“IT LOOKS LIKE your back is a bit tight today. Let’s modify your workout.” The voice is gentle yet commanding, the instructions rolled out in the signature cadence of a physiotherapist. It is also unmistakably robotic. The AI physio issues her commands straight from a smartphone’s speakers. A phone with a camera is all she needs to do her job: select the exercises to suit the patient’s injury, guide him through each session and order corrections when he is not doing something right (bending a knee at the wrong angle, for example). An AI algorithm marks up his body as he moves, which is how it knows when a joint is hurting or the back is stiffer than the day before.

This digital therapist, developed by Kaia Health, a German startup, is, by many measures, as good as a human therapist. One trial, which involved 552 exercises by osteoarthritis patients, found that human therapists agreed with the corrections to exercise suggested by Kaia’s app as often as they agreed with the corrections suggested by other human therapists. In clinical trials patients with back pain using the app improved more than those who got in-person physiotherapy. Making people with injuries bend and twist carries some risks. On that, too, Kaia’s app is no worse than human experts. Less than 0.1% of nearly 140,000 users of the app in studies reported adverse events.

The app is registered as a medical device by America’s Food and Drug Administration (FDA) and also in the European Union. Since 2017 the FDA has approved more than 40 other health apps for problems as varied as diabetes, back pain, opioid addiction, anxiety, ADHD and asthma. They are reviewed under the rules for medical devices, usually in the moderate-risk category (which covers things such as pregnancy tests and electric wheelchairs).

Some European countries are designing special approval pathways that also stipulate how health apps are paid for through their health systems. In Germany health apps can get provisional approval for a year based on preliminary evidence of benefits, which obliges health insurers to pay for them. Apps that provide solid evidence from clinical trials get permanent approval. Twelve have already done so and another 19 are on the provisional list. France and Belgium are copying the German model.

Such apps, known as “digital therapeutics”, hold great promise for common and rare diseases alike. Some are stand-alone products, for which you need nothing more than a smartphone. Others are paired up with wearables and other devices that feed them data from users’ bodies, like continuous glucose monitors. Some of them are available only on prescription or through referral from a health professional.

Brent Vaughan, a veteran entrepreneur in digital health who is currently the chief executive of Cognito Therapeutics, a Boston-based startup, says there have been three waves in the evolution of digital therapeutics. The first was mostly what he calls “nagware”—apps that help patients with diabetes and other chronic diseases manage mundane tasks, such as taking their medication, moving around more, eating suitable food or measuring their blood sugar. The second wave digitised existing therapeutic interventions that had almost no safety risk, such as cognitive-behavioural therapy for insomnia and various mental-health problems (“Repackaging things that we’ve done before face-to-face, and moving them to face-to-screen,” as Mr Vaughan puts it). It is the latest wave of digital therapeutics that he says are genuine medical breakthroughs. These therapies may change the progression of a disease by altering the underlying biological mechanisms, such as rebuilding neural connections within the brain.

Within a year most heart attack survivors prescribed cholesterol drugs are not even picking up their drugs

The nagware apps for chronic conditions may sound boring but can have a big impact on health at population level. “If you want to change behaviour, that’s what it’s going to take,” says Steven Driver, a cardiologist and medical director of digital therapeutics at Advocate Aurora Health, a big hospital group in America. “It’s not going to be me in the office 30 days later, reminding them to eat less and exercise more. It’s going to be someone on Tuesday saying, ‘It’s 5pm, you’re probably home from work and you only have 3,500 steps. If you get up and walk for a half hour you’ll meet your goal.’ That’s what we need and that’s what a digital therapeutic allows us to do.”

Poor adherence to medication regimens is a huge problem, too. “Nobody is more eager to take their medicines than someone who just survived a heart attack,” says Dr Driver. But in just three weeks, adherence to most treatment regimens begins to wane. Within a year most heart-attack survivors prescribed cholesterol drugs are not even picking up their drugs. And because chronic conditions like diabetes and heart disease often come with other health problems mixed in, even conscientious patients struggle to keep track of everything they must do.

AI-coaches help with all that, some with the level of personal detail typical of an organised, doting or even despotic spouse. Blue Star, an app for diabetes, combines data on a patient’s diet, activity, sleep, social interactions and state of mind, as well as medications and tests, automatically bringing in data from the laboratories and pharmacies the patient uses. People can connect it to all sorts of devices such as smart weight scales, continuous glucose monitors, fitness trackers and blood-pressure cuffs. They are told daily what a particular meal, bedtime schedule or exercise does to their blood sugar, with advice on what they should change. Clinical trials have found that BlueStar, when added to patients’ usual care, reduces the amount of haemoglobin A1C (a biomarker of long-term blood-sugar level) by two to four times more than it is reduced by the most widely used diabetes drugs alone.

Such apps could also make a difference for chronic conditions for which existing treatments do not always work. Perfood, a German startup, is testing an AI-based app for migraine that has a personalised-nutrition component. Some studies suggest that for some sufferers a low-glycemic diet may provide as much relief as some of the commonly used migraine medications.

And breathe

Though chronic-care apps are likely to become the blockbuster category of digital therapeutics, some of the most exciting innovation is directed at less common health problems, including some debilitating conditions for which existing therapies are of limited benefit. One example is Freespira, a digital therapy for panic attacks and post-traumatic stress disorder. It consists of a breathing sensor placed in the nose and connected to a tablet that patients use twice a day for four weeks. People with panic disorders breathe in a particular way that leads to a build-up of carbon dioxide, thought to set off the physiological chain that causes panic.

Freespira trains them to normalise their breathing. Acacia Parks, a user whose panic attacks began when her husband was hospitalised after a car accident, is a trained psychologist. She says current treatments available for panic are awful. “You’re essentially pushing yourself towards the thing that’s causing your panic, so that you could purposely induce a panic attack and then learn to cope with it. Nobody wants to do that.” Several clinical trials, though small, suggest that most users have decreased symptoms or are in remission after six months or longer. Ms Parks says the app has helped her a lot.

Even the most ground-breaking therapeutics, digital or otherwise, will not help many patients if they are not a good business proposition for health-care payers, such as insurers, national health-care systems and employers who provide health benefits to their employees. That message is not lost on some of the more established digital-therapeutics companies. They are starting to invest in studies that show their products deliver good value.

In profit-oriented health systems like America’s, some doctors see prescribing digital therapeutics as a way to be more efficient. Such apps provide them with data at a glance on what their chronic-disease patients have been up to, so a consultation can be shorter and focus on pressing concerns.

The smartness of such apps means that, effectively, they set their own dose. That means they can be prescribed to a large population without worrying too much about an individualised response. Some American health-care systems are integrating these therapeutic apps into the system-specific apps that patients already use to book appointments and see test results.

This allows them to “push” new digital therapeutics that may be relevant to patients with certain conditions straight onto their smartphones. Dr Driver’s team at Advocate Aurora Health, for example, recently batch-prescribed an app for pre-diabetes to thousands of patients who had received worrying test results. The team had thought it might need up to five weeks to enrol 250 people, but so many people signed up that enrolment had to be closed after just 36 hours.

Connectivity needed

The two biggest problems that have emerged for health-care providers exploring these possibilities are finding the most suitable apps and making various computer systems—from the patient’s watch to the records system in his doctor’s office—talk to each other. In recent years a cottage industry of specialised firms has emerged to help with that. Orcha and AppScript are just two of the companies that review and rank apps for effectiveness, user experience and privacy, and prepare bespoke digital formularies for specific conditions or patient groups. Companies like Xealth specialise in integrating a range of health-software systems and devices for clients like Welldoc and Advocate Aurora Health to make the flow of data seamless.

Marc Sluijs, a consultant, estimates that about $11bn has been invested so far in the 349 digital-therapeutics companies that he has identified. Most of them are small. The top 20 of them have raised $7bn of that between them. But some of the pioneers in this nascent industry are already going public. Pear Therapeutics, developer of the first FDA-approved health app, a cognitive-behaviour therapy for addiction called reSet, went public in December 2021 in a deal that increased its valuation to $1.6bn. Akili Interactive, which makes video-game based therapies for ADHD, plans to go public in the summer.

The regulator-approved digital therapeutics behind such valuations are being positioned as high-margin products. Some of their features have been patented and their algorithms are proprietary, which protects the firms from having their products copied. In America, Pear Therapeutics sells a three-month course of its insomnia app Somryst for $899. Akili’s EndeavourRx therapy for ADHD is $450 for a three-month course. In Germany most of the digital therapeutics that have been vetted by regulators are priced at €400-500 ($450-560) per course.

For all their promise, digital therapeutics are still a novelty among doctors. Matteo Berlucchi, a “serial digital entrepreneur” who founded Healthily, an AI self-care app and website, reckons that it may take as much as 15 years for digital therapies to be used as much as pills are today. It often takes a decade or two for innovative drugs to become widely used because clinicians and insurers are very conservative groups, says Murray Aitken from IQVIA, a research firm.

That is why the makers of some digital therapeutics are partnering with pharmaceutical companies, which have the sales teams to market their products to doctors. Financially this is small beer for big pharma. Analysts put the digital-therapeutics market at $3.3bn in 2020, when pharmaceutical sales reached $1.1trn. But the market is expected to grow by about 20% a year for the next five to ten years. And such deals offer other benefits. Some pharma companies think digital therapeutics could increase their drug sales by boosting the efficacy of their drugs and the brand loyalty of their customers. If patients exercise, sleep and adhere to their drug regimen, they feel better and may have fewer side-effects. In these medication-centred models the app often has the company’s branding. For some drug makers, the biggest draw may be the fact that the apps generate vast amounts of real-time intelligence on their customers.

Even in a supporting role to conventional health care, though, digital therapeutics could eventually transform medicine. Those that target the workings of the brain are some of the most exciting. MedRhythms, based in Boston, has produced a therapy that uses music to restore movement-related brain connections in people undergoing rehabilitation after a stroke. Sensors attached to patients’ shoes measure gait and feed the results into an AI-based algorithm that mixes a customised beat into a playlist chosen by the patient—like a personalised DJ of sorts. The therapy is in clinical trials for strokes and there are plans to test it for Alzheimer’s, multiple sclerosis and Parkinson’s disease.

Chrissy Bellows, a 74-year-old stroke survivor in Maine, is one of the early-trial participants. She had been told that improvement in her condition was highly unlikely more than two years after her stroke. That meant being dependent on her husband to walk, a few steps at a time. After the Med Rhythms treatment, however, she can walk up to 100 metres without support. Her husband recalls a therapy session in which she was walking towards a set target but suddenly stood still, as if her leg couldn’t move. It turned out the music had stopped. Such is the power it has over the brain’s control of the body.

Cognito’s Mr Vaughan thinks that digital therapeutics will play a big role in conditions that are now hard to treat, such as brain disorders. “I think that 15-20 years from now the idea that you give someone a systemic drug, either a pill or an injection, hoping that some small amount of it gets into the brain and then it actually changes electrical activity in the brain before the rest of it piles up in your liver or kidneys and causes a problem…I think we’re going to look back on that as going after flies with hammers.” ■