The artificial intelligence paradigm shift is taking what was once a wish into reality, specifically, cross-regulatory agency coordination and harmonization.  The latest publication in this area is a joint effort of the U.S. Food and Drug Administration (FDA), U.K.’s Medicines and Healthcare products Regulatory Agency (MHRA), and Health Canada to define guiding principles in AI titled “Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles” (here).  This publication is a follow-up from the 2021 joint publication titled “Good Machine Learning Practice for Medical Device Development: Guiding Principles.” 

The need for this type of guidance has been demonstrated by the commercial/financial plans as frequently discussed in annual reports and enterprise risk assessments, which are converging at an accelerated rate.  The economic impact of AI is undeniable, as a recent quote published in CPhI online states, “A recent study estimated that global GDP could be up 14% higher in 2030 as a result of AI – the equivalent of an additional $15.7 trillion – making it the biggest commercial opportunity in today’s economy” (here). 

A different view of these recent publications is warranted since, not only are they developing quickly, but cross-agency cooperation is also at a high level, which is critical for a global economy.  This is very much aligned with the increase in Mutual Recognition Agreements (MRA).  A thumbnail timeline is outlined below for convenience: 

  • October 2021:  Good Machine Learning Practice for Medical Device Development:  Guiding Principles 
  • October 2023:  Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles 
  • Ongoing:  The call for public comment via the FDA public docket (here) 

As tempting as it may be to be dismissive if your organization does not engage in Software as a Medical Device (SaMD), the trend is remarkably clear and obvious.  An examination of the content of the most recent publication is as interesting for what is different as it is for what is NOT changing.  Understanding the principles is a potential informal test in QMS maturity.  For instance, the items listed are: 

  • Focused and Bounded 
  • Risk-Based 
  • Evidence-Based 
  • Transparent 
  • Total Product Lifecycle Perspective 

When it comes to change management at the principle level, have expectations really changed?  These items are desirable traits for any change process.  This highlights the need to truly understand the principles of any QMS program rather than just “executing” a process.