Dr Mary Coghlan. Image: EY
Wearables are far from new, but the evolution of generative AI could help take the analysis of their data to the next level.
In 2009, Fitbit launched its first wearable step counter, doing exactly what it said on the tin and encouraging users to keep track of how much they moved on any given day. Within just a few years, the Samsung Galaxy Gear and the Apple Watch hit the market as smartwatches.
Since then, the wearables market has exploded, with many companies jumping on the watch or wristband-type device to track any number of metrics from calories burned to hours slept. The medtech industry in particular has been capitalising on the use of wearables to track patients’ blood pressure, glucose, heart activity and many more.
As technology continues to advance, there’s no doubt that the wearables industry shows huge promise for improving healthcare. Dr Mary Coghlan is a partner at EY Ireland where she leads the health data and analytics team. She told SiliconRepublic.com that the increasing use of these wearables can improve both individual and population-level health outcomes.
“Population analysis powered by the data from wearable technology can enable risk stratification according to specific lifestyle and risk factors and ‘give advice’ based on an individual’s data-driven wellbeing profile,” she said. “In certain cases, wearable devices could even be lifesavers if enabled to detect and alert critical incidents to appropriate clinical services.”
‘Siloed data is all too common across the health system with monitoring devices’
While adoption is still at an early stage globally, Coghlan said the devices can be seen to improve medical intervention with individual data and analytics able to pick up irregularities quickly. But, as the greatest value lies in the data and analytics side of things, a siloed approach to healthcare could stand in the way of reaping the biggest rewards. “This is a technical infrastructural challenge as well as a privacy and data management issue, with a bit of clinical governance thrown in for good measure,” she said.
“While members of the population can wear a 24-hour blood pressure monitor to guide treatment of hypertension for example, typically the device won’t ‘speak’ directly to health ICT infrastructure, so the data generated cannot be leveraged in real time or easily scaled for population level analysis. Instead, it needs to be downloaded, uploaded and analysed. This type of siloed data is all too common across the health system with monitoring devices.”
A siloed system
Ireland’s current health IT infrastructure doesn’t allow the required flexible data sharing needed to effectively capitalise on this data. But challenges of privacy and security remain a challenge, particularly since health data is considered more valuable to cybercriminals than financial data.
Despite these concerns – particularly following the 2021 HSE cyberattack – Coghlan said EY’s Consumer Health survey showed that people are ready for their data to be shared in the right way to maximise health outcomes.
“If anything, I believe that there is currently a potential level of frustration that data is not adequately shared, for example, through having to repeat the same details to numerous clinicians in health settings given the lack of connectedness of the data and systems,” she said.
“Of course, different people will have different appetites for the extent to which their data is shared. The right balance can be found through transparency about data usage and assurance that the appropriate level of investment has taken place in data security – these combined with a clear articulation of the direct benefits of data-driven healthcare innovation are a compelling proposition.”
The enormity and complexity of the healthcare system means clinical decision-making requires a deep understanding of many variables. With this mind, Coghlan said the application of data and analytics need to be core components of a data-driven health service.
“These cutting-edge tools can also help address the fundamental issue of the all too frequent mismatch between demand and capacity in key areas of the healthcare system for example. They can incorporate expected future evolving demand patterns and the re-configuration of capacity requirements in a relatively safe digital environment,” she said.
“Capacity modelling evaluates workforce availability, the physical capacity of facilities and the availability of specialist equipment to align resources with anticipated demand. These models are effectively digital twins and can identify potential bottlenecks and inefficiencies within a health ecosystem. They can also incorporate new interventions to enable an understanding of what will work in the real world to solve system or service challenges.”
However, she warned that demand and capacity models are “only as good as those who fine-tune them with historical and emerging data”.
The advances of generative AI
While the tech and data behind wearables is not new, the advances in generative AI are and they can bring about the next stage of evolution to data analytics in healthcare.
“While AI has been deployed in areas such as mammography for some time, generative AI can potentially detect and explore new and more complex patterns in the outputs from wearables that can be applied to a range of different scenarios to see if there’s a correlation between them,” said Coghlan.
“Data beyond pure diagnostics, lifestyle risk factors for example, can be incorporated into algorithms and combined with data. For example, with mammography, to enable us to understand far earlier the likelihood of developing breast cancer – and therefore the ability to intervene through lifestyle or other factors to manage this risk optimally.”
Coghlan added that generative AI also has the potential to accelerate the adoption of digital twins as core planning tools in healthcare, which could have a positive impact on the cost of health services. “Using intelligently deployed digital twins to inform planning has the potential to ensure that the right return on investment in terms of health outcomes and patient experience are achieved more effectively and at a lower cost.”
Outside of generative AI, Coghlan said she’s also excited about the development of wellness analytics. “This is where the integration of population level insights combined with individual health data can guide lifestyle decisions and choices to maintain and improve wellness. Given our ageing population, extra years of good health can lead to a direct benefit for our health services and of course, the happiness and prosperity of society as a whole,” she said.
“Related to this, an area I believe we can make huge advances in the coming years is screening analytics. This is where we get better at leveraging data assets to predict the highest risks to individuals and therefore create the potential for intervention.”
Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.
