AI Innovation: Implications for Medtech & Patient Diagnostics

In this exclusive Q&A, MD+DI continues its conversation with Talia Grace Haller, a consultant who specializes in AI strategizing for clients in the fields of life sciences and healthcare. Haller will present the session “Innovating Diagnostics: AI-Driven Biomarker Discovery for Medical Device Start-ups” during the upcoming MEDevice conference in Boston, MA, Sept. 25-26. For Part 1 of this Q&A, click here.

The ongoing rapid advancement of artificial intelligence (AI) will continue to have a profound impact on the development and performance of medtech devices, particularly as they relate to patient diagnostics. From improving healthcare diagnoses through the utilization of digital tools to making more efficient and effective recommendations for care planning with the assistance of an increasing volume of data that will be available for analysis, providers and designers alike will be offered various opportunities to use AI innovation to their advantage for the benefit of patients. There will also be challenges to navigate, however, both from a technical and ethical perspective.

What specifically are the most meaningful roles of AI in the use of devices for diagnostic purposes from your standpoint as a consultant in the medtech domain?

Haller: There are many ways AI can add value to diagnostic devices. Examples include:

How should medtech companies prepare for the continued advancement of AI and the larger amounts of data it is expected to create, as well as the need to analyze that data for meaningful, actionable insights?

Haller: The advancement of AI will indeed deepen the capabilities of data collection and analytics, particularly through aggregation and broader sharing of health information. This includes, for example, the patient sharing data with various third-party providers to seek insights, or companies sharing de-identified patient data for specific purposes. As AI systems become increasingly sophisticated at processing vast and complex datasets, such as genomic, transcriptomic, and proteomic data, the volume of information will exceed what can be managed without advanced computational support.

Key challenges and strategies for medtech companies will include:

Simply put, no human has the intellectual capacity to make sense of the volume of data being generated for even a single patient, wherein we may want to combine and analyze multiple whole genome sequences (WGS), proteomic sequences, transcriptomic sequences, metabolomic sequences, microbiome sequences, their electronic health record (EHR), wearables data, exposome data, and more.

To me, the clear benefit of AI is that it allows us to make sense of and extract insights from the astronomical amounts of data we will generate per individual in the coming years.

As it pertains to device metadata tracking, what types of opportunities could exist to improve the healthcare ecosystem?

Haller: The concept of device metadata tracking with AI involves gathering and analyzing data from medical devices about their usage and performance. This data can include information on how frequently a device is used, under what conditions it operates best, and any potential failures or maintenance issues. The benefits of this approach include the ability to optimize device design based on real-world usage and the potential to preemptively address device malfunctions before they lead to adverse outcomes.

For instance, consider a company that manufactures heart rate monitors. By using AI to track and analyze metadata from these devices, the company can identify patterns such as frequent breakdowns under certain conditions or discrepancies in data accuracy that occur after extended use. This insight might allow the company to make necessary adjustments to the product design or user guidelines to improve reliability and user experience.

Moreover, this process can create an opportunity for a feedback loop with regulatory bodies. Typically, medical device companies are required to report adverse events or potential safety issues to FDA. However, with AI tracking, companies will have the ability to go a step further by continuously sharing device performance data with FDA, thereby enhancing the regulatory oversight process.

While continuous data sharing with the agency through AI-driven device metadata tracking might not align perfectly with a company’s interests due to potential exposure of product flaws or increased regulatory scrutiny, it is something that regulatory bodies could advocate for. This approach enhances transparency and patient safety, helping to ensure that devices on the market perform safely and effectively. Companies need to prepare for and navigate these requirements carefully, balancing regulatory compliance with their strategic interests.

This proactive approach to sharing data can help to monitor safety and efficacy of medical devices more effectively. For example, if AI data analysis reveals a pattern of device malfunctions that could potentially compromise patient safety, the manufacturer can be alerted promptly to take corrective action, possibly before any serious adverse events occur. This not only helps in ensuring that medical devices on the market are safe, but also supports the goal of protecting public health. By addressing these challenges proactively, medtech companies can harness the power of AI to transform healthcare data analytics while ensuring patient safety and compliance with regulatory standards. This will enhance the efficacy of medical devices and bolster the overall healthcare ecosystem.”

With what you understand to be the expectations that patients have about AI, what are the challenges that you see device developers facing as they try to meet expectations and improve healthcare?

Haller: From my perspective, the greatest challenges facing device developers will lie in the realm of compliance and ethical responsibility. As technology allows developers to do more with patient data than ever before, the biggest challenge will be to avoid the temptation to pursue actions that, although technically feasible, may not comply with stringent regulations. For example, an EHR company could obtain many insights by combining all of their patients’ data and leveraging AI to extract insights. However, can this actually be done in a way that’s compliant with privacy, identification, bias/fairness, data integrity, and transparency policies? Potentially, but this would take more time. This is what happens when technical innovation comes face to face with compliance.

Although perhaps counterintuitive, the obstacles facing device developers are not just technical but profoundly ethical — really affecting people's lives. The task for developers extends beyond mere innovation; it involves carefully navigating a complex landscape of ethical, regulatory, and political challenges. They must operate within both fixed and evolving frameworks to use AI responsibly, ensuring that patient data is meticulously managed, and that both privacy and patient preferences are upheld. This careful balance is essential as we advance in creating AI tools that not only meet the highest standards of data privacy, transparency, and ethical responsibility, but also significantly improve patient outcomes.

How do you think today’s patients view AI as it pertains to their own healthcare?

Haller: Do they want to know when something is wrong? Perhaps they do, so they can take proactive steps to address it. Or maybe they don’t, preferring to live without the burden of worrying about potential health issues. If informed with high probability that they will develop a health condition unless they take specific actions, will they take those actions? They might, if they believe the benefits of the actions outweigh the inconvenience or discomfort. Conversely, they might choose not to, if they feel the tradeoffs are not worth it for them.

Ultimately, AI provides individuals with the power of informed choice. It equips them with the knowledge and options to make decisions that align with their personal values and lifestyle preferences. AI doesn't dictate how one should live their life; it gives the insights and tools needed to make those critical decisions themselves. This empowerment is the real gift of AI in healthcare, offering the freedom to choose one’s own path to well-being.