Pulled from the Infectious Economics archives, these articles provide recommendations for the design and development of robust mathematical models of COVID-19 epidemiology and economics to reduce uncertainty in decision-making and inform evidence-based public policy.
Associations are easy; valid causal inference is hard
- Tutorial: how to estimate person-centered treatment effects so that public policy to control an infectious disease does not make things worse for the poor
- Instrumental variables are an econometric tool that can be used to strengthen causal inference in observational studies of COVID trying to estimate the effectiveness of policies
- Propensity scores can reduce bias in studies that compare the effectiveness of treatments for COVID during hospitalization
- Interrupted time series and difference-in-differences analyses are strong approaches for the design of quasi-experimental studies of the impact of national policy on the control of an infectious disease
If we want to be honest in evaluating the winners and losers in healthcare policies, then we argue more people and disciplines should learn and use these methods if the necessary conditions are satisfied. Failure to consider these methods could result in unintended consequences and exacerbate existing inequalities in health between patients who are “average” and “outliers.”
Patient-centered treatment effects, March 2018
- Approach to validating forecasts, enhancing forecasts with more recent supplemental data from Google Trends, and correlation between Google Trends and uptake of preventative drug infectious disease
- Visualization of progress toward infectious disease goals with vector representing trajectory over time
- Financial incentives to individuals for preventative behaviors can reduce the spread of infectious disease, a cost-effective way to maximize the spillover benefit to society
…my stepwise approach to predictive validation. We fit the model with outcomes data only up to a certain point in time (I data up to 36, 48, and 60 months) and then predict outcomes each month in the following 12 months. As you follow the panels from left to right, you can see how closely the predictions match the real observed data…
How to validate forecasts, Nov 2017
Enjoy the throwback posts. Random skills in forecasting epidemics, math modeling infectious diseases, causal inference in observational studies of policy effectiveness, and decision analysis are urgently needed to fight through this pandemic together and make smart decisions along the way.
Update 5/4/23
By popular demand, I’ve added my UW HSERV 525 Advanced Methods Final Exam Cheat Sheet that was a well-used reference for me in the first few years analyzing EHR data.
