pith. sign in

arxiv: 2605.17388 · v1 · pith:JGIKQ62Enew · submitted 2026-05-17 · 💰 econ.TH

The partial adoption trap: Coordination failure, trust, and cultural lock-in in health AI adoption

Ari Ercole This is my paper

Pith reviewed 2026-05-19 22:37 UTC · model grok-4.3

classification 💰 econ.TH
keywords health AI adoptionpartial adoption trapevolutionary game theorycoordination failuretrust failurecultural normsbistabilitysystemic benefits
0
0 comments X

The pith

Health AI adoption often stabilizes in a partial adoption trap rather than achieving full systemic transformation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper models doctors choosing between genuine adoption of health AI, partial adoption, and rejection using evolutionary game theory. It shows the system is generically bistable, meaning a stable equilibrium of partial adoption can coexist with full genuine adoption. Three failure modes enlarge the basin of the partial trap: inability to capture systemic benefits due to non-appropriability, lack of trust in sharing productivity gains, and negative cultural coordination norms among doctors. These effects are strongest for high-value systemic technologies, and standard policies focused only on individual incentives are likely to reinforce the trap instead of overcoming it.

Core claim

An evolutionary game theoretic model of health AI adoption reveals generic bistability, with a stable partial adoption equilibrium alongside full genuine adoption. Genuine adoption is necessary for systemic benefits to emerge above a population threshold. The partial adoption trap's basin is enlarged by threshold coordination failure from non-appropriable systemic benefits, trust failure from the organization's inability to credibly commit to sharing gains, and cultural failure from negative coordination norms. This Value-Adoption Paradox means the most valuable technologies are most prone to the trap. A cost ratchet allows failed attempts to lower barriers permanently, but rapid trust erosi

What carries the argument

Evolutionary game model with three strategies—genuine adoption, partial adoption, and rejection—where dynamics produce bistability and the partial adoption trap is amplified by coordination, trust, and cultural factors.

Load-bearing premise

The model assumes that genuine adoption is required for systemic benefits to materialize above a certain population threshold and that doctors' choices follow evolutionary dynamics leading to bistability.

What would settle it

Empirical observation of whether health organizations that implement coordinated team-level adoption and credible gain-sharing mechanisms achieve full adoption while those relying only on individual incentives remain stuck at partial levels would test the claim.

read the original abstract

Health artificial intelligence (AI) adoption presents a paradox: point-solution tools diffuse readily through clinical populations, yet system-change AI, which carries the greatest potential for pathway-level transformation, consistently stalls at partial adoption. An evolutionary game theoretic model is developed to explain this pattern. Doctors choose among three strategies: genuine adoption, partial adoption, and rejection, where genuine adoption is required for systemic benefits to materialise above a population threshold. The system is shown to be generically bistable, with a stable partial adoption equilibrium coexisting alongside full genuine adoption. The basin of attraction of the partial adoption trap is enlarged by three compounding failure modes: a threshold coordination failure arising from the non-appropriable nature of systemic benefits; a trust failure arising from the organisation's inability to credibly commit to sharing productivity gains; and a cultural failure arising from negative coordination norms among doctors. These failure modes are shown to be most severe precisely for the technologies with the greatest systemic value: the Value-Adoption Paradox. A cost ratchet dynamic implies that failed adoption attempts permanently lower barriers even when embedding fails, but this benefit is offset when trust erosion is rapid. Conditions are derived under which sustained but imperfect adoption pressure is welfare-improving, and the policy architecture required to escape the trap (targeting trust, sequencing, and team-level adoption) is characterised. Standard health system digital transformation policy, which typically addresses only the threshold failure through individual incentives, is predicted to systematically produce the partial adoption trap.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The paper develops an evolutionary game-theoretic model explaining the partial adoption trap in health AI, particularly for system-change technologies. Physicians choose genuine adoption, partial adoption, or rejection, with systemic benefits requiring a threshold of genuine adopters. The model establishes generic bistability between full genuine adoption and a stable partial adoption equilibrium. Three failure modes—threshold coordination failure, trust failure, and cultural failure—enlarge the basin of attraction of the partial trap, most severely for high-value technologies (Value-Adoption Paradox). The paper also analyzes a cost ratchet dynamic and derives conditions for welfare-improving policies and escape strategies from the trap.

Significance. If the bistability and the effects of the failure modes are robust, this provides a valuable theoretical lens on why health AI adoption often stalls, integrating coordination, trust, and cultural factors. It offers specific policy recommendations that go beyond standard incentives, such as targeting trust and sequencing. The identification of the Value-Adoption Paradox and conditions for imperfect adoption pressure being welfare-improving are particularly insightful contributions to the literature on technology diffusion in healthcare.

major comments (2)
  1. [§2.2 (Payoff Structure), Eq. (4)] §2.2 (Payoff Structure), Eq. (4): The step-function threshold for systemic benefits to materialise is load-bearing for the coordination failure and the claimed generic bistability. The skeptic's concern is valid here: if benefits accrue continuously rather than at a sharp threshold, the partial adoption equilibrium may lose stability or the basin enlargement by the other modes may not compound in the same way. The manuscript should include a robustness analysis replacing the threshold with a continuous function to confirm the central claims.
  2. [§3 (Evolutionary Dynamics)] §3 (Evolutionary Dynamics): The incorporation of the trust failure (inability to commit to sharing gains) and cultural failure (negative norms) into the payoff matrix or dynamics needs to be shown explicitly. It is unclear from the description whether these are modeled as parameter shifts or structural changes, and whether they independently enlarge the basin or only interact with the threshold.
minor comments (3)
  1. [Abstract] Abstract: The abstract claims the system is 'generically bistable' without specifying the parameter ranges or conditions under which this holds, which would help readers assess the scope.
  2. [Notation] Notation: Ensure consistent use of symbols for the three strategies (genuine, partial, rejection) throughout the text and figures.
  3. [References] References: Consider adding references to related work on evolutionary games in technology adoption or health economics to contextualize the contribution.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and insightful comments. We address each major comment below and describe the revisions that will be incorporated into the next version of the manuscript.

read point-by-point responses

  1. Referee: [§2.2 (Payoff Structure), Eq. (4)]: The step-function threshold for systemic benefits to materialise is load-bearing for the coordination failure and the claimed generic bistability. The skeptic's concern is valid here: if benefits accrue continuously rather than at a sharp threshold, the partial adoption equilibrium may lose stability or the basin enlargement by the other modes may not compound in the same way. The manuscript should include a robustness analysis replacing the threshold with a continuous function to confirm the central claims.

    Authors: We agree that the step-function threshold is central to generating the coordination failure and the generic bistability result. In the revised manuscript we will add an explicit robustness subsection that replaces the step function with a continuous approximation (a logistic sigmoid function calibrated to produce a comparable transition). We will verify analytically and numerically that bistability is preserved and that the basin-enlarging effects of the trust and cultural failure modes remain qualitatively unchanged. This analysis will be reported in the main text or an appendix as appropriate. revision: yes

  2. Referee: [§3 (Evolutionary Dynamics)]: The incorporation of the trust failure (inability to commit to sharing gains) and cultural failure (negative norms) into the payoff matrix or dynamics needs to be shown explicitly. It is unclear from the description whether these are modeled as parameter shifts or structural changes, and whether they independently enlarge the basin or only interact with the threshold.

    Authors: We thank the referee for highlighting the need for greater transparency. In the revised manuscript we will explicitly derive and display the modified payoff matrices under each failure mode in turn. Trust failure is introduced as a parameter shift that lowers the fraction of systemic benefits appropriated by genuine adopters. Cultural failure is introduced as a structural addition of a negative payoff term attached to genuine adoption. We will provide both the resulting replicator dynamics and numerical phase portraits demonstrating that each mode enlarges the basin of the partial-adoption equilibrium independently of the others, while their joint presence produces the strongest compounding effect. These clarifications will be added to §3 and the associated figures. revision: yes

Circularity Check

0 steps flagged

Model derivation is self-contained with no reduction to inputs by construction

full rationale

The paper sets up an evolutionary game-theoretic model with explicit assumptions on strategies (genuine adoption, partial adoption, rejection) and a threshold condition for systemic benefits to materialize. Bistability and the partial adoption equilibrium are derived directly from the resulting replicator dynamics and payoff structure under these assumptions. No load-bearing step reduces by construction to a fitted parameter, self-citation chain, or redefinition of the target result; the Value-Adoption Paradox and policy conditions follow as consequences within the stated framework rather than being presupposed. The analysis is therefore self-contained as a theoretical model.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

Only the abstract is available, preventing extraction of specific free parameters, axioms, or invented entities from the full derivations. The model rests on standard evolutionary game theory assumptions such as strategy updating via replicator or imitation dynamics and the existence of a population threshold for systemic benefits.

free parameters (2)
  • payoff matrix entries for genuine, partial, and rejection strategies
    Values chosen to illustrate bistability and enlargement of the partial adoption basin; not specified in abstract.
  • population threshold for systemic benefits
    Critical value above which genuine adoption yields non-appropriable gains; central to the coordination failure.
axioms (2)
  • domain assumption Doctors update strategies according to evolutionary game dynamics based on relative payoffs
    Standard background assumption for evolutionary game models; invoked to derive bistability.
  • domain assumption Systemic benefits from genuine adoption are non-appropriable by individual adopters
    Foundation for the threshold coordination failure; stated in abstract.

pith-pipeline@v0.9.0 · 5793 in / 1596 out tokens · 56280 ms · 2026-05-19T22:37:01.994103+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages

  1. [1]

    Fit for the future: the 10-year health plan for England

    NHS England. Fit for the future: the 10-year health plan for England. London: NHS England; 2025 Jul 3. Available from: https://www.england.nhs.uk/long-read/fit-for- the-future-the-10-year-health-plan-for-england/

    work page 2025

  2. [2]

    Adopting ambient scribing products in health and care settings

    NHS England. Adopting ambient scribing products in health and care settings. London: NHS England Digital; 2025. Available from: https://digital.nhs.uk/services/ambient-scribing

    work page 2025

  3. [3]

    Improving drug-therapy decisions through educational outreach

    Avorn J, Soumerai SB. Improving drug-therapy decisions through educational outreach. N Engl J Med. 1983;308(24):1457–1463. doi:10.1056/NEJM198306163082406

    work page doi:10.1056/nejm198306163082406 1983

  4. [4]

    The grammar of society: the nature and dynamics of social norms

    Bicchieri C. The grammar of society: the nature and dynamics of social norms. Cambridge: Cambridge University Press; 2006

    work page 2006

  5. [5]

    Workarounds in electronic health record systems and the revised sociotechnical electronic health record workaround analysis framework

    Blijleven V, Hoxha F, Jaspers M. Workarounds in electronic health record systems and the revised sociotechnical electronic health record workaround analysis framework. J Med Internet Res. 2022;24(3):e33046. doi:10.2196/33046

    work page doi:10.2196/33046 2022

  6. [6]

    Vested interests and resistance to technology adoption

    Canton EJF, de Groot HLF, Nahuis R. Vested interests and resistance to technology adoption. CentER Discussion Paper 1999-106. Tilburg University; 1999

    work page 1999

  7. [7]

    Perceived usefulness, perceived ease of use, and user acceptance of information technology

    Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 1989;13(3):319–340. doi:10.2307/249008

    work page doi:10.2307/249008 1989

  8. [8]

    Beyond adoption: a new framework for theorizing and evaluating nonadoption, abandonment, and challenges to the scale-up, spread, and sustainability of health and care technologies

    Greenhalgh T, Wherton J, Papoutsi C, Lynch J, Hughes G, A’Court C, et al. Beyond adoption: a new framework for theorizing and evaluating nonadoption, abandonment, and challenges to the scale-up, spread, and sustainability of health and care technologies. J Med Internet Res. 2017;19(11):e367. doi:10.2196/jmir.8775 16 The Partial Adoption Trap Ercole (2025)

    work page doi:10.2196/jmir.8775 2017

  9. [9]

    Why does the quality of health care continue to lag? Insights from management research

    Nembhard IM, Alexander JA, Hoff TJ, Ramanujam R. Why does the quality of health care continue to lag? Insights from management research. Acad Manage Perspect. 2009;23(1):24–42. doi:10.5465/AMP.2009.37008001

    work page doi:10.5465/amp.2009.37008001 2009

  10. [10]

    Coordination dynamics in technology adoption: lessons from an evolutionary game theoretical analysis

    Ogbo NB, Han TA. Coordination dynamics in technology adoption: lessons from an evolutionary game theoretical analysis. In: Burrell DN, editor. Multisector insights in healthcare, social sciences, society, and technology. Hershey: IGI Global; 2024. p. 295–326

    work page 2024

  11. [11]

    Workarounds to intended use of health information technology: a narrative review of the human factors engineering literature

    Patterson ES. Workarounds to intended use of health information technology: a narrative review of the human factors engineering literature. Hum Factors. 2018;60(3):281–292. doi:10.1177/0018720818762546

    work page doi:10.1177/0018720818762546 2018

  12. [12]

    Diffusion of innovations

    Rogers EM. Diffusion of innovations. New York: Free Press; 1962

    work page 1962

  13. [13]

    Population games and evolutionary dynamics

    Sandholm WH. Population games and evolutionary dynamics. Cambridge, MA: MIT Press; 2010

    work page 2010

  14. [14]

    Evolutionary games on graphs

    Szab´ o G, F´ ath G. Evolutionary games on graphs. Phys Rep. 2007;446(4–6):97–216. doi:10.1016/j.physrep.2007.04.004

    work page doi:10.1016/j.physrep.2007.04.004 2007

  15. [15]

    Evolutionary game theory

    Weibull JW. Evolutionary game theory. Cambridge, MA: MIT Press; 1995

    work page 1995

  16. [16]

    Tracking medicine: a researcher’s quest to understand health care

    Wennberg JE. Tracking medicine: a researcher’s quest to understand health care. Oxford: Oxford University Press; 2010. 17

    work page 2010