New Tool Helps Hospitals Decide If AI Is Worth Its Cost

Hospitals are under pressure to adopt AI, but most AI tools are judged only on accuracy — not on whether they actually save money or improve care. The new FLARE framework changes that by calculating the full financial picture: what the AI costs to build and run, how it changes staff workflows, and whether those changes are worth it. It uses simple logic and cost analysis to handle real-world uncertainty, like varying patient numbers and hidden expenses.

In a real test, researchers applied FLARE to AI that detects a type of stroke caused by large blood vessel blockages. They found that the AI becomes profitable only when a hospital treats roughly 4,000 stroke patients per year. For hospitals seeing about 5,000 patients a year, the AI pays for itself in the first year. Below that threshold, the technology might actually lose money — even if it works perfectly.

The key insight is that AI's value depends on much more than accuracy. Patient volume, how long staff take to verify results, what computer systems are already in place, and how the AI fits into daily routines all matter. A hospital with fewer patients or slower workflows might get less benefit. The framework makes these trade-offs visible, so administrators and doctors can decide together whether to invest.

FLARE is a decision-support tool, not a replacement for human judgment. It gives hospitals a clear, evidence-based way to evaluate AI before spending money — potentially saving millions in wasted technology purchases. For patients, it means hospitals will adopt AI that genuinely improves care and avoid AI that only adds costs.