ML-Powered Triage and Queue Optimization for Resource-Constrained Free Clinics

Armaan Grewal
Proceedings of the 7th Conference on Health, Inference, and Learning, PMLR 333:905-923, 2026.

Abstract

Free clinics serve about 1.7 million uninsured Americans annually in the US, yet operate under severe resource constraints that lead to missed urgent cases, inefficient patient flow, and long waits. Our machine learning (ML)-powered triage and queue optimization system is designed as a decision support tool specifically for resource-constrained free clinics. Our system combines a Random Forest (RF) classifier trained on MIMIC-IV-ED data with a multi-objective queue optimization algorithm presented via a staff-facing interface. Our triage model, when simulating free clinic deployment without vital sign equipment, achieves an 83.6% critical case detection rate and no dangerous misses on the test set (0.012% on holdout); optimizing for patient safety over raw accuracy. Monte Carlo simulation across 1,000 clinic sessions demonstrates a 72% reduction in wait times ($p<0.001$) for critical patients compared to first-come-first-served (FCFS) queue ordering. Unlike commercial triage systems that can be prohibitively expensive, our solution is built entirely on free and free-tier tools and designed for volunteer-staffed environments lacking trained intake nurses, vital sign monitors, and electronic health records (EHR). We developed the system as a tablet-optimized web application with real-time queue updates and physician queue overrides. Our work is entirely retrospective evaluation and simulation. Prospective studies in free clinic settings, in collaboration with clinicians and domain experts, are planned as a next step.

Cite this Paper

BibTeX
@InProceedings{pmlr-v333-grewal26a, title = {ML-Powered Triage and Queue Optimization for Resource-Constrained Free Clinics}, author = {Grewal, Armaan}, booktitle = {Proceedings of the 7th Conference on Health, Inference, and Learning}, pages = {905--923}, year = {2026}, editor = {Healey, Elizabeth and Fries, Jason and Pollard, Tom and Tang, Shengpu and Zink, Anna and Hartvigsen, Tom and Agrawal, Monica and Finlayson, Sam and Glicksberg, Benjamin and Beaulieu-Jones, Brett and Wang, Kai and Fontalvo, Daseyra and Sarker, Tasmie and Chen, Irene and Alsentzer, Emily}, volume = {333}, series = {Proceedings of Machine Learning Research}, month = {29--30 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v333/main/assets/grewal26a/grewal26a.pdf}, url = {https://proceedings.mlr.press/v333/grewal26a.html}, abstract = {Free clinics serve about 1.7 million uninsured Americans annually in the US, yet operate under severe resource constraints that lead to missed urgent cases, inefficient patient flow, and long waits. Our machine learning (ML)-powered triage and queue optimization system is designed as a decision support tool specifically for resource-constrained free clinics. Our system combines a Random Forest (RF) classifier trained on MIMIC-IV-ED data with a multi-objective queue optimization algorithm presented via a staff-facing interface. Our triage model, when simulating free clinic deployment without vital sign equipment, achieves an 83.6% critical case detection rate and no dangerous misses on the test set (0.012% on holdout); optimizing for patient safety over raw accuracy. Monte Carlo simulation across 1,000 clinic sessions demonstrates a 72% reduction in wait times ($p<0.001$) for critical patients compared to first-come-first-served (FCFS) queue ordering. Unlike commercial triage systems that can be prohibitively expensive, our solution is built entirely on free and free-tier tools and designed for volunteer-staffed environments lacking trained intake nurses, vital sign monitors, and electronic health records (EHR). We developed the system as a tablet-optimized web application with real-time queue updates and physician queue overrides. Our work is entirely retrospective evaluation and simulation. Prospective studies in free clinic settings, in collaboration with clinicians and domain experts, are planned as a next step.} }
Endnote
%0 Conference Paper %T ML-Powered Triage and Queue Optimization for Resource-Constrained Free Clinics %A Armaan Grewal %B Proceedings of the 7th Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2026 %E Elizabeth Healey %E Jason Fries %E Tom Pollard %E Shengpu Tang %E Anna Zink %E Tom Hartvigsen %E Monica Agrawal %E Sam Finlayson %E Benjamin Glicksberg %E Brett Beaulieu-Jones %E Kai Wang %E Daseyra Fontalvo %E Tasmie Sarker %E Irene Chen %E Emily Alsentzer %F pmlr-v333-grewal26a %I PMLR %P 905--923 %U https://proceedings.mlr.press/v333/grewal26a.html %V 333 %X Free clinics serve about 1.7 million uninsured Americans annually in the US, yet operate under severe resource constraints that lead to missed urgent cases, inefficient patient flow, and long waits. Our machine learning (ML)-powered triage and queue optimization system is designed as a decision support tool specifically for resource-constrained free clinics. Our system combines a Random Forest (RF) classifier trained on MIMIC-IV-ED data with a multi-objective queue optimization algorithm presented via a staff-facing interface. Our triage model, when simulating free clinic deployment without vital sign equipment, achieves an 83.6% critical case detection rate and no dangerous misses on the test set (0.012% on holdout); optimizing for patient safety over raw accuracy. Monte Carlo simulation across 1,000 clinic sessions demonstrates a 72% reduction in wait times ($p<0.001$) for critical patients compared to first-come-first-served (FCFS) queue ordering. Unlike commercial triage systems that can be prohibitively expensive, our solution is built entirely on free and free-tier tools and designed for volunteer-staffed environments lacking trained intake nurses, vital sign monitors, and electronic health records (EHR). We developed the system as a tablet-optimized web application with real-time queue updates and physician queue overrides. Our work is entirely retrospective evaluation and simulation. Prospective studies in free clinic settings, in collaboration with clinicians and domain experts, are planned as a next step.
APA
Grewal, A.. (2026). ML-Powered Triage and Queue Optimization for Resource-Constrained Free Clinics. Proceedings of the 7th Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 333:905-923 Available from https://proceedings.mlr.press/v333/grewal26a.html.

Related Material