Quantifying Regional Variation in Ivermectin Use During the Pandemic using Regularized Synthetic Controls

Abstract: We analyze weekly U.S. prescription claims data over a 36-month period spanning the COVID-19 pandemic to investigate patterns of overconsumption of the antiparasitic drug Ivermectin (IVM). To quantify the overconsumption of IVM use following the heightened public attention on IVM as a candidate treatment for COVID-19, we adopt a causal framework based on synthetic controls, comparing IVM prescription trends to those of a large set of control medications. We employ a regularized synthetic control (RSC) method using continuous spike-and-slab shrinkage priors to estimate state-level deviations in IVM consumption. This approach offers decision-theoretic guarantees on predictive risk, supporting its application in downstream policy analysis at multiple-time points after the intervention. Its empirical robustness is demonstrated through extensive validation checks. Our findings reveal a modest increase in IVM prescriptions following early reports of its potential therapeutic use, with no significant surge over the subsequent eight months. This was followed by a pronounced increase that coincided with the peak in COVID-19 cases. Strikingly, elevated IVM use persisted even after COVID-19 vaccines became widely available and multiple federal countermeasures were implemented. Furthermore, we find that state-level political affiliation significantly explains variation in overconsumption, even after accounting for COVID-19 incidence rates. These results highlight deep regional disparities in the effectiveness of public health messaging and suggest a need for more targeted and trusted communication strategies.

Website: https://gmukherjee.github.io/