We’re pleased to share the latest in our series of posts from our interns. Daniel Ohrenstein has just completed an extended 5-month stay with us working on some exciting new techniques for multiple risk factor modelling.
I recently completed an MSc in Machine Learning and Data Science at UCL, having graduated with a BA in Natural Sciences from the University of Cambridge in 2019. Working for HealthLumen as a statistical modelling and machine learning intern has been a fantastic way to gain some experience in applying statistical and analytical techniques to real-world public health data. During my time here, I have been involved with various interesting projects, authored and co-authored a number of academic papers, and hopefully made a small contribution to the incredible research the team are doing at the cutting edge of public health modelling.
One project which Dr Lise Retat (Senior Health Economist & Mathematical Modeller) and I have worked on for the last few months is the development of multiple risk factor models for non-communicable diseases (NCDs) using the copula method. Our research (awaiting peer-review) suggests that it is possible to accurately estimate joint relative risks for NCDs e.g. the risk of developing coronary heart disease while being both obese and having a high alcohol consumption relative to a tee-total person of healthy body mass index (BMI) using only cross-sectional data, without the need for an expensive and time-consuming longitudinal study (which tracks subjects over a number of years). This result has exciting implications for mathematical models which use these joint relative risks as parameters for predicting future incidence of NCDs, as interference effects between the risk factors can be much more accurately represented.
Along the way, I have had the pleasure of working with some truly outstanding colleagues, with skills and backgrounds ranging from machine learning engineers to public health specialists. As such, expert advice is never too far away at HealthLumen! The team is remarkably close, with regular check-ins, news updates and occasional meme-sharing. My five months of working at HealthLumen have been hugely enjoyable and I wish the whole team the very best for the future.
If you are interested in learning more about working with HealthLumen, please see our Team page to learn more.