analytics to further Alzheimer's disease clinical trial designs

Written by Mary Rechtoris | April 05, 2017 | Print  |

Researchers spanning the nation are working to employ analytics to drive real changes for patients with various conditions. Joshua Grill, PhD, co-director of the Memory Impairments Neurological Disorders Institute at University of California Irvine, details how he and fellow researchers are using analytics systems for clinical trials on Alzheimer's disease research and how data may shape trials moving forward.

Question: What makes analytics actionable to physicians?

 

Dr. Joshua Grill: Medicine is becoming increasingly reliant on information and analytics. Some refer to this as a transition from the bedside to the desktop. In the research setting, analytics has long been essential, not only to assessing clinical trial data to decide if a new treatment achieves necessary standards for improved safety and efficacy, but also to take as sophisticated approach as possible to study design.


 
Q: Do you have programs that use analytics to improve clinical decision making?

 

JG: We have been focused on using analytics to instruct Alzheimer's disease clinical trial designs, specifically as they pertain to considering who should be enrolled, the quality of the data they can expect to produce and how changing the criteria for participation may change the overall trial expectations as they relate to data integrity. It is possible that by limiting trials to specific populations — potentially making it more difficult to find appropriate participants — we could simultaneously reduce the overall number of trial participants needed to test a study hypothesis, potentially making it easier and faster to complete a study. Clearly, there is a need to find the optimal balance.  


 
Q: What are some of the challenges in using analytics?
 

JG: In Alzheimer's disease research, we are fortunate that a number of very large, well-designed studies for which the entirety of study data are publicly available for researchers to perform additional analytic research. These include the public-private partnership known as the Alzheimer's Disease Neuroimaging Initiative. While these databases are tremendous tools, their designs don't entirely replicate trial designs, they don't include treatment and placebo groups and we don't know if the people who enroll are equivalent to the people who enroll in actual trials. For these reasons, actual trial datasets provide the greatest validity in this type of analytic research, and fewer of those are currently publicly available.


 
Q: What progress have you made?


 
JG: We've been able to model trials based on natural history studies, like ADNI, and find interesting observations that we hope may be instructive to the designs of future trials. Specifically, it's clear that some patients in Alzheimer's disease clinical trials, specifically those who lack a spouse (Alzheimer's disease trials always require patient participants to enroll with another person known as a study partner), may be associated with more variable data. This associated variability reduces trial power, therefore requiring more participants in the study to demonstrate a drug effect. We're eager to further this work to identify the optimal trial design, which requires the fewest participants possible, ensuring high data integrity, but also being realistic in the ability to recruit quickly and test the studies' scientific hypothesis.

 

More articles on practice management:
Physicians increasingly lean toward single-payer healthcare system
10 statistics on how millennials use digital technology for healthcare
UBMD Neurology moves to new location: 3 notes

© Copyright ASC COMMUNICATIONS 2018. Interested in LINKING to or REPRINTING this content? View our policies here.

Top 40 Articles from the Past 6 Months