Can tech save healthcare? – Investors’ Chronicle

David Beckam

If there’s one part of the economy ripe for technological assistance, it’s the healthcare sector – with rising patient numbers and not enough doctors to treat them. More than 6mn people were on the waiting list for preplanned NHS treatments in England in December 2021, equivalent to over a tenth of the population. Healthcare workers, meanwhile, are overworked and exhausted. A study by the British Medical Association last year found that over a fifth of doctors were considering leaving the NHS for another career. 

The good news is that interacting technologies are nearing the point at which they can transform healthcare processes and outcomes. Cardiologist, geneticist and digital medicine researcher Dr Eric Topol in his review for the NHS ‘Preparing the healthcare workforce to deliver the digital future,’ published in 2019, said that the world is at a “unique juncture in the history of medicine, with the convergence of genomics, biosensors, the electronic patient record and smartphone apps, all superimposed on a digital infrastructure, with artificial intelligence (AI) to make sense of the overwhelming amount of data created”.

There are three key trends driving innovation. The first relates to scientists’ understanding of molecular biology and genetics. The human genome was first sequenced in 2003, a process that took 13 years and cost $3bn (£2.3bn). Today, it can be done in under an hour for less than $800. Most diseases have a genetic component, from cancer to heart disease, and pioneering diagnostics and treatments are increasingly based on an improved ability to sequence (read out) and even edit DNA. The ease with which gene sequencing can now be performed is set to result in enormous volumes of data being collected for individual patients.

At the same time, people are becoming much more digitally connected. The vast majority of people in the developed world now have a smartphone, and many also have devices such as Alphabet’s (US:GOOGL) Fitbit, Apple’s (US:AAPL) Apple Watch or Amazon’s (US:AMZN) Alexa, which record personal data. As sensor accuracy, portability and costs have come down, enormous volumes of data can now be collected on individuals, which can be put to many uses – good and bad.

AI and machine learning have become much more powerful, supercharging analytical capabilities and enabling computers and algorithms to find meaningful patterns in data (from genome sequences to heart rates and step counts collected from a smart watch). Advances in hardware, driven by Moore’s law, have decreased the cost and increased the speed of computation. Meanwhile, advances in AI algorithms, most notably in deep learning (an advanced form of pattern recognition) and reinforcement learning (a trial and error process), have enabled computers to understand ever more sophisticated patterns of data.

The difficulty for investors is that the companies at the forefront of emerging health and biotech can be prone to excess hype, are often lossmaking, and have a high probability of failure. In the case of biotech, even top scientists can’t predict if their innovations will work, and on top of that there is significant regulatory risk. There has been a big sell-off in many healthcare technology and biotechnology companies over the past year, as the prospect of higher interest rates reduced risk appetite. That has also created a difficult environment for companies to raise money – which many rely on to succeed. 

Investor exuberance over health and biotech led to an IPO boom in 2020 and 2021, which was “an opportunity for an awful lot of shit to reach the market” as Paul Major, manager of Bellevue Healthcare Trust (BBH) put it.  

But back the right players and the rewards could be big. Success for any one of these three trends would be news enough, but the possible synergies between the three has led some to talk of a new technological revolution at the intersection of AI and biotech.


Therapeutic advances from genomics 

Decades of basic science research and innovation in scientists’ understanding of genetics and molecular biology is coming to fruition in a number of therapeutic applications – from diagnostics to new treatments and approaches to chemical manufacturing.

The genetic code contained in DNA can be thought of as a blueprint for how cells grow and behave, holding rich information that could help with diagnosis, prognosis and treatment in a host of disorders. Gene sequencing technologies, such as those developed by Illumina (US:ILMN), permit scientists to read out genetic sequences whether it be in healthy human cells, tumour cells, bacteria or viruses.

There are lots of therapeutic advances stemming from cheap, accurate and fast gene sequencing. Sequencing the Sars-Cov-2 genome enabled Moderna (US:MRNA) and BioNTech (US:BNTX) to create messenger RNA vaccines for Covid-19 in 2020 for the first time. These vaccines work by instructing cells within the body to produce a harmless version of key proteins, derived from the viral genome, in contrast to traditional vaccines which typically inject an inactive virus into the body. 

The success of mRNA vaccines has opened up a whole new treatment avenue with potential application in other infectious diseases, and even some cancers. Zhiqiang Shu, senior biotech analyst at Berenberg Capital Markets, says that BioNTech has had “encouraging early-stage results” using an mRNA-based therapy for lung cancer – the most common type of cancer. The goal here is to help people live with cancer as a chronic disease, rather than cure it, Shu says.

Moderna, meanwhile, is building out its Covid vaccine franchise to treat a host of diseases from influenza to respiratory syncytial virus, and is developing a “very promising” HIV vaccine, which is currently in human trial, according to Jack Torrance, a specialist on the Baillie Gifford Health Innovation Fund (GB00BMVLY145). Similar RNA-based technologies are being examined by companies such as Anylam (US:ALNY). 

Gene editing is also a technology that holds therapeutic potential. These techniques use tools such as CRISPR-Cas9 and viral vectors to permanently alter the DNA in living cells. For the most part, therapeutic applications remain in early preclinical studies, although clinical trials are under way for a number of genetic conditions. Scientists are hopeful that, in time, gene editing will be able to treat genetic conditions such as cystic fibrosis, muscular dystrophy and Huntingdon’s disease, as well as other conditions such as some epilepsies and cancers.  

Gene editing is one of the core pillars of synthetic biology – a way to engineer or re-programme cells to produce new chemicals, for manufacturing and therapeutics, an approach espoused by companies such as Gingko Bioworks (US:DNA)

While therapeutic possibilities are unquestionably exciting in their potential, biotechnology in general is a very difficult area to invest in if you are not a medical expert. Investors should never forget that biology and medicine are complex areas of expertise, and it’s easy for statements such as ‘engineering biology’ or ‘reprogramming the cell’, which leverage an analogy to computer programming or manufacturing, to underestimate or even downplay the challenges involved.


Personalised and digital medicine

Some transformations are closer to reality. Perhaps the greatest impact of synergistic innovations in healthcare will be a clear shift from the ‘one size fits all’ approach that has long dominated modern medicine to a more personalised approach. As data collection grows simpler – aided by everything from gene sequencing to wearable biosensors – doctors will have more and more information pertaining to the biology, medical history and behaviour of each patient. Some foresee a new era of ‘precision medicine’, when this data can be analysed by AI-powered algorithms to yield personalised risk scores and treatment recommendations. Such an approach could even be used to advise on personalised diet and exercise regimes.

Personalised or precision medicine can come in many forms. One is the increased use of smart devices, such as smartwatches, to collect data for patients at home, and extend clinical assessment into the home environment (leveraging advances in both hardware and AI-assisted diagnostics). 

Wearable at-home devices are increasingly being used to help with diagnosis, prevention and treatment too. Apple Watch checks for unusually high or low heart rates and Fitbit aims to help people “build healthy habits that change behaviour and improve health”.  

Some wearables have more clinical uses, such as iRhythm Technologies’ (US:IRTC) Zio device, which received funding from the NHS in 2020, worn to detect irregular heart rhythms quickly and accurately. Anthony Ginsberg, manager of HAN-GINS Indxx Healthcare Megatrend Equal Weight UCITS ETF (WELL), says that company is “redefining the way cardiac arrhythmias are clinically diagnosed”. 

Another form of technology-driven personalised medicine is in the delivery of care itself. Teladoc Health (US:TDOC), for example, aims to provide “whole-person Telehealth” with round-the-clock video-conferencing access to doctor visits for a broad range of concerns. NHS England says that one in four GP appointments are “potentially avoidable” in the UK and telemedicine (the remote diagnosis and treatment of patients) is critical for relieving the burden on the health system.

Digital medicine as applied to mental health is another big boom area. The area currently encompasses more than 10,000 apps, according to an estimate by the American Psychiatric Association, although inevitably some of these offer more serious solutions than others. Apps range from mood tracking, to guided meditation, to text therapy by guided counsellors. Ieso digital health, a private company, is being used by the NHS to deliver a free cognitive behavioural therapy service.   

Some mental health apps, such as Wysa and Woebot, are taking digital medicine a step further by trying to deliver cognitive behavioural therapy via algorithms. Given the apparent early successes of these apps, and the rising number of people meeting criteria for depression, Topol says that although virtual counsellors will never replace human ones “this could turn out to be one of the most important AI booster functions in medicine”. 


AI-assisted drug discovery

The average cost to bring a new drug to market is estimated at $1.3bn, with an average of seven to 12 years from pre-clinical testing to drug approval. To make matters worse, progressing from a phase 1 trial to approval has a success rate of around 5 to 10 per cent, Berenberg’s Shu says. Improvements in the speed with which new drugs are discovered, with an increased probability of success and the identification of likely failures ahead of time, would be a game changer. 

Lots of companies claim to be using AI to realise this goal, including Recursion (US:RXRX), Tempus Labs, Exscientia (US:EXAI) and Benevolent AI, which announced a merger with a European special purpose acquisition company (Spac) in December. These companies use some combination of AI and automated experiments to guide the design and modification of new drug candidates, in effect taking over and speeding up the job of a human scientist. The combination of automation and AI-assisted pattern discovery promises to drastically cut the cost of drug discovery, and potentially identify new classes of drug action.

Clearly, this could have huge implications for the biopharma sector, but the technical challenges are enormous and it’s hard to predict how long it might take. 

Google subsidiary DeepMind has developed an algorithm, called AlphaFold, which may also have significant implications for how drugs are designed in the future. AlphaFold helps predict the 3D shape of a protein, which could make it easier for drug developers to design new drug candidates. DeepMind has made its software ‘open source’, which means it is freely available for others to use. This is allowing researchers to undertake projects that would otherwise be impossibly expensive for them, but also hinders the opportunity for rivals developing similar software to make money out of it.  


AI for medical data and diagnostics

Progress is not just about medicinal advances. AI’s pattern-recognition capabilities can also be put to use in medical diagnostics. One of the earliest research applications of AI to healthcare was in diagnosis from X-rays and other radiology images. The first FDA approval for AI medical imaging was in 2017, but several algorithms focused on different areas have been approved since. There are a number of companies now undertaking deep learning of medical images, such as Wright Medical, which was taken over by medical technology company Stryker (US:SYK) in 2020.  

Similar image-recognition technology is also finding its way into medical devices themselves. For example, Butterfly Network (US:BFLY) offers a portable ultrasound tool that plugs into an iPhone, enabling healthcare practitioners to make more informed decisions earlier, regardless of location or care setting. The company was one of 53 named in Fortune‘s 2021 Change the World list.

More recently, AI has also been applied to less standardised data formats, which computers have traditionally struggled with, such as electronic health records and speech transcripts, leveraging advances in natural language processing algorithms. This could help free up caregivers’ time, 20 to 30 per cent of which is currently spent on paperwork, Major says. 

One issue with AI is that regardless of how effective it is at picking up patterns in data, it can only be as good as the quality of that data. Medical data, particularly electronic health records, are notoriously incomplete and often contain mistakes.  

Deane Donnigan, manager of Polar Capital Global Healthcare Trust (PCGH), says a company she thinks looks particularly promising for the way in which it harnesses AI in diagnostics is Renalytix (RENX), which spun out of EKF Diagnostics Holdings (EKF) in 2018. Renalytix uses a machine learning algorithm that checks blood markers and compares them against patient medical histories to provide a kidney health risk score. The company only recently became revenue-generating, but it has a large addressable market, with diabetes a major cause of kidney disease, providing an incentive for diabetics to test the health of their kidneys. 

High expectations are also being placed on liquid biopsies – blood tests for early pan-cancer detection – with the first test having been approved by the FDA in 2020. Torrance describes this area as “pretty exciting”, citing Illumina’s Grail and Exact Sciences (US:EXAS) as two notable players in the field.

Ed Radkiewicz, chief executive at specialist healthcare investor MARCOL Health, thinks it is speeding up detection processes that will make the biggest difference to healthcare processes. He says around 80 per cent of healthcare spend is currently focussed on treating chronic conditions, and early detection will make the treatment process “easier, less expensive and more effective”.


Medical analytics

Meanwhile, medical analytics companies are helping to digitise administrative patient data for hospitals, a key ingredient for improving care and hospital efficiency. This is an area in which big tech has been making inroads. Recent acquisitions by Oracle (US:ORCL) (Cerner), Alphabet (HCA Healthcare) and Microsoft (US:MSFT) (Nuance) reflect the growth of cloud computing in the medical space – and how hospitals are embracing telehealth and digitisation. 

Ginsberg believes the healthcare analytics sector is still “ripe for consolidation”, anticipating more big tech takeovers, with Medical Data Vision (JP:3902)JMDC (JP:4483)Phreesia (US:PHR) and Personalis (US:PSNL) among the potential targets. With big tech under increased antitrust scrutiny, expanding into new markets looks increasingly attractive. Healthcare – which represents almost 20 per cent of US gross domestic product – is an obvious target.   

While some healthcare investors can become wedded to seeing their investments reach medical fruition, the road to success is a long one and acquisitions often represent a good exit. Finding the right companies is not easy, particularly if they are early-stage in nature. Experts say the greatest predictor of success in companies at the cutting edge of innovation is past success, so look closely at management teams. The downfall of Theranos, led by Elizabeth Holmes, provides a cautionary tale. The macroeconomic outlook is difficult for early-stage growth companies, with the US lifting its benchmark rate for the first time since 2018 last week and policymakers forecasting six more interest rate rises in 2022. But we do know that health innovation is a trend that is not going away, and while many companies will fail, some have the potential for enormous growth. 


Fund / Index (total return) % 1yr 3yr 5yr 10yr
Baillie Gifford Health Innovation -34      
Syncona (share price) -31 -26 31  
HAN-GINS Indxx Healthcare Megatrend Equal Weight UCITS ETF  -25      
L&G Healthcare Breakthrough UCITS ETF -23      
Worldwide Healthcare Trust  -16 15 31 347
Bellevue Healthcare Trust  -11 29 69  
Polar Capital Global Healthcare Trust (share price) 17 34 40 179
Index: MSCI ACWI/Health Care Technology -37 29 72 245
Index: Nasdaq Biotechnology -16 12 18 287
Index: MSCI ACWI/Health Care 15 42 58 292
Index: MSCI World 9 45 58 216
Source: FE Analytics, 16.03.22

Next Post

Opinion | For Lent, I Gave Up English

At first, the idea terrified me. But my apprehension convinced me that this would be a good test both of who I was and what I could do. Giving up English would deconstruct not only an integral part of my identity, but also the way in which I daily build […]
Opinion | For Lent, I Gave Up English

You May Like

Subscribe US Now