Are Current Medical Device Regulations Obsolete in an AI-driven, Digital Health Future?

HealthSpark, Episode 26: Ariel Stern, Alexander von Humboldt Professor for Digital Health, Economics, and Policy at the Hasso Plattner Institute and Professor at the University of Potsdam, examines how rapidly evolving algorithms and digital therapeutics are challenging legacy oversight systems, and what smarter, more adaptive policy approaches may require.

Ariel Stern for HealthSpark

Medical Device Development graphic

How should health systems respond to rapidly evolving medical devices?

Breakthrough innovations, like AI-driven diagnostics and digital therapeutics, are becoming critical for patient care, but they are advancing much faster than the regulatory frameworks designed to oversee them. Many of these new tools are software-based and continuously updated, and a single software change may impact clinical performance. This raises important questions about how these technologies are classified, evaluated, and monitored to make sure they remain safe and effective.

Can regulatory frameworks evolve through real‑world experimentation?

Some health systems are beginning to treat regulation itself as something to be tested and refined. They are testing regulatory sandboxes, pilot pathways, and time‑limited authorizations for certain types of digital products. These approaches allow new technologies to be introduced in controlled settings under close observation to collect data on performance, safety, and unintended consequences. The resulting evidence can then be used to revise approval pathways, post‑market requirements, and guidance so that regulation adapts based on real‑world use.

How can incentives be redesigned to reward better evidence and outcomes?

For many digital and device-based interventions, success depends on whether payment systems recognize and reward their value. Some countries are testing payment models that tie regulatory decisions to reimbursement and require manufacturers to demonstrate ongoing safety, effectiveness, and real-world impact. In practice, this could mean time‑limited coverage linked to data collection, higher payment for tools that improve outcomes, and withdrawal of payment for those that do not.

Key question to take forward:

As you watch the video and consider your own setting, you might reflect on: 

What is the greatest opportunity to better align regulation, incentives, and evidence on new technologies to improve patient care?

TECD Course Ad

 

Sign Up to Receive HealthSpark Weekly

Explore Related Programs:

To learn more about how evolving regulations, global approval pathways, and reimbursement strategies impact the success and safety of AI, software, and other medical technologies, explore the Medical Device Development and Commercialization course in HealthXcelerate. Ariel Stern also leads the Technology-Enabled Care Delivery: From Digital Medicine to AI certificate program, which builds the clinical, operational, and regulatory perspective needed to lead the adoption and scaling of digital and AI-driven care models.

Medical Device Development and Commercialization

Examine the structure of the medical device industry, including the major types of devices on the market today, how companies are typically organized, their economics, and the keys to success in the industry.

  • HealthXcelerate
  • Online; Self-Paced
  • Group Enrollment Option
  • Start Today

Dates: Always Available

For: Health care professionals throughout the industry seeking better understanding of the health care system

Technology-Enabled Care Delivery: From Digital Medicine to AI

Develop the clinical, operational, and regulatory skills required to lead the adoption and scaling of digital and AI-driven care models.

  • Postgraduate Certificate
  • Blended
  • Registration Deadline Approaching:

Dates: -

For: Individuals who are evaluating, implementing, managing, or supporting new approaches to care, including digital health, virtual care, remote patient monitoring, hospital-at-home, artificial intelligence, data, and other technology-enabled models.