To Issue 166
Citation: Bubb T, “Navigating the AI Revolution in the Medical Devices Sector: Between Innovation and Regulation“. ONdrugDelivery Online, October 3rd, 2024
Timothy Bubb discusses the potential impact of artificial intelligence and machine learning on the healthcare and drug delivery industries, with an overview of the differing regulatory approaches taken by the EU, US FDA and UK MHRA towards this novel area of digital health.
The advent of digital health technologies, from wearable sensors that monitor vital signs to artificial intelligence (AI)-powered diagnostic tools, is bringing about a vast range of innovations and promises to revolutionise the healthcare landscape. When it comes to drug delivery systems, there is exciting potential around the capability of digital and AI-powered delivery systems to sense physical cues, such as peaks in specific hormones, and modify drug delivery responses based on known past patient responses to drug delivery and dosage under similar circumstances.
AI and machine learning (ML) can dramatically impact healthcare delivery by facilitating early disease detection and personalised treatment approaches, as well as augment remote consultations through telehealth solutions. AI-powered algorithms can have the capability to analyse vast troves of health data to identify disease biomarkers, predict disease trajectories and tailor interventions to individual patients.
SOFTWARE AS A MEDICAL DEVICE: A DEFINITION
Digital health encompasses a spectrum of technologies, ranging from non-medical devices designed to monitor wellbeing through to medical devices tailored for specific medical purposes. The many mindfulness, sleep-tracking and other wellbeing apps flooding the internet, for example, are usually not medical devices.
A consensus, but not exhaustive, definition for “software as a medical device” (SaMD) is provided by the International Medical Device Regulators Forum (IMDRF), which describes it as “software intended to be used for one or more medical purposes that perform these purposes without being part of a hardware medical device.”
“Understanding the difference between digital health and digital medical devices, although not always easy, is crucial to comprehend regulatory nuances.”
Understanding the difference between digital health and digital medical devices, although not always easy, is crucial to comprehend regulatory nuances. When evaluating sensor and monitoring devices, such as those used by drug delivery systems, this is a critical definition that determines how strictly devices are regulated and that could severely impact patient outcomes.
In its June 2023 Roadmap, the UK MHRA confirmed that it is developing guidance to help clearly identify SaMDs, differentiating them from the wide and confusing range of other tools, such as wellbeing and lifestyle software products, in vitro diagnostics software and companion diagnostics.1
AI AND ML APPLICATIONS IN HEALTHCARE
More and more AI/ML tools are being developed in the healthcare sector to help manage vast amounts of data and interpret it speedily and accurately. Whether this data is in text form, video or imagery, AI/ML can help save hours of manual analysis and cross-checking and suggest interpretations that would otherwise take human reviewers years to complete. AI/ML can rapidly analyse radiology images, histological data, posture, eye movement, speech speed, pitch and sound, as well as a whole range of other types of input.
“AI and ML have the potential to revolutionise several areas of healthcare, from diagnostic images to remote patient monitoring, CDSS, predictive analytics for healthcare management and even robotic surgery.”
AI and ML have the potential to revolutionise several areas of healthcare, from diagnostic images to remote patient monitoring, clinical decision support systems (CDSS), predictive analytics for healthcare management and even robotic surgery. Among these, drug delivery represents a substantial segment of the healthcare sector where connectivity, AI and ML are poised to make a significant impact.
However, the application of AI and ML in commercial drug delivery products is currently limited. Drug delivery systems require greater assurance of performance and safety due to the inherent risk of harm from inappropriate delivery and dosing, and therefore higher levels of clinical evidence and longer regulatory review times are required. In the future, we may see increasing miniaturisation of systems gathering data or used for drug delivery. Nanomedicinal techniques currently in emergent research – such as nanorobots programmed to identify a target, navigate to target sites based on physiological conditions to deliver a drug and finally be expelled by the body – are becoming more technically feasible with the advent of new and innovative ML techniques.
This would enable highly targeted and controlled drug delivery – potentially increasing efficacy and reducing side effects. This may be especially effective when there is a combination drug therapy or one that involves very specific dosage timings, such as cancer treatments, requiring a six-to-seven drug combination therapy.2
“AI for drug delivery also holds promise within the companion diagnostics field, where AI algorithms can accurately predict the most suitable patient-specific dosing regimens and patient suitability for a particular drug delivery modality.”
AI for drug delivery also holds promise within the companion diagnostics field, where AI algorithms can accurately predict the most suitable patient-specific dosing regimens and patient suitability for a particular drug delivery modality. However, navigating the evolving regulatory landscape for these devices and the platforms that house and analyse the data they collect can be challenging for medical device manufacturers.
Although it is currently not possible to attach sensors directly to injected liquids or inhaled powders, significant advancements are being made by integrating electronics for sensors and connectivity into the delivery devices themselves, such as syringes, inhalers, autoinjectors, eyedroppers and nasal sprays. This innovative approach transforms these delivery systems into smart devices, capable of collecting and transmitting critical data to healthcare providers.
By embedding sensors within these delivery devices, it becomes possible to monitor correct administration of medication, track usage patterns and even gather new physiological data from the patient. For example, a smart syringe could record the time and dose of an injection, while an intelligent inhaler or chamber might track the frequency and effectiveness of inhalation therapy based on breath volumes and velocities. Using monitoring devices, it becomes possible to track the therapeutic and side effects of the medicine received with greater precision, which can then inform future dosing decisions.
This approach has been used to great effect with continuous glucose monitoring devices used in combination with insulin pumps to create an “artificial pancreas”, automatically delivering insulin when required as part of the management of Type 1 diabetes. The technique of combined monitoring and automated drug delivery has similar applications in other chronic conditions that require a personalised and time-critical dosing schedule. This data can subsequently be transmitted to healthcare professionals, enabling more precise and personalised treatment plans.
Smart drug delivery systems can also play a critical role in improving patient adherence to medication regimens. Reminders and alerts can be integrated into the devices, encouraging patients to take their medications as prescribed. Adherence support is particularly valuable for managing chronic conditions where consistent medication adherence is vital for effective treatment and, sadly, a number of chronic conditions, such as Type 2 diabetes and glaucoma, are becoming ever more common.
Glaucoma treatment, specifically, offers an interesting case study of how AI can help improve drug delivery systems, as recent research shows that AI can accurately predict an effective sequence of amino acids that would bind to a particular chemical in animal eye cells and safely dispense medications over several weeks. This resulted in US FDA approval of an implantable device that can be placed in the eye and release drugs to treat glaucoma. However, this case demonstrates a consideration for any risk/benefit assessment needed as part of choosing a suitable drug delivery method as, although that device worked for longer periods than drops or injections and does not require a dosing schedule, prolonged use was shown to cause eye cell death in some cases.3
These application areas demonstrate the versatility and potential impact of AI and ML in medical device development and healthcare delivery. In underfunded areas of medical research, this could prove life-changing by helping detect comorbidities, or environmental or genetic factors that place particular individuals at higher risk of disease.
EU, UK AND US: DIFFERENT REGULATORY APPROACHES
As the industry embraces these advancements, regulatory bodies the world over are grappling with the complexities of evaluating and approving medical devices that do not conform to traditional paradigms and do not have a physical presence in the traditional sense. This is while also addressing concerns regarding data security, potential bias and potential safety impacts from poorly performing clinical software tools, which underscores the need for robust laws to support proportionate regulation of AI across a range of sectors, including healthcare.
The EU has published the final text for the AI Act,4 which will regulate AI systems in multiple industries, including medical devices, and specifies the core objectives and requirements for healthcare AI regulation. Additionally, the legislation specifically calls out certain digital health technologies as being high risk – for example, AI systems that are used for emergency healthcare patient triaging and systems used in evaluating eligibility for certain healthcare services. These are classified as high risk “since they make decisions in very critical situations for the life and health of persons and their property.” These digital health products will therefore now require notified body assessment and CE marking as an AI system, which will be based on different criteria from existing requirements for medical devices.
In the US, the FDA has significant experience of successfully regulating AI/ML-enabled devices and has gone so far as to compile a publicly available list of such AI-enabled tools with FDA marketing clearance. The FDA currently applies its “benefit-risk” framework and confirms that devices must conform to some basic principles, such as the demonstration of sensitivity and specificity for devices used for diagnostic purposes, the validation of intended purpose and stakeholder requirements against specifications and development that ensures repeatability, reliability and performance. The FDA has also considered the need for some AI/ML systems to be adaptively retrained on new data or context-specific data, and has introduced processes to enable certain pre-authorised software changes to be agreed by the manufacturer and the FDA that can then be deployed without the need for further regulatory assessment.5
In the UK, a new regulatory sandbox, which was launched in pilot in May 2024, has been designed to provide a safe space for AI tool developers in healthcare to trial innovative AI products in view of regulators before they are implemented,6 in an effort to help simplify the confirmation of clinical efficacy and safety for AI systems. To respond to a need to keep pace with evolving ML methods and applications, the MHRA will also develop a system based more on guidance than regulation in the UK, allowing for more frequent updates to account for the speed of innovation. Alongside the FDA and Health Canada, the MHRA has outlined 10 guiding principles that can inform the development of good ML practice (GMLP) that are safe, effective and promote high-quality medical devices that use AI and ML.7
In 2023, the MHRA also updated the “Software and AI as a Medical Device Change Programme” to ensure that future regulatory requirements for software and AI are clear and patients are protected. In a bid to address bias and inequalities, the MHRA also confirmed that it recognises that SaMD and AI as a medical device must perform across all populations within the intended use of the device and serve the needs of diverse communities.
CONCLUSION
As the medical device landscape continues to evolve, given the heightened complexity of AI-based medical devices – particularly in drug delivery systems that integrate or use AI algorithms – manufacturers should consider trusting regulatory experts and embracing multidisciplinary approaches to navigate these challenges and ensure compliance. Partnering with professionals who possess in-depth knowledge of regional regulations and emerging trends in medical devices can expedite market entry and ensure compliance with evolving standards and requirements. By embracing collaboration and expertise, innovators can navigate regulatory challenges and pave the way for transformative solutions in medical devices and drug delivery systems.
REFERENCES
- “Software and AI as a Medical Device Change Programme – Roadmap”. UK MRHA, Jun 2023.
- Bose P, “Optimizing Drug Delivery Using AI”. AZO Life Sciences, May 2022.
- “AI Used to Advance Drug Delivery System for Glaucoma and Other Chronic Diseases”. John Hopkins Medicine, May 2023.
- “Proposal for a Regulation laying down harmonised rules on artificial intelligence”. European Commission, Apr 2021.
- “Draft Guidance for Industry: Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions”. US FDA, Apr 2023.
- “MHRA launches AI Airlock to address challenges for regulating medical devices that use Artificial Intelligence”. Press Release, UK MRHA, May 2024.
- “Good Machine Learning Practice for Medical Device Development: Guiding Principles”. Joint Press Release, US FDA, Health Canada and UK MRHA, Oct 2021.