Despite decades of technological progress, healthcare organizations face remarkably similar AI implementation challenges today as they did in 1987.
The Research That Predicted Today’s Challenges
In March 1987, Dr. Edward Shortliffe and colleagues published a prescient analysis in the New England Journal of Medicine titled “Artificial Intelligence in Medicine: Where do we stand?” Their findings were startling not for what they revealed about technology, but for what they uncovered about human organizations.
The central insight: The primary barriers to medical AI implementation were organizational, not technological.
Fast Forward 39 Years
Today, medical AI has evolved beyond the wildest dreams of 1987 researchers. We have:
- Large language models that can process medical literature at superhuman speed
- Computer vision systems that rival radiologists in diagnostic accuracy
- Predictive algorithms that can identify sepsis hours before clinical symptoms appear
- Natural language processing that can extract insights from vast medical records
Yet healthcare organizations in 2026 report the same fundamental implementation challenges:
- Integration Complexity: Fitting AI tools into existing clinical workflows
- Change Management: Getting staff to adopt and trust new AI systems
- Training Requirements: Building organizational AI literacy across roles
- Evidence Evaluation: Distinguishing between vendor hype and proven effectiveness
- Resource Allocation: Justifying AI investments to budget-conscious leadership
The SMART Perspective: Evidence Over Enthusiasm
At SMART, we see this historical continuity as a crucial insight for healthcare leaders. The lesson isn’t that medical AI is doomed to repeat past failures. Instead, it’s that successful AI implementation requires as much organizational science as computer science.
Our approach focuses on:
- Evidence-Based Evaluation: Rigorous assessment of AI tools before deployment
- Implementation Science: Systematic study of how to integrate AI into healthcare workflows
- Organizational Readiness: Preparing healthcare teams for AI adoption through targeted training
- Realistic Expectations: Setting achievable goals based on organizational capacity, not just technological capability
What Organizations Can Learn
The 1987 NEJM paper offers three timeless principles for medical AI implementation:
- Start with Workflow Analysis: Understand current processes before introducing AI
- Invest in Change Management: Technology adoption is fundamentally a human challenge
- Measure Impact Systematically: Use rigorous metrics to evaluate AI effectiveness
Looking Forward
As we advance deeper into the AI revolution in healthcare, the lessons from 1987 remain strikingly relevant. The organizations that will succeed with medical AI aren’t necessarily those with the most advanced technology—they’re the ones that approach implementation with the same rigor they apply to clinical research.
The technology has evolved dramatically. The implementation science has remained constant.
Ready for Evidence-Based AI Implementation?
SMART specializes in helping healthcare organizations navigate the organizational challenges of medical AI deployment. Our evidence-based approach ensures that your AI investments deliver measurable improvements to patient care.
Published: March 20, 2026 | Reading Time: 4 minutes Research and Analysis by HERBIE Scientific Content Intelligence