The end of captive demand

Much of healthcare's commercial architecture rests on demand that could not question, compare or leave. AI is dismantling the ignorance that held it captive. What remains is what deserves to.

The quiet assumption underneath the business model

Healthcare economics has known for sixty years that this market is not like other markets. Arrow's foundational analysis identified the reason: information. The patient cannot easily judge what they need, whether it worked, or what it should cost; the clinician and the institution know more, and the whole architecture of the sector, from professional licensure to insurance itself, grew up around that asymmetry (Arrow, 1963). Later economics sharpened the picture: when one side of an insurance market knows more than the other, the pool itself becomes unstable (Rothschild and Stiglitz, 1976), and real markets reveal that private information runs in more than one direction at once (Finkelstein and McGarry, 2006).

What is said less often is what that asymmetry did commercially. It made demand captive. A patient who cannot evaluate the recommendation does not shop. A referral that flows through habit and hierarchy is not chosen. A price that cannot be compared is not negotiated. Whole revenue lines, in every health system we work in, were built not on being demonstrably better but on being structurally unquestioned. That is not an accusation of bad faith. It is simply what any commercial system does when its customers cannot see: it stops having to earn what it collects.

What AI is actually disrupting

The current wave of AI is routinely described as disrupting diagnosis, documentation or discovery. Underneath those, it is disrupting something more structural: the distribution of knowing.

Continuous monitoring has begun moving clinically meaningful information to the patient first. In the Apple Heart Study, more than 419,000 ordinary smartwatch users were screened for irregular pulse; among those alerted and assessed, a third had atrial fibrillation confirmed, and 57% of notified participants contacted a health professional as a result (Perez et al., 2019). Whatever one thinks of the sensitivity of consumer devices, the direction is unambiguous: the first signal increasingly arrives with the person, not the institution. Accessible AI does the same for interpretation. The patient who once arrived with a symptom now arrives with a differential, a summary of the evidence, and questions about the tariff.

Every one of those arrivals is a small subtraction from captivity. The asymmetry Arrow described does not vanish, but it narrows, and it narrows fastest exactly where margins were most protected by it: elective and shoppable care, referral-routed volumes, and insurance pools priced on the assumption that the insurer knows the risk better than the insured. When knowledge moves to the individual, the insurance problem inverts too: the conditions for adverse selection reappear in modern dress, with the monitored patient understanding their own risk in ways the community-rated pool cannot price (Rothschild and Stiglitz, 1976; Finkelstein and McGarry, 2006).

What the exposure reveals

When demand stops being captive, the question every board should sit with is uncomfortable and clarifying: which of our revenues were earned by excellence, and which merely by opacity?

The sector's own supply side gives a warning about what happens when claims outrun evidence. An analysis of 224 digital health companies found 44% had completed no clinical trials or regulatory filings at all, and that the strength of a company's clinical evidence bore essentially no relationship to the claims it made or the funding it raised (Day et al., 2022). Captive demand tolerated that. Informed demand will not, on either side of the market: not from vendors selling to health systems, and not from health systems selling to increasingly knowledgeable patients and payers.

This is why we read the disruption as an opportunity rather than a threat, on one condition. The organisations that thrive will be the ones that treat the newly informed patient as a design input, not a channel problem: evidence as the organising discipline, outcomes that survive comparison, prices that survive daylight, and an encounter good enough that a person who could go anywhere chooses to stay. Demand that stays because it wants to is worth more than demand that stayed because it could not see. It refers, it returns, and it compounds.

The test

The strategic work, then, is not defending the asymmetry. That era is closing, and defending it converts quiet advantage into visible exposure. The work is redesigning the model so it no longer needs the asymmetry: care models built around the encounter, economics built around demonstrated value, and technology woven in where it earns its place. Healthcare, human by design, is also a commercial thesis. What is designed for the patient who can see survives the patient who can see.

References

1. Arrow, K.J. (1963) 'Uncertainty and the welfare economics of medical care', American Economic Review, 53(5), pp. 941-973.
2. Day, S., Shah, V., Kaganoff, S., Powelson, S. and Mathews, S.C. (2022) 'Assessing the clinical robustness of digital health startups: cross-sectional observational analysis', Journal of Medical Internet Research, 24(6), e37677. https://doi.org/10.2196/37677
3. Finkelstein, A. and McGarry, K. (2006) 'Multiple dimensions of private information: evidence from the long-term care insurance market', American Economic Review, 96(4), pp. 938-958.
4. Perez, M.V., Mahaffey, K.W., Hedlin, H., Rumsfeld, J.S., Garcia, A., Ferris, T., Balasubramanian, V., Russo, A.M., Rajmane, A., Cheung, L., Hung, G., Lee, J., Kowey, P., Talati, N., Nag, D., Gummidipundi, S.E., Beatty, A., Hills, M.T., Desai, S., Granger, C.B., Desai, M. and Turakhia, M.P. (2019) 'Large-scale assessment of a smartwatch to identify atrial fibrillation', New England Journal of Medicine, 381(20), pp. 1909-1917. https://doi.org/10.1056/NEJMoa1901183
5. Rothschild, M. and Stiglitz, J. (1976) 'Equilibrium in competitive insurance markets: an essay on the economics of imperfect information', Quarterly Journal of Economics, 90(4), pp. 629-649.