Ed Mills, ki’s trial services lead, held the first of what we expect to be many learning sessions at the foundation to help familiarize our MNCH-D&T colleagues with various aspects of data science so that they better understand how they can work with us to do their work most effectively.

Ed focused on how combining approaches used by pharmaceutical companies, harnessing new computational approaches, and involving end users in the design of clinical trials can make trials faster, cheaper, and better.

In global health, there tends to be a fair amount of uncertainty about the disease, the setting, and the interventions being investigated. As a result, it can be difficult to optimize RCT designs in the planning stage before patients are enrolled to the trial. Moreover, during the course of a trial, information may come to light that suggests that the trial could be less expensive, shorter, or otherwise different and better than originally planned.

However, since most RCTs hold their designs fixed, investigators are stuck spending more and waiting longer for results that will be less helpful than they could be. Ed explained what’s possible using data, modeling, and simulations to improve trial design and adapt designs during the trial based on preliminary evidence. See the handout from the session to learn more, and stay tuned for information about the next ki learning session.

Ki is currently supporting design of several trials in MNCH-DT. We are working towards a goal of simulating all trials in MNCH-DT before making a funding decision. What we learn should be readily scalable: the Gates Foundation spends hundreds of millions of dollars every year on trials, and by using data we can increase the quality of decision making on trial investments, optimize the efficiency of trials, and get to high-quality answers faster.