Paula Bellostas Muguerza, Partner at Kearney, explores the growing significance of Artificial Intelligence in the healthcare sector.
It has been a landmark year for Artificial Intelligence. What was once the reserve of science fiction is now becoming an intrinsic part of our everyday lives. From voice-controlled digital assistants in our homes to customer service chat bots, AI is now entrenched in the mass market. Most significantly, it has also been a year in which AI in healthcare has put down roots for a more radical transformation.
AI and machine learning have been quietly revolutionising the health sector for years by delivering everything from robotic surgery and 3D image analysis to intelligence biosensors that allow diagnoses and treatments to be managed remotely. But while the COVID-19 pandemic has been devastating, it has also catalysed technological developments in and awareness of healthcare AI. In the first quarter of 2020 alone, almost $1bn was invested in AI-focused healthcare start-ups and a recent projection shows the global industry growing at a rate of 44% until 2026.
The potential uses of Artificial Intelligence in the healthcare sector are vast, and the technology is rapidly gaining momentum with investors as a result. With its applications ranging from disease prevention and diagnostics to acute care and long-term disease management, the industry is reaching a tipping point in 2020 and AI is finally becoming mainstream.
Yet it still seems we have only scratched the surface; and like any revolution witnessed in real time, the possibilities are seemingly limitless. For healthcare providers and associated organisations, it remains a real challenge to turn vision into reality. To move from testing to regular use, and to change the patient experience more fundamentally, organisations wanting to engage with AI must approach the issue strategically.
The ‘art’ in Artificial Intelligence
The technology behind Artificial Intelligence is evolving at breakneck speed, but the real test of an organisation is how it can harness and implement that technology for its own ends. The pressure of the pandemic has no doubt accelerated innovations, but before we look at how they can be put into practice, it is useful to consider what AI actually is – and what it looks like in a healthcare setting.
At its core AI is machine learning, which is comprised of three cognitive nodes: computer vision, natural language processing and data inference. Computer vision is the ‘eyes’ of AI, as it is capable of recognising visual patterns, objects, scenes and activities in digital imagery far quicker humans. Natural language processing refers to the technology that recognises and understands spoken language. Structured data inference is the technology that uses data, most often numerical, to solve problems. We have seen exciting developments for healthcare in all three in 2020.
Take natural language processing, which has come under the spotlight during the pandemic as healthcare providers have been forced to move operations online. The ‘telehealth’ industry has grown exponentially because it has enabled providers to automate and streamline basic services in order to free up resources to deal with the crisis. In France, for instance, telemedicine appointments increased from 10,000 to a staggering 500,000 per week during the initial peak of the pandemic.
Recent developments in AI show that ‘telehealth’ can be more than a platform for consultation. One startup, Vocalis Health, is exploring the use of voice data as a biomarker for disease progression. Using AI, the technology can detect signs of pulmonary hypertension in specific segments of speech, which can be recorded into a smartphone. Similar efforts are being focused on voice-based COVID-19 screening apps and also on using data to track neurological conditions like Parkinson’s disease. The potential for this is significant and it promises to elevate telehealth to whole new level.
Huge strides in healthcare AI have been made by larger operations too, such as Alphabet’s AI subsidiary DeepMind. In November, DeepMind’s AlphaFold project revealed it had in large part resolved a half-century-old challenge for scientists by understanding how a protein folds into a unique three-dimensional shape. This paves the way for a much greater understanding of diseases and the creation of designer medicines. On a wider scale, it even can help break down plastic pollution. Once more, the implications are enormous and not only for research scientists but for the role of Artificial Intelligence in the healthcare sector as a whole.
AI’s ability to solve incredibly complex problems using huge sets of data far surpasses our own; and for the decades ahead, the sky really is the limit for the businesses pioneering change – so how can a healthcare provider think about effectively building-in such developments into strategy?
The elusive blueprint
Artificial Intelligence is a vast field with many potential applications. There is no single, fool proof blueprint for its implementation, so healthcare organisations looking to harness its potential must make choices that fit their financial and technical capabilities.
The first key question that providers should ask themselves before embarking on their AI journey is: do we have the capacity to build out these capabilities in-house? Having the internal resources, proprietary data and capital to develop AI solutions in-house comes with obvious benefits in terms of control, but businesses will need to decide for themselves whether it’s realistic given their goals and timeline.
Next, should we consider partnerships or acquisitions? Even with the best resources and in-house capabilities, partnerships can rapidly increase the development and deployment of AI systems and tools. Investments in AI start-ups or acquisitions of smaller companies can also give an organisation fast access to development phases and provide greater expertise and capabilities.
Finally, businesses will need to think about which key enablers will accelerate their AI strategy. This means thinking about everything from building or acquiring new technologies, to leadership alignment and team allocation.
The common factor
We know that AI can transform many aspects of healthcare; and as we have seen this year, it is evolving rapidly on a global scale. However, healthcare providers engaging with AI face specific challenges, especially when implementing it.
Data is AI’s raison d’être: without a continuous supply of data, AI technology simply could not have achieved what it has to date. However, it can also be a nuisance for organisations which are grappling with the challenge of ‘dirty data’, which is not yet standardised and remains disparate. Privacy protocols and security requirements present additional barriers to progress, but as they concern protections for patient rights, these are hills that must be climbed. Consent for the use of patients’ data and the need to address perceived bias in algorithms are additional ethical issues of which all organisations must be wary.
Necessity is the mother of invention, which explains in part why so much ground has been made this year. However, the healthcare business model could do more to incentivise innovation. While there is a broad range of industry players in this sector, larger technology companies are known to lure talent away from start-ups, who also face difficulties scaling up their products without partnerships.
These challenges are certainly real, but they are by no means insurmountable. While the success of engaging with AI relies on careful preparation, it is an innovation that is not just worth pursuing, but one that will be integral to healthcare’s story in the years to come. As such, organisations need to prioritise AI initiatives and plan for implementation. On a basic level, this means ensuring leadership is on board and the right talent is being supported.
Many organisations throughout the healthcare chain are already deep into their digital transformation journey. While some of these will have well-developed AI strategies in play, others will not. It is worth bearing in mind that the road to AI-enabled healthcare is long, which makes having a strategy to turn vision into reality key to a successful journey.
Overall, approaches may vary and will be dependent on specialism and sub-sector. But what sets healthcare ahead of other industries is the universal recognition of the power of AI and machine learning, and the sheer scale – from start-ups to multinational companies – involved.
The medical landscape of tomorrow is likely to look very different, but it is down to healthcare organisations across the board to steer their own path in a future defined by Artificial Intelligence.