Introduction
This blog post uses quilts and duvets as a metaphor for the regulation of AI in healthcare. It entreats those involved in AI Regulation within Healthcare to- think Duvet not Quilt
On Quilting
The Process
My wife has recently taken up quilting, producing spectacular results despite the painstaking nature of the craft. The process begins with her developing a concept or theme for the project. Next, she meticulously designs and creates each square. After that, she arranges the squares and stitches them into the pattern. Finally, she applies a fabric back and edges to complete the quilt.
The Technical and Physical Challenge
Quilting presents technical challenges, requiring mathematics, equipment (a proper sewing machine and attachments) and 3D planning. Depending on the quilt’s size, it can also be physically demanding. Wrestling with a double bed sized quilt is not for the faint hearted!To develop the necessary expertise, my wife invested significant time in online training. This is a fulfilling pastime but not a solution for an immediate imperative
The Immediate Imperative
Now, imagine facing a pressing need to get warm. Will people wait for a quilt to be designed and made? Most people would buy a duvet to address the immediate requirement.
Similarly, healthcare faces present-day clinical safety challenges that AI-based systems and devices can address. If harm is occurring now, we should adopt a “duvet mindset” for regulatory solutions, prioritising immediate, effective action.
Introducing Technology
The NHS Digital Clinical Safety Strategy, published in 2021, sets out two key aims:
1. Improve the safety of digital technologies in health and care, both now and in the future.
2. Identify and promote the use of digital technologies as solutions to patient safety challenges.
These aims are entirely sensible and could inspire a global aspiration for introducing healthcare technologies.
Regulators: Avoid Designing Quilts
In my view, there is a significant risk that regulators, when addressing AI, focus excessively on assessing and assuring the mitigation of all conceivable risk at the expense of enabling solutions. This results in a “regulatory quilt,” painstakingly crafted while ignoring the urgent need to stay safe and warm.
Even worse, the individuals designing these regulatory frameworks are often disconnected from frontline clinicians and the patients who face these challenges daily.
Engage with the Sleepers- Duvet or Quilt?
The more I look at AI opportunities, the more I am convinced that its safe implementation will be a product of training and engagement. Clinicians and patients need to understand the strengths and weaknesses of the current technologies and to view it as a partner or valued colleague. Not as a replacement or superior intellect. AI is Robin to your Dr Batman, This also represents an imperative as clinical training establishments need to be sowing the seeds for this revolution.
Safety Challenges in Healthcare
The safety challenges in healthcare are profound, with harm affecting patients today. The World Health Organization (WHO) highlights troubling statistics:
• One in every ten patients is harmed in healthcare, leading to over 3 million deaths annually due to unsafe care.
• In low-to-middle-income countries, up to 4 in 100 people die from unsafe care.
• Over half of this harm is preventable, with half attributed to medication errors.
• Common adverse events include medication errors, unsafe surgeries, diagnostic mistakes, and healthcare-associated infections.
Patient harm reduces global economic growth by 0.7% annually, with indirect costs amounting to trillions of US dollars.
https://www.who.int/news-room/fact-sheets/detail/patient-safety
Current NHS Pressures- A shivering service
The NHS faces significant pressures driven by increasing demand, workforce shortages, and constrained resources. Rising patient volumes, exacerbated by an aging population and post-pandemic healthcare backlogs, have strained capacity across services, including primary care, emergency departments, and elective procedures. Workforce challenges, including recruitment and retention issues, have been compounded by strikes and burnout among healthcare professionals. Financial pressures further limit the system’s ability to expand services or innovate effectively. These issues are amplified by broader socioeconomic factors, such as health inequalities, which exacerbate the gap between demand and capacity, leaving the NHS struggling to meet both immediate and long-term healthcare needs.
Patient Impact
These pressures on the NHS significantly impact patient safety. Overcrowded emergency departments, longer waiting times for treatment, and delayed diagnoses increase the risk of adverse outcomes, including preventable harm and deterioration of patient conditions. Workforce shortages lead to overburdened staff, resulting in fatigue, burnout, and a higher likelihood of errors in care delivery. Financial constraints can limit the implementation of essential safety measures, such as technology upgrades or staff training. Additionally, the strain on resources reduces the time healthcare professionals can spend with each patient, compromising thoroughness and the quality of care. Together, these factors create an environment where patient safety is increasingly at risk.
Technology as a Requirement
These current global and NHS status underlines the urgent need for solutions. Technology, particularly AI, is not a luxury—it is a fundamental necessity. AI must become an integral part of our technology development strategy.
The Inevitability of AI
Many, including Mustafa Suleyman, argue that AI’s progression is inevitable, carrying profound implications for the future. This belief drives healthcare technology suppliers to invest heavily in AI solutions that enhance existing products or create entirely new ones.
Despite regulatory concern, there is good evidence that Doctors are already using AI, despite a lack of guidance about these tools and unclear work policies, an estimated 1 in 5 GPs report using generative AI to assist with their job. The medical community will need to find ways to both educate physicians and trainees and guide patients about the safe adoption of these tools.
There is, therefore, a risk that AI will be used in healthcare in advance of regulations.
https://informatics.bmj.com/content/31/1/e101102
Equitable, Accessible, and Safe Care Despite Global Challenges
AI-driven products can not only improve safety in healthcare but also help address the staggering supply-demand mismatch. According to WHO, approximately 4.5 billion people lack access to essential healthcare, and there is a global shortage of over 10 million healthcare workers.
AI Regulation in Healthcare- think Duvet not Quilt
Since AI healthcare solutions are inevitable, our regulatory approach must address the current access and safety challenges. Patients and clinicians are metaphorically shivering—they need immediate relief. Regulatory frameworks must “think duvet,” enabling safe, accessible, personalised, and equitable healthcare solutions now.