Cleared for Takeoff: Transforming Health Care Pilots into Scalable Innovation

Written by Feby Abraham, PhD, executive vice president and chief strategy and innovations officer, Memorial Hermann Health System

Healthcare innovation is at an inflection point. Providers are constantly working to develop and test new technologies and models of care with the intent to improve patient outcomes, reduce costs, enhance workforce productivity, improve patient experience and advance health. It is a complex endeavor, and to ensure these initiatives are successful, many organizations first pursue pilot programs — controlled experiments designed to test viability before full deployment. 

Just like pilot programs pursued in biosciences, the space program and other innovative organizations striving to push the boundaries of what we know and don’t know, healthcare pilot programs can stretch our understanding of how to achieve new, better and more successful outcomes for patients. However, they can also stall — falling into what is known as “pilot-isis,” where promising ideas remain stuck in experimentation without progressing to broader implementation.

In fact, according to the The New England Journal of Medicine, 68% of digital health pilots never progress beyond the test phase. Additionally, 78% of AI-driven health pilots fail to scale. While pilots are essential for de-risking innovation, their design and execution determine whether they become launchpads or dead ends.

For those of you in the innovation space: Do you feel as though your pilot program is in a perpetual cycle of experimentation? If so, the problem isn’t necessarily the concept or execution of the pilot program itself, it may lie in the design. When structured as hypothesis-driven experiments, pilots act as launch pads rather than dead-ends. Here are a few reasons your pilot may not be “taking off,” and what you can do about it. 

• Misaligned Priorities: Some pilots are solution drivers in search of a problem. If they don’t address a clear, urgent need tied to organizational priorities, decision makers can hesitate to invest in scaling.
• The Fix: Start with a problem, not a solution. Define a clear, unmet need before selecting the innovation to test. Also, assess if a pilot is necessary. Some innovations may not require a test phase and should move directly to implementation especially if they have been validated elsewhere. 
• Lack of Clear Success Metrics: Without well-defined, measurable outcomes, even promising pilots can fail to justify further investment.
• The Fix: Define clear metrics for success. Replace vague goals with measurable predictions (e.g., shift a metric like test if AI improves diagnostics to in ERs with longer than eight-hour wait times, AI triage will reduce time-to-diagnosis by 25% without increasing misdiagnoses). 
• Overly Controlled Testing Environments: Pilots often succeed in ideal conditions but fail when introduced to the complexities of real-world health care settings.
• The Fix: Test in real-world conditions. Pilots must account for operational complexities — resource variability, staffing constraints and diverse patient demographics — to ensure scalability. 
• Lack of Integration with Existing Workflows: If an innovation does not seamlessly fit into electronic medical records, reimbursement models or clinician workflows, it’s unlikely to succeed.
• The Fix: Ensure alignment with existing workflows from the start. Scalability plans should also address policy, reimbursement models and equity considerations from the start. 
• Premature Scaling (enthusiasm vs. evidence): Some organizations scale pilots too soon, without sufficient data to prove long-term viability — leading to costly failures.
• The Fix: Structure pilots as hypothesis-driven scientific experiments. Use rigorous methodologies such as control groups, pre-registered hypotheses and design of experiments to ensure objective evaluation. Establish clear go/no-go criteria with predefined benchmarks for scaling, redesigning or terminating the pilot. 
• Lack of Executive Ownership and Resource Allocation: Even well-designed pilots fail if there is lack of ownership and providing the resources in scaling if the pilot succeeds.
• The Fix: Address cultural and operational resistance early on by engaging a program leader from the onset to achieve buy-in and create a plan to allocate appropriate financial and human resources when the pilot scales. Work to incorporate change management strategies in tandem that will drive adoption at all levels of the organization. 

Beyond the Pilot: Engineering for Scale

Pilots that succeed don’t just validate ideas — they set the foundation for system-wide transformation. To bridge the gap between testing and scaling, organizations need structured processes:

1. Make pilots part of a broader innovation portfolio
   – Pilots should be embedded in a programmatic innovation framework, not treated as one-off projects.
   – A structured stage-gate model can ensure pilots systematically transition to scalable solutions.  
2. Streamline processes to accelerate experimentation
   – Establish lean approval pathways for small-scale pilots to enable rapid testing, while parallelizing longer compliance processes for broader rollout.  
3. Implement change management for adoption
   – Pilots require frontline staff buy-in for successful implementation. Early engagement, training, and transparent communication are critical.  
4. Measure pilot impact on system-level outcomes
   – The impact of pilots should be measured not just in isolation but in their ability to integrate across organizational priorities and improve operational, financial and patient outcomes at scale.
5. Incentivize learning from successes and failures 
   – Recognize and reward teams not just for successful pilots but also for smartly discontinuing those that are not viable. Learning from failure is critical for a high-functioning innovation ecosystem.  

Reframing the Role of Pilots

Pilot programs are a crucial practice for innovators across multiple fields looking for answers to any number of questions. Good pilot programs must ask smart questions, pursue evidence-based answers and not fear failure. 

Pilots are not where innovation goes to die — they are where it learns to fly. But for healthcare pilot programs to achieve their intended outcomes they must be designed as structured experiments with clear objectives, success criteria and pathways to scale.

For healthcare leaders, the message is clear: Move away from episodic, unstructured pilots. Build disciplined, scalable innovation programs. With the right design, pilots will not only validate new solutions but accelerate their adoption, driving real transformation in healthcare.

Feby Abraham, PhD, is executive vice president and chief strategy officer for Memorial Hermann.