PREDICARE Medico-Economic Note · v2 · May 2026 · 12 pages
The French healthcare system is no longer principally organized around the acute episode: it is organized around the chronic trajectory. Long-term conditions (ALD) now concentrate approximately 67% of compulsory health insurance expenditure for 17% of the population, and this proportion is progressing at constant behavior. In this context, the cost of non-surveillance is no longer a diffuse externality — it becomes an identifiable envelope of €2 to €3 billion annually on potentially avoidable hospitalizations alone, excluding medico-social downstream effects.
This note defends three linked propositions: (i) the principal lock on predictive prevention in France is institutional, accounting-related, and conventional, not technological; (ii) a territorial predictive vigilance infrastructure is an infrastructure of continuous medico-economic visibility production, not an additional digital device; (iii) no systemic transformation is viable without an explicit operational vehicle — and such a vehicle can be built by coherent composition of existing instruments, without constitutional upheaval.
An economy that does not know how to measure its avoidances can only finance its occurrences.
Public decision-makers (DGOS, DGS, ministerial cabinets), hospital financial directors, regional health authorities (ARS), the National Health Insurance Fund (CNAM, strategy directorate), and long-term healthcare infrastructure investors — particularly regulated infrastructure funds and long-term insurance actors whose risk profile matches the actual nature of this asset class.
The system’s central pathology is not a moral malfunction. It is a rational equilibrium: the operator who invests in surveillance, coordination, and longitudinal follow-up bears immediate costs whose profitability does not return to them. The savings their practice produces appear on lines they do not control — hospital envelopes, downstream ALD expenditure, medical transport, rehospitalizations, medico-social dependency. The economic beneficiary is not the economic producer. Until this lock is lifted by an explicit mechanism of capture and redistribution, no preventive reform will hold.
Five structural locks articulate this pathology: time misalignment between clinical horizons (5 to 15 years) and institutional horizons (political mandates of 2 to 5 years, annual ONDAM); activity-based pricing (T2A) as an intrinsically anti-preventive mechanism; the trap of derogatory devices (Article 51, FIR) that absorb reformist energy without permitting systemic transformation; the budgetary prudence circle that demands proofs whose production is rendered structurally difficult by the architecture itself; the fragmentation of arbitration authorities. None resolves itself independently of the others.
The note proposes a concrete mechanism articulating five coherent elements. A ring-fenced ONDAM line over 5 to 7 years — ring-fencing is not a fiscal exception; it is a condition of longitudinal observability. Territorial contractualization through the existing ARS and CPOM instruments. An HAS-CEESP imputation doctrine adapted to longitudinality, which admits the informational value of comparisons between observed trajectories and simulated counterfactual trajectories without conflating it with causal proof. A territorial return mechanism for captured savings (functional analogue of ACO shared savings, adapted to French conventional grammar). A trajectory of progressive T2A reallocation across the chronic-disease lines where it produces its most counterproductive effects.
None of these instruments is new. It is their articulation that constitutes the suspended decision.
The value capturable by a private actor is not principally technological: it is positional, contractual, and regulatory. Software commoditizes. What does not commoditize is the position of territorial infrastructure operator holding the longitudinal data, the stable methodological protocol, the multi-year contract, and the recognized imputation doctrine. This asset belongs to the profile of regulated infrastructure concessions (water, energy, transport, territorial telecoms), not to the SaaS publisher. Confusing the two investor profiles is a frequent cause of strategic failure in this asset class.
PREDICARE · predictive medicine · health economics · activity-based pricing (T2A) · shared savings · Medicare Advantage · Accountable Care Organizations · risk-adjusted capitation · longitudinal vigilance · health infrastructure · avoidable hospitalizations · private capturability · concession of economic observation