Impact of artificial intelligence in operations

Abstract

India’s rural healthcare system continues to face critical challenges, particularly limited infrastructure and a shortage of medical professionals. While Community Health Workers (CHWs) play a vital role in bridging these gaps, the absence of direct performance data hinders evaluation. This study addresses that limitation by using patient consultation data as a proxy to assess CHW effectiveness in maternal healthcare. Through statistical and machine learning analyses (Factor analysis, clustering, ANOVA, Cramér’s V, correlation) and machine learning models (Random Forest, Logistic Regression, XGBoost), key CHW activities—such as visit continuity, counseling and assistance frequency—were analyzed for their impact on patient outcomes. Results reveal strong associations between these CHW activities and improved maternal health indicators, including stable vitals and reduced complications. Predictive models further confirm the impact of these CHW activities on maternal health, achieving high accuracy and recall in identifying high-risk patients. These findings support a scalable, data-driven framework for evaluating CHW performance where traditional supervision data is lacking. By transforming routine health records into targeted strategies—such as prioritizing follow-ups for high-risk patients, reinforcing counseling interventions, and improving access to financial aid—this study contributes a novel, scalable method for CHW evaluation, combines predictive analytics with field-relevant features, and provides actionable pathways to inform policy and improve maternal health outcomes.