Nov 20, 2024

Nov 20, 2024 • Stephen Hurford Telemedicine didn't solve the fundamental cost problem in healthcare, but new AI models may be tackling this. The CEO of Abi Global Health thinks that AI is the key to making telemedicine work as a business model. Telemedicine emerged as an exciting new prospect before the pandemic, with insurers adding it to the services they could offer customers. But, says Kim-Fredrik Schneider, nobody stopped to consider a crucial flaw in the model. “None of [the telemedicine startups] directly attacked that fundamental unit of cost, which is the time of a healthcare professional,” he says. Covid saw healthcare providers eagerly turn to telemedicine and digital health as ways to see patients while enforcing social distancing. In 2021, at the height of the pandemic, funding rounds for telemedicine startups surged, with investors putting in more cash even as there were early signs in the US that demand might be falling as restrictions eased. Beneficiaries of the boom period include Cerebral, a US mental health telemedicine unicorn that raised $300m in 2021 in a round led by SoftBank Vision Fund 2. In the same year, Capsule, a digital pharmacy startup, raised $300m; Kry, a Swedish provider of virtual and in-person healthcare, raised $312m. But in 2022 the sector saw a downturn. Cerebral laid off about 20% of its workforce, Capsule 13%, while Kry pulled its operations out of Germany and made around 400 redundancies. Schneider says that if investors in 2021 saw telemedicine as the future of post-pandemic healthcare, they were mistaken. “What’s wrong with that business model that it took a global pandemic for people to actually use the service?” Kim-Fredrik Schneider, Abi Global Health “I would look at it the other way around,” he says. “[And I would ask] what’s wrong with that business model that it took a global pandemic for people to actually use the service?” Schneider believes that companies like his, Abi Global Health, can use artificial intelligence to finally achieve the efficiencies the telemedicine sector originally hoped for. The company uses a machine learning algorithm to match patients with doctors in their area whose specialties suit their needs. It also uses AI to support healthcare professionals during the consultation and to help with administration and quality assurance afterwards. “We’re delivering on the promise of technology, which is that you get more for less.” It is just one startup among many that are using AI in healthcare provision. With advances in sophistication coming rapidly, the hope is that trained models can provide more efficient or effective care at a time when healthcare systems across the world are under pressure from ageing populations and an undersupply of medical personnel. Uniqa Ventures, the CVC arm of the Austrian insurance company Uniqa is an investor in Abi Global Health. It co-led its $4.9m series A funding round in April 2023. Bori Farkas-Fozy, principal at Uniqa Ventures, says she sees a lot of potential for AI medical technology to improve the offerings of insurance companies. “Speeding up the [healthcare provision] process has multiplying positive effects,” she says. “We know that if care is delayed there can be complications, making it costly both for the patient and the insurer.” Her fund has also invested in Ifeel, a Spanish startup that provides employee assistance mental health care. It surveys a workforce and uses AI to triage them based on the results to decide what level of care, if any, is necessary. But she says that the insurance industry needs to move slowly when considering the offerings of healthcare companies using new technologies. “In healthcare, we see tremendous application opportunities. But it is happening with a great deal of caution, and rightfully so.” Kim-Fredrik Schneider similarly stresses that there is a danger of getting caught up in hype. “AI is not a panacea,” he says. “It’s something that should be selectively deployed to solve a good problem.” A disruptive cure AI-assisted diagnostic tools could also make a difference for the life insurance industry. Lapetus, a US startup that uses machine learning to analyse a patient’s face as a biomarker for their health, is targeting the life insurance providers for its services. Lapetus CEO, Karl Ricanek, says it provides a way of assessing a person’s longevity and giving early signs of conditions they may be suffering. “We take a medical selfie and then quickly extract information about your health and wellness, as well as demographic information. So that’s available in a matter of seconds versus sending someone over to your home to do that.” Ricanek says the technology could assist medical underwriting, where a patient’s life insurance policy is valued according to their health and life expectancy. But he says the life insurance sector has been slow to take on novel technologies like AI. “We’re the poster child for what’s wrong with the insurance industry. Adoption should have been worldwide by now.” Karl Ricanek, Lapetus “We’re the poster child for what’s wrong with the insurance industry,” says Ricanek. “We’re 10 years into [development of Lapetus’s technology] and adoption should have been worldwide by now.” “But what we’re now seeing is that insurance companies have no choice but to start thinking about reducing the cost of insurance. That’s where we come in.” Gen Re, the US multinational reinsurance company, is one of the companies that has partnered with Lapetus. It has offered Lapetus’s technology to life and health insurance applicants. The startup has deployed in Africa, the Middle East and the US so far. Ricanek says the training data has been specifically chosen to contain a broad coverage of skin tones and socioeconomic backgrounds, both of which can affect the inferences drawn from a facial analysis. And it goes beyond lower costs. There’s also a need to keep pace with the kind of technological convenience seen in other sectors. “We have very sophisticated end users,” says Ricanek. “They are not going to settle for a paper form. They’re not going to settle for [having to visit] a medical clinic. They expect things to be easy.” Schneider says that insurance companies have three “chronic” problems that a startup can make itself very valuable by solving. These are customer engagement, product differentiation, and the loss ratio. In other words, a startup needs to offer a strong, unique end user experience and should improve the insurance company’s profits. Farkas-Fozy of Uniqa Ventures agrees. “I think the more you can offer a policyholder, the higher your retention or acquisition of new customers.” Elsewhere, there are signs that the telemedicine market is entering a new AI-driven phase. Hippocratic AI, a US startup building generative AI healthcare chatbots, raised $53 in series A funding in March. In a September extension round it received another $17m, this time led by NVentures, the CVC arm of Nvidia, the world’s leading AI chip designer which has been recently taking an increased interest in healthcare startups. The Swedish startup Kry, which had to scale back its workforce and shut down its German operations in 2022, offers a hopeful sign. After that turbulent year, it managed to apply generative AI techniques to its business model to drastically reduce admin costs associated with each virtual visit. In February, Sifted reported that the company had turned a profit for the first time. Stephen Hurford