Evaluation of government medical program

A Syrian Scientific Achievement at the U.S. National and Governmental Level Dr. Mohammad Hassan Al-Makdash

Imagine the government asks you this question:

"We have a national medical program that has been serving millions of veterans across hundreds of centers and hospitals for years.

But… is this program truly better than standard medical care?

Does it accelerate diagnosis and treatment?

And where exactly is this impact observed — in heart disease, lung disease, cancer, or other conditions? Or in patterns of medication and treatment prescriptions?"

This is the central question that guided our research, fully supported and competitively funded by the AHOBPR program under the U.S. Department of Veterans Affairs (VA). The goal was to evaluate the effectiveness and impact of the VA Toxic Exposure Medical Evaluation Program on a national scale by developing a modified analytical algorithm with innovative methodologies applicable at the public policy level.

In this large-scale national study, Dr. Mohammad Hassan Al-Makdash, a statistician and data scientist, led the design and analysis of longitudinal data from the VA Toxic Exposure Medical Evaluation Program, which is implemented across hundreds of VA centers and hospitals throughout the United States.

The core research question was:

"Is this national toxic exposure evaluation, conducted in multiple VA centers, superior to traditional primary care in accelerating veterans’ access to appropriate care, tests, and diagnoses? And if so, in which specific areas? Does the benefit appear in particular diagnostic domains among thousands, or in patterns of medication prescriptions, treatments, and medical procedures?"

To tackle this complex, longitudinal question, Dr. Al-Makdash, together with a team of leading researchers from top U.S. universities and government centers, developed an advanced modified analytical algorithm based on High-Dimensional Propensity Score (HDPS) methodology.

HDPS was used as the initial filtering stage to generate variables from ICD and CPT codes, as well as medication prescriptions, aiming to identify the factors most likely to influence outcomes. This was followed by a second stage using the LASSO algorithm to select the most impactful and statistically significant variables from those filtered by HDPS, enabling the construction of a robust and balanced model.

The results were then evaluated using a Cox Proportional Hazards Model to estimate the effect of the medical evaluation on the time to diagnosis or treatment initiation for each code selected by the algorithm, employing iterative automated analyses on government servers.

This research was fully competitively funded by the AHOBPR (Airborne Hazards and Open Burn Pit Registry Pilot Project Program) of the U.S. Department of Veterans Affairs, reflecting its national significance.

This study represents an exemplary application of advanced analytical techniques to government-level programs, enabling intelligent and sustainable evaluation of national health policies.

Published in: Annals of the American Thoracic Society https://www.atsjournals.org/doi/abs/10.1513/AnnalsATS.202408-835OC