AI in healthcare is a subsector of a fast-expanding industry. In 2020-2021, the market was estimated to be worth between $7-9 billion, and it is anticipated to rise at a CAGR of up to 48% to more than $60-90 billion by 2030.
Investors have acted swiftly. In India, healthcare investment is surpassing AI investment in other industries. However, this is likely due to a rise in larger deals, as it appears disproportionate to the rise in total deals.
Globally, 2021 was a record-breaking year, with about $16 billion raised in VC alone. There was a rise in M&A activity (as observed), but a significant portion of this may be attributable to the current trend of SPAC listings, which may be termed IPO activity.
And there are no indications of an end in sight. Even though key events, such as the Russia-Ukraine war, continuous supply chain shocks, and rising prices, have dampened expectations for this year, the first quarter of '22 raised $3.3 billion to match investments made during the same period in 2018.
In 2018, AI ventures accounted for 8.2% of all funds collected in the healthcare industry. By 2021, this share will have climbed to 11.4%. This gain is negligible and is likely mitigated by the concomitant expansion of the $1 trillion healthcare sector.
The Exit?
On the other hand, Healthcare AI departures are beginning to increase. There were 83 global company closures in the year 2021. The median exit size, excluding distress and liquidity, was $52.9 million.
Only one AI departure of over $1 billion since 2013 has been in healthcare. This makes sense, given AI's head start in other digital areas and that healthcare enterprises require a longer time to market and greater initial investment and research and development.
We anticipate a rise in this number on the strength of current momentum.
That's great, but…
What are the Risks?
To address this issue, we must remember that healthcare is a vast sector comprised of numerous industries. The industry has defensive characteristics that can contribute diversification to a broader portfolio, but there are inherent and specific risks. In terms of danger, the application of AI in healthcare as a medical device or service is akin to biotechnology.
Nonetheless, the significant risks associated with AI in healthcare are as follows:
- Highly regulated sector
- Data confidentiality and accountability
- A high standard of proof is necessary for MVP
- Front-loaded R&D expenses and a high rate of attrition
Let’s discuss these considerations further in the following themes.
Understand the AI
Before diving in, it is beneficial to evaluate the big picture and ask the following questions:
What is the necessary solution to the problem?
AI = answer or not?
Do you require an adaptive and learning data interpretation system, or would automation (or another solution) suffice?
How will AI enhance existing best practices?
The next step is to gain knowledge of AI and the viability of the technology by considering:
Inputs
– Data is key.
Process
– How does the platform undergo training?
Output and evaluation
– Is the algorithm visible, and what oversight may be provided?
There may be a requirement for expert advice to comprehend the technology and how it is distinct. In contrast, early-stage organisations may emphasise the team's expertise, experience, and network. Have the concerns above been explored, and can the team create the technology despite obstacles?
Data, data, and data
It is essential to comprehend the data's source, quality, protection, and application to design robust algorithms with low bias based on sensitive information.
Is the data private? Or merely de-identified, and if so, what are the implications of that?
A data breach could drastically affect an individual's insurance rates, career opportunities, and relationships, so privacy is of the utmost importance. This lack of transparency is perilous, and any breach might be catastrophic.
To collect the necessary data, alternatives outside traditional de-identification are being considered. To address the demand side, synthetic data platforms and federated learning, which eliminate the need to transmit and place data in the hands of consumers, are examples of alternatives.
On the supply side, data ownership remains a contentious issue. Could blockchain be the solution, allowing individuals to retain ownership of and potentially monetise their healthcare data?
The value of data to firms and their intellectual property is still up for dispute, but it is already being utilised in clinical studies and business. This is promising, but widespread adoption is unlikely until digital identity is well defined.
Regulation and market strategy
The healthcare industry is highly regulated, and all medical-grade standard gadgets must adhere to stringent clinical criteria. Regulation is a minefield due to the high burden of proof required for approval, and there is tremendous variability between nations.
Does the startup have an awareness of the upcoming obstacles, and does it have a plan to address them?
In each nation, regulatory obstacles are evaluated on two levels.
- First, regulatory permission at the local level.
- Second, payers must be involved.
P.S.
– The healthcare system in the target market has a big impact on this, and sales cycles for each of these are distinct.
Individual permission is difficult to obtain, and the National Institute for Health and Care Excellence's (NICE) generic advice is a help. However, approval here can be lengthy and is not assured.
Clearly, these factors can substantially impact financial estimates, and they should inform the selection of the market, the commercialization approach, and the timeframe.
The long game - or that’s it?
Given these protocols and the substantial weight of data required for success in the healthcare industry, it seems likely that Healthcare AI will follow the same extended timeframes as conventional biotechnology.
The potentially catastrophic repercussions of making a mistake necessitate foregoing the luxury of agile iteration, resulting in expensive and front-loaded R&D. In summary, income or profits may not be realized for some time, and the life-stage of the company should be evaluated in the context of the investment strategy and portfolio as a whole.
However, this approach could be disrupted by the exponential rate of AI evolution. Developing technology may be used to benefit health directly and to revolutionize supporting systems. For instance, the capacity to simplify data gathering and trials could substantially cut the amount of time required to overcome regulatory obstacles.
Watch this sector!
AI and healthcare together are a potent growth combo. Even while there are risks and obstacles to managing as an investor in this sector, the quick rate of development is cause for hope. Definitely, a sector to keep an eye on; exciting times lie ahead!
