Regulations to Enable Safe, Effective & high-Quality AI/ML/Digital Solutions in Drug Development

Shrishaila Patil
Vice President, Statistical Programming, Navitas Data Sciences

From mobile medical apps and fitness trackers, to software that supports clinical decisions doctors make every day, digital technology has been driving a revolution in healthcare. As we are learning to live the virtual new normal in Covid19 era, adoption of digital solutions has witnessed further acceleration.

To keep pace with this promising innovation, the regulators must modernize its approach to regulation. Question to ask ourselves is “How to ensure high-quality, safe and effective digital health products”.

This article highlights changing Regulatory landscape (European/FDA Guidelines) for enabling Safe, effective, high-Quality & trustworthy AI/Digital solutions in Pharma.

What’s Changing (Covid19 acting as a catalyst)

We have witnessed many innovations & novel approaches across Clinical trial processes in past few years per below.

In some sense, Covid19 has acted as a catalyst to some of these great innovations and adaptations.

  • Real word data/evidence
    • To Leverage diverse data sources and analytics tools to enhance study design and protocol optimization, capture clinical trial data more efficiently.
    • To make healthcare data tangible and manageable using novel data analytics to extract insights, better understanding of disease. RWD can come from several sources, for example:
      • Electronic health records (EHRs)
      • Claims and billing activities
      • Product and disease registries
      • Patient-generated data including in home-use settings
      • Data gathered from other sources that can inform on health status, such as mobile devices
  • Digital Data collection methodologies (mobile technology, Wearables, ePRO, eCOA etc.)
    • To Enable robust data capture whether reported by the patient or leveraging novel sensors. Example: We have seen Apple Heart Study on 4 lakh subjects to identify irregular heart rhythms.
  • Conventional clinical trial to Virtual Trials design
    • To Improve Patient Recruitment, retention, real-time access to Data, better Quality
    • Sponsors need to ensure that vendors have enough industry, technological and regulatory expertise to create a virtual trial environment.
    • Training needed to personnel and patients involved in the trial on Tools used.
    • Hybrid model improves the patient experience and gives sponsors a low-risk jumping off point to become familiar with the virtual trial environment.
  • Routine monitoring to Risk based and centralized monitoring
    • Improved and Efficient Approach to Clinical Trial conduct, oversight by moving away from 100% source data verification. Thanks to ICH E6 (R2) for encouraging this approach.
    • Real-time remote monitoring enabling decentralized healthcare
  • Digiceuticals or Digital Therapeutics
    • Digital therapeutics are evidence-based therapeutic interventions driven by high quality software programs to prevent, manage, or treat a medical disorder or disease. Examples:
      • FDA Approved “Bluestar” Phone App (from Company Welldoc) for managing diabetes and is the “first FDA-cleared mobile prescription therapy”
      • FDA Approved “reSET” to treat Substance Use Disorders.
  • Personalized Medicine
    • Making sense of Genomic data and Personalized adaptive treatment, dosing plan for each patient
    • Medical treatment tailored and adaptive to patient condition
  • Leveraging Big Data, Robotic Process Automation, Artificial Intelligence, Machine Learning, Deep Learning and Natural Language Processing
    • Using technologies to unlock valuable new insights from our rich and diverse data

One needs to know how this Technological Transformation impact on

  • Patients
  • Study/Protocol Design
  • Regulatory Compliance
  • Data Quality
  • Investigators
  • Pharma Companies
  • CROs
  • Biometrics Process/CDM functionality
  • Scientific/Technical/Data Standards

As a core part of Clinical Trial Team, each of us will have a significant role to play in this exciting time of transformation.

We need to have new mindset & adapt our SOPs, Processes, approach to accommodate these changes in the way Clinical trials are conducted, Data is captured, analysed and Reported.

As we adapt to changes, we need to continue to ensure Data Integrity, Quality and Security by ensuring compliance to changing requirements from Global regulatory bodies.

FDA - Key Guidelines and Initiatives:

The 21st Century Cures Act (Cures Act), signed into law on December 13, 2016, is designed to help accelerate medical product development and bring new innovations and advances to patients who need them faster and more efficiently.

As of now, more than 30 artificial intelligence (AI) algorithms have been approved by the US Food and Drug Administration including those for the detection of diabetic retinopathy, stroke, brain hemorrhage and atrial fibrillation.

FDA released its “Digital Health Innovation Action Plan” in year 2017 & it offers Clarity about FDA’s role in advancing safe and effective digital health technologies and addresses Key provisions of 21st Century Cures Act.

FDA released its “Policy for Device Software Functions and Mobile Applications” in year 2019. This guidance explains how Agency plans to regulate software that aids in Clinical decision support (CDS), including software that utilizes machine-learning-based algorithms.

FDA plans to focus regulatory oversight on “higher-risk software functions”, such as those used for more serious or critical health circumstances and Software that utilizes machine-learning based algorithms, where user might not readily understand program’s “logic and inputs” without further explanation

Digital Health Software Precertification (Pre-Cert) Pilot Program: Intention of Pre-cert Pilot program is to build confidence in stakeholders that participating organizations have demonstrated capabilities to build, test, monitor and proactively maintain and improve safety, efficacy, performance & security of medical device software products, so that they meet or exceed existing FDA standards of Safety and Effectiveness.

Pre-certified developers could then qualify to be able to market their lower-risk devices without additional FDA reviews or with a more streamlined pre-market review faster by CDRH staff.

Software Pre-Cert Program would pre-certify organizations, not individual products based on five excellence principles.

  • Patient safety
  • Product quality
  • Clinical responsibility
  • Cybersecurity responsibility
  • Proactive culture

Achieving FDA Approval

FDA Approval Status could be put at risk after each update or new iteration for AI-powered therapeutic/diagnostic tool. Example: Security Updates, adding new features or functionalities or updating an algorithm, etc. Planning to take version-based approach to the FDA approval process might be good idea. In this approach, new version of software is created each time the software's internal ML algorithm(s) trained by new set of data, with each new version being subjected to independent FDA approval.

FDA's total product lifecycle (TPLC) approach on AI/ML workflow:

The highly iterative, autonomous, and adaptive nature of these tools requires a new, total product lifecycle (TPLC) regulatory approach that facilitates a rapid cycle of product improvement and allows these devices to continually improve while providing effective safeguards.

To fully realize the power of AI/ML learning algorithms while enabling continuous improvement of their performance and limiting degradations, the FDA’s proposed TPLC approach is based on the following general principles that balance the benefits and risks, and provide access to safe and effective AI/ML-based SaMD (Software as Medical Device):

  • Establish clear expectations on quality systems and good ML practices (GMLP)
  • Conduct premarket review for those SaMD
  • Expect manufacturers to monitor the AI/ML device and incorporate a risk management approach in development, validation, and execution of the algorithm changes (SaMD Pre-Specifications and Algorithm Change Protocol); and
  • Enable increased transparency to users and FDA using postmarket real-world performance reporting for maintaining continued assurance of safety and effectiveness.

European Commission on Artificial Intelligence: EU has adopted a three-step Approach per below.

1. Setting out requirements for achieving trustworthy AI

  • Human Agency and Oversight: AT systems should enable equitable societies by supporting fundamental rights and not decrease, limit or misguide Autonomy
  • Robustness and Safety: AI should be secure, reliable and robust enough to deal with errors or inconsistencies during all life cycle phases of AI systems
  • Privacy and Data governance: Citizens should have full control of their own data. Data should not be used to harm or discriminate against them.
  • Transparency
  • Diversity, non-discrimination and fairness
  • Societal and environmental well-being: AI systems should be used to enhance positive social change, enhance sustainability and ecological responsibility
  • Accountability for AI Systems and their Outcomes: Mechanisms should be put in place to ensure responsibility, Accountability for AI systems & their outcomes

2. European AI Alliance:

  • Large-scale pilot with partners involving wide range of stakeholders for feedback
  • Members of AI high-level expert group will help present and explain the guidelines to relevant stakeholders in Member states

3. Building international consensus for human-centric AI

  • Play an active role in international discussions and initiatives

Addressing AI Black Box Issue:

  • Can users understand the Root cause of the negative outcome in AI solutions? Can users identify the training data or Machine Learning paradigm that led to AI application’s specific action? Incorrect Training Data can lead to misdiagnosis, incorrect treatment recommendation. Thus, Clinical adoption is slow and there is a lack of trust by Patients too.
  • At this stage, AI can help provide with data for better decision making (AI will not be fully replace decision making process). AI software developers will have to demonstrate to Clinicians that with this tool they can do better job.

Formalized AI Use cases by American College of Radiology (ACR) Data Science Institute (DSI)

  • To increase utilization of AI adoption in Medical Imaging, American College of Radiology (ACR) Data Science Institute (DSI) started releasing formalized use cases for how AI tools can be reliably used.
  • Use cases empowers AI developers to produce algorithms that are Clinically relevant, ethical and effective.
  • The use cases are designed to guarantee that algorithms are applied to address clinical questions and allow for quality assessment measures and comply with legal, regulatory and ethical measures.
  • At present, ACR DSI has use cases for breast imaging, cardiology, musculoskeletal, neurology, oncology, pediatrics and thoracic.

Approved AI solution in Pharma – Case Study - Example

IDx-DR is an AI diagnostic system that autonomously analyzes images of the retina for signs of diabetic retinopathy.

This is the first ever autonomous AI system cleared by FDA (announced on 11Apr2018) that provides screening/diagnostic decision without need for a clinician to interpret the image or results.

This device is a software program that uses an AI Algorithm to analyze images of eye taken with a retinal camera called Topcon NW400. Once the Doctor uploads digital images of patient’s retina to a cloud server on which IDx-DR software is installed, Software provides doctors with one of two results:

  • More than mild diabetic retinopathy detected; refer to eye care professionals
  • Negative for more than mild diabetic retinopathy; rescreen in 12 months

Conclusion:

With ongoing Global health concerns & pandemic situation, we need to identify, encourage forward-looking ideas, facilitate innovations, bring them to life and grow the ideas into potential solutions for our customers & most importantly to save precious lives of needy patients.

We need to ensure Patient Safety & Data Integrity when dealing with too much of Technology. Working professionals need to understand changing landscape, Data sources, Metadata, ensure Data Quality during processing & upskill themselves to manage new technology/tasks.

Global Regulatory bodies are encouraging implementation of Technology for faster Drug development and are coming-up with Key Guidance documents and Initiatives.

Sponsors, Vendors need to think of innovative Study design & execution plan to catch-up with Technology driven elements. One needs to think of developing innovations beyond current boundaries.

References: