Predicting the Individual Treatment Effect of Neurosurgery for Patients with Traumatic Brain Injury in the Low-Resource Setting: A Machine Learning Approach in Uganda
Written by: Adil SM, Elahi C, Gramer R, Spears CA, Fuller AT, Haglund MM, Dunn TW
DGNN’s communications team recently sat down with Adil Syed to talk about their recent publication in the Journal of Neurotrauma. We share a few relevant details below:
DGNN: What are the main takeaways? (What were the key results or conclusions from the research?)
Adil: We provide the first machine learning-based prediction of the impact of surgery for TBI patients in low and middle income countries (LMICs). Results suggest that current patient selection for surgery may be suboptimal, and a tool such as ours may enable better decision making and subsequently more beneficial neurosurgical interventions.
DGNN: What does this research mean for the scientific community?
Adil: Even in LMICs where data collection is less robust, there is enough data to create accurate machine learning models for TBI prognostication. These models are even more essential in the low-resource settings where questions of triage are key. Specifically, calculating “individual treatment effects” may serve as an important way to select TBI patients for neurosurgery. This lays the groundwork for deeper efforts in such causal inference challenges.
DGNN: What does this research mean for the common man? (Why should we care about your work?)
Adil: If we can continue to refine tools such as that presented here, and they are adopted by the medical community in LMICs, we can better allocate limited resources and thus enhance the benefit of neurosurgical intervention after TBI. We aim to ensure that those patients who would most benefit from neurosurgery do indeed receive it.
If you would like to read the whole article, follow this link.