There's no denying it's a big buzzword in digital health right now, but what impact could AI have on the future of healthcare?
We spoke to four leaders from across the sector to share their views on AI's transformative power, from revolutionising personalised medicine to reshaping operational efficiency. They also explore how AI impacts healthcare, while highlighting key challenges in adoption at scale.
Martin Sandhu, CEO & Founder at Nuom, shared his views focusing on how AI will improve personalised care: “AI will have a bigger impact in healthcare than we can possibly imagine. It represents the future of our collective health. However, I feel the trend that will impact patient care most, by being new, novel, and intersecting with us as patients, will be the use of digital twins and precision medicine.
“Digital twins will be AI models of our organs (or our entire body) using our own personal data, to represent how our individual physiology and pathology works. In order to test the safety and efficacy of drugs, or give direction in surgery (for example, what valve to use in heart surgery, and exactly where). As a result, precision medicine will offer optimal personalised suggestions for bespoke treatment.“
Tamsin Chislett, COO at Hyfe, also touched upon the possibilities for AI to improve the personalisation of healthcare: “One of the historical challenges to scaling personalised medicine was the need for specialised infrastructure (high resolution imaging etc) and the lack of sensitive longitudinal data. Acoustic AI has the potential to eliminate both of these barriers. By detecting things like coughing and wheezing, passively and continuously via wearable technologies, a new generation of AI models can design and continuously adjust treatment plans for chronic sufferers (COPD, asthma, IPF) based on real time data. AI will monitor health indicators continuously and unobtrusively and provide risk screening, early detection, personalised advice, personalised treatment regimen and can even alert healthcare providers in emergencies.”
Turning her attention towards using AI to boost healthcare efficiency, Kelly Klifa, CEO & Co-Founder at Ally Health, added: “I believe a lot of value can be created by initially deploying AI to increase efficiency of operational processes involved in delivering healthcare. At Ally Health where we look to optimise in person nursing care delivery at home, AI can be applied in preference-based planning - a form of automated planning and scheduling which allows us to satisfy several preference plans at once (e.g. location for care delivery, clinical skills required, timings etc.).”
Tamsin also drew on the efficiency of AI, sharing her predictions on how AI will speed up drug development and be used for predictive models: “AI is already drastically reducing the time and cost of drug development by predicting how different drugs interact with various biological systems as well as the best molecular structure for the purpose. This will expedite the introduction of new, life-saving drugs to the market. Additionally, new, precise ways to measure things like cough in the real world will accelerate the trials and increase the quality and reliability of data.
“In the recent pandemic we learned the hard way just how consequential it is to be unprepared. AI will play a significant role in predicting and managing future epidemic outbreaks. Using data from various sources, AI can forecast outbreaks and guide public health responses, potentially saving thousands of lives. Acoustic AI lends itself particularly well to this purpose, as cough frequency changes can be detected in real time at scale. This can help predict and respond to things like a flu epidemic as well as the next respiratory pandemic.”
While AI may promise to change healthcare delivery, Noel O’Kelly, Medical Director & Co-Founder at Clinitouch raised concerns that AI should be treated as an enabler, not something that replaces the clinician: “We cannot exclude the importance of the human element in communication between a patient and their clinician. AI is an enabler in healthcare – it’s not a replacement. In the world that we live in, we need a human touch and humanity should always be a part of care. My personal feeling is that we will still require clinical validation and discussion with a patient to design the most appropriate management plan based on their particular needs, wants and their own personal situation.”
Kelly raised a similar point, predicting pushback in adoption if “human touch” is minimised: “It is important to remember that AI does not constitute a product on its own, but rather a capability. By standardising and increasing accessibility, solutions leveraging AI can take on the largest challenge associated with preventative care: the sheer volume and frequency of interactions required between patients and “care providers”. The key challenge here will be to create the trust required to drive behavioural change from patients, a feat that might prove challenging if in-person interactions are minimised.”
Noel highlighted some of the challenges that may arise as we introduce AI at scale to healthcare: “We already see that much of the data from different healthcare systems are contradictory and not aligned. The quality of what you get out is very much dependent of the quality of what you put in. The integration challenge of systems will need to be addressed to get the full benefit that AI might provide within healthcare, especially in relation to the opportunity to build robust risk prediction tools into our systems.
“An impending challenge to deploying AI within clinical practice will be the adoption of this technology by clinicians. By their nature clinicians tend to be sceptical of new technology, especially in relation to how it might impair their relationships with patients. Making case for AI within clinical care and supporting clinicians in training needs careful consideration. Deploying new technology within clinical pathways has always been problematic. We’ve had the ability to have paperless records for 30 years and only now are some hospitals getting EPRs. This demonstrates how long it takes to get technology embedded within clinical pathways and healthcare systems.
“There is also, not unsurprisingly, lots of privacy concerns around AI. At the moment, the output of AI often needs clinical validation because we don’t know if it can fully be trusted. Trust in AI and its ability to be accurate and protect patient confidentially are real concerns – however clinician validation can inform on machine learning and thus help AI systems to become more accurate. However, like anything new(ish), the more we use the technology the more we will learn what works and the potential pitfalls and then we can address them to reduce anxiety regarding these issues.”
The implementation of AI and its impact on the industry was also touched upon by Martin, who commented: “The way this will impact the industry is the need for a patient-centric approach, otherwise there could be scepticism or refusal from patients. We need to make sure that the AI that is patient-facing takes a service design approach to understand patients’ needs, perspectives, and concerns. To collaborate with patients, for them to trust in the process being put forward, in order to effectively implement the benefits of AI for them. It almost has to be symbiotic relationship as AI starts to work for us in new ways.
“Part of this now is the need for robust data and bullet-proof AI governance frameworks. AI models that currently use historical data show racial and other biases. Patient data security is also hugely important. Therefore, patient involvement in policy and processes at every step will help to create a relatively smooth transition and buy-in, as AI becomes more central to patient care and experience.”
Similarly to Martin, for Noel the impact of AI on healthcare will depends on how it is embraced: “Change is difficult within healthcare. But in the future, embracing this technology will be a given. We’ve got rising prevalence of long-term conditions with patients with multiple comorbidities and an increasing ageing population. Added to this, we have better and more innovative technology, including better medications and improved procedures within healthcare. That is all causing a drain on the capacity of services.
“I don't know what level of AI will be embraced within the next 5 years but the pace of change in this area is astounding. Whether we get the full benefit of AI within healthcare will depend on how it's embraced by the people who pay for healthcare, by the people who train clinicians, by the clinicians, and by patients who can see the benefits for themselves.”
As we navigate the potential of AI in healthcare, one thing seems certain: the integration of AI into healthcare at scale won’t be easy - but the opportunities for AI to be used to improve healthcare delivery are vast. These insights have also highlighted that for AI to be adopted at scale, there must be engagement and adoption from key stakeholders. As the healthcare industry steers toward embracing AI, the impact will hinge on whether people feel they can trust it – or will it always need clinical validation? We’d love to know what you think.
Do you want to add your thoughts to a future article about digital health? Email us at marketing@clinitouch.com or drop us a message on Linkedin.
There's no denying it's a big buzzword in digital health right now, but what impact could AI have on the future of healthcare?
We spoke to four leaders from across the sector to share their views on AI's transformative power, from revolutionising personalised medicine to reshaping operational efficiency. They also explore how AI impacts healthcare, while highlighting key challenges in adoption at scale.
Martin Sandhu, CEO & Founder at Nuom, shared his views focusing on how AI will improve personalised care: “AI will have a bigger impact in healthcare than we can possibly imagine. It represents the future of our collective health. However, I feel the trend that will impact patient care most, by being new, novel, and intersecting with us as patients, will be the use of digital twins and precision medicine.
“Digital twins will be AI models of our organs (or our entire body) using our own personal data, to represent how our individual physiology and pathology works. In order to test the safety and efficacy of drugs, or give direction in surgery (for example, what valve to use in heart surgery, and exactly where). As a result, precision medicine will offer optimal personalised suggestions for bespoke treatment.“
Tamsin Chislett, COO at Hyfe, also touched upon the possibilities for AI to improve the personalisation of healthcare: “One of the historical challenges to scaling personalised medicine was the need for specialised infrastructure (high resolution imaging etc) and the lack of sensitive longitudinal data. Acoustic AI has the potential to eliminate both of these barriers. By detecting things like coughing and wheezing, passively and continuously via wearable technologies, a new generation of AI models can design and continuously adjust treatment plans for chronic sufferers (COPD, asthma, IPF) based on real time data. AI will monitor health indicators continuously and unobtrusively and provide risk screening, early detection, personalised advice, personalised treatment regimen and can even alert healthcare providers in emergencies.”
Turning her attention towards using AI to boost healthcare efficiency, Kelly Klifa, CEO & Co-Founder at Ally Health, added: “I believe a lot of value can be created by initially deploying AI to increase efficiency of operational processes involved in delivering healthcare. At Ally Health where we look to optimise in person nursing care delivery at home, AI can be applied in preference-based planning - a form of automated planning and scheduling which allows us to satisfy several preference plans at once (e.g. location for care delivery, clinical skills required, timings etc.).”
Tamsin also drew on the efficiency of AI, sharing her predictions on how AI will speed up drug development and be used for predictive models: “AI is already drastically reducing the time and cost of drug development by predicting how different drugs interact with various biological systems as well as the best molecular structure for the purpose. This will expedite the introduction of new, life-saving drugs to the market. Additionally, new, precise ways to measure things like cough in the real world will accelerate the trials and increase the quality and reliability of data.
“In the recent pandemic we learned the hard way just how consequential it is to be unprepared. AI will play a significant role in predicting and managing future epidemic outbreaks. Using data from various sources, AI can forecast outbreaks and guide public health responses, potentially saving thousands of lives. Acoustic AI lends itself particularly well to this purpose, as cough frequency changes can be detected in real time at scale. This can help predict and respond to things like a flu epidemic as well as the next respiratory pandemic.”
While AI may promise to change healthcare delivery, Noel O’Kelly, Medical Director & Co-Founder at Clinitouch raised concerns that AI should be treated as an enabler, not something that replaces the clinician: “We cannot exclude the importance of the human element in communication between a patient and their clinician. AI is an enabler in healthcare – it’s not a replacement. In the world that we live in, we need a human touch and humanity should always be a part of care. My personal feeling is that we will still require clinical validation and discussion with a patient to design the most appropriate management plan based on their particular needs, wants and their own personal situation.”
Kelly raised a similar point, predicting pushback in adoption if “human touch” is minimised: “It is important to remember that AI does not constitute a product on its own, but rather a capability. By standardising and increasing accessibility, solutions leveraging AI can take on the largest challenge associated with preventative care: the sheer volume and frequency of interactions required between patients and “care providers”. The key challenge here will be to create the trust required to drive behavioural change from patients, a feat that might prove challenging if in-person interactions are minimised.”
Noel highlighted some of the challenges that may arise as we introduce AI at scale to healthcare: “We already see that much of the data from different healthcare systems are contradictory and not aligned. The quality of what you get out is very much dependent of the quality of what you put in. The integration challenge of systems will need to be addressed to get the full benefit that AI might provide within healthcare, especially in relation to the opportunity to build robust risk prediction tools into our systems.
“An impending challenge to deploying AI within clinical practice will be the adoption of this technology by clinicians. By their nature clinicians tend to be sceptical of new technology, especially in relation to how it might impair their relationships with patients. Making case for AI within clinical care and supporting clinicians in training needs careful consideration. Deploying new technology within clinical pathways has always been problematic. We’ve had the ability to have paperless records for 30 years and only now are some hospitals getting EPRs. This demonstrates how long it takes to get technology embedded within clinical pathways and healthcare systems.
“There is also, not unsurprisingly, lots of privacy concerns around AI. At the moment, the output of AI often needs clinical validation because we don’t know if it can fully be trusted. Trust in AI and its ability to be accurate and protect patient confidentially are real concerns – however clinician validation can inform on machine learning and thus help AI systems to become more accurate. However, like anything new(ish), the more we use the technology the more we will learn what works and the potential pitfalls and then we can address them to reduce anxiety regarding these issues.”
The implementation of AI and its impact on the industry was also touched upon by Martin, who commented: “The way this will impact the industry is the need for a patient-centric approach, otherwise there could be scepticism or refusal from patients. We need to make sure that the AI that is patient-facing takes a service design approach to understand patients’ needs, perspectives, and concerns. To collaborate with patients, for them to trust in the process being put forward, in order to effectively implement the benefits of AI for them. It almost has to be symbiotic relationship as AI starts to work for us in new ways.
“Part of this now is the need for robust data and bullet-proof AI governance frameworks. AI models that currently use historical data show racial and other biases. Patient data security is also hugely important. Therefore, patient involvement in policy and processes at every step will help to create a relatively smooth transition and buy-in, as AI becomes more central to patient care and experience.”
Similarly to Martin, for Noel the impact of AI on healthcare will depends on how it is embraced: “Change is difficult within healthcare. But in the future, embracing this technology will be a given. We’ve got rising prevalence of long-term conditions with patients with multiple comorbidities and an increasing ageing population. Added to this, we have better and more innovative technology, including better medications and improved procedures within healthcare. That is all causing a drain on the capacity of services.
“I don't know what level of AI will be embraced within the next 5 years but the pace of change in this area is astounding. Whether we get the full benefit of AI within healthcare will depend on how it's embraced by the people who pay for healthcare, by the people who train clinicians, by the clinicians, and by patients who can see the benefits for themselves.”
As we navigate the potential of AI in healthcare, one thing seems certain: the integration of AI into healthcare at scale won’t be easy - but the opportunities for AI to be used to improve healthcare delivery are vast. These insights have also highlighted that for AI to be adopted at scale, there must be engagement and adoption from key stakeholders. As the healthcare industry steers toward embracing AI, the impact will hinge on whether people feel they can trust it – or will it always need clinical validation? We’d love to know what you think.
Do you want to add your thoughts to a future article about digital health? Email us at marketing@clinitouch.com or drop us a message on Linkedin.
Shaping the future of healthcare by making it easy and accessible for everyone, no matter their location.
