Data/AI + Healthcare
My work is organized under two main streams.
How do AI or data-driven recommendations shape decisions?
It seems like providing more information, like explanations for the AI recommendations, should improve decisions. We find that’s not always the case. I employ mixed methods combining qualitative (interviews, case studies) and quantitative (experiment, retrospective data analysis, meta-analysis) approaches.
Real-world data provides important, timely feedback about healthcare decisions and policy.
We can use data, like pharmaceutical claims data, to look at the impact of regulatory actions (like FDA letters), policy changes, or events (like COVID-19) on patient outcomes like drug adherence. Similarly, reviewing patient records helps with EMS resource allocation. Modern causal inference techniques (difference-in-differences, regression discontinuity) allow us to draw conclusions from real-world data.
Assistant Professor, Information Systems
Augsburg University
Research
I strive to do impactful research yielding actionable insights for data in healthcare.
Were patients able to access medication after Covid-19 hit?
Medication access and adherence are crucial to health outcomes. Using a large dataset covering 93% of the US retail market, we found increased discontinuation of some (but not all) medications and a likely drop in new patients accessing critical therapies.
How can organ transplant teams move toward Meaningful AI use?
Through a systematic literature review, we assess the state-of-the-art in transplant AI. We identify the key roadblocks to implementation and how transplant clinicians can start to address them.
How do explanations for AI recommendations impact clinician decisions?
“Explainable AI” is a hot topic. It seems like providing an explanation for black box AI systems should increase use of AI recommendations and improve clinical decision quality, but we find that’s not necessarily the case.
Recent Posts
I explore topics in data analysis, statistics and data visualization. Putting together a brief, intuitive explanation usually forces me to understand the concepts better.