Skip to main content
Article

The Government Shutdown, the Healthcare Spending Crisis, and the Case for Artificial Intelligence in Healthcare

An essay by Dr. Omar Costilla Reyes links U.S. healthcare spending pressures to the case for transparent, explainable artificial intelligence in healthcare.

Written by Dr. Omar Costilla Reyes
The AI for All session in progress at the 2025 CHCI Tech Summit in Washington, DC, on November 19, 2025. Related photograph from AI for All: Building an Inclusive Future in Health Tech. Credit: Congressional Hispanic Caucus Institute; image courtesy of Dr. Omar Costilla Reyes.

Healthcare spending and operational pressure

Dr. Omar Costilla Reyes published this essay on LinkedIn in November 2025 as founder and chief executive officer of Equ Healthcare. It uses the 2025 U.S. government shutdown, and the budget debate over health-coverage subsidies, as a lens on a larger problem: healthcare is the largest and fastest-growing part of the national budget. Citing figures from the Centers for Medicare & Medicaid Services, the essay notes that national health expenditures now amount to several trillion dollars a year and a large share of gross domestic product, and argues that access and outcomes remain uneven. It also says that research estimates a substantial part of that spending comes from administrative overhead, redundant procedures, and preventable inefficiencies, and that physicians spend a large part of their workday on electronic health records and desk work.

A case for explainable healthcare AI

Many machine-learning systems are black boxes: they give a prediction without a visible line of reasoning. The essay proposes neurosymbolic AI, which pairs neural networks, which learn patterns from data, with symbolic reasoning, which applies explicit rules and domain knowledge, so that a system can both learn and explain. It illustrates the idea with a hypothetical emergency-department scenario: a patient with mild chest pain, for whom a recommendation records a low Wells score (a clinical rule that estimates the probability of a blood clot in the lungs) and a negative D-dimer (a blood test used to help rule such a clot out), so a clinician can review each step and override it. The scenario shows how an explainable system can lay out its reasoning in terms clinicians already use, which is what makes a recommendation reviewable at the bedside.

From analysis to responsible use

The essay names administrative automation, clinical decision support, and policy forecasting as areas for careful AI use, and treats transparency, privacy, and clinical oversight as design requirements. It says evidence from decision-support systems shows measurable reductions in unnecessary imaging and faster prior-authorization cycles, depending on integration and workflow design, and argues that efficiency gains should come from removing waste rather than cutting care. It closes by describing Equ Healthcare’s aim of building explainable, privacy-first AI, and it places that aim within the wider debate over how health systems can lower costs while keeping clinicians in charge of decisions. In the same month Dr. Costilla Reyes joined a panel on inclusive health technology at the Congressional Hispanic Caucus Institute Tech Summit in Washington, DC.

Topics