Just by taking notes during patient visits, generative AI could save a five-physician primary care practice $291,200 in one year. The practice would see a return on its AI investment of 94.13% and reach the breakeven point in a bit more than six months.
The figures are based on a number of variables, including an outlay of $150,000 for the technology, overhead costs of $250,000 per year and time savings of seven minutes per patient visit (three minutes with AI vs. 10 minutes without).
Rubin Pillay MD, PhD, MBA, of the University of Alabama came up with the numbers using the Time-Driven Activity-Based Costing (TDABC) model. He shares the results at his Substack blog, RubinReflects.
Looking at modeled scenarios involving radiology as well as primary care, Pillay notes ROI increases along with scale. A one-physician primary care practice would see direct ROI of just 16.48%, for example, and take more than 10 months to break even on its AI investment.
Pillay is a family physician and clinical pharmacologist whose UAB titles include chief innovation officer at the Heersink School of Medicine. Along with the ROI calculations, he presents six observations gleaned from the exercise.
1. Scale matters.
The ROI for both modeled specialties, radiology and primary care, improved dramatically when scaled from a single practitioner to a group of five, Pillay found while conducting the analysis. He comments:
‘This suggests that larger healthcare organizations may be better positioned to benefit from AI implementations.’
2. Faster breakeven for radiology.
The radiology use case showed a quicker path to breakeven and higher ROI compared to the primary care scenario.
‘This could be due to the higher volume of discrete tasks (X-ray readings) and the more significant time savings per task.’
3. Improved efficiency.
In both specialties modeled, Pillay observed, AI significantly reduced the time required for key tasks.
‘The time savings may allow healthcare providers to see more patients or focus on more complex cases.’
4. Non-financial benefits.
While the analysis focused on objective financial metrics, it’s important to note the potential for subjective but meaningful gains from AI implementation.
‘These potentially include improved patient care, reduced burnout and increased job satisfaction.’
5. Variability in implementation costs.
The assumed costs for AI systems are estimates and may vary significantly based on the specific solution and scale of implementation, Pillay points out.
‘Healthcare organizations should carefully assess these costs in their own contexts.’
6. Time savings translate to cost savings.
The TDABC model clearly illustrates how time savings directly impact costs, Pillay emphasizes.
‘The modeling provides a clear rationale for AI adoption in time-intensive tasks.’
Pillay suggests healthcare leaders use his ROI framework “as a starting point for evaluating AI solutions, while also considering the broader implications for their organizations and patients.”
Read the whole thing.
