From cognitive overload to institutional escalation: formalizing the supply-demand production condition in patient-safety failure.

Physician exit, unit closures, and regulatory interventions are conventionally treated as separate workforce and governance problems. The workload-patient safety literature documents the empirical relationship between system strain and deterioration in outcomes; the agency literature describes the information asymmetry governing physician behavior; the organizational safety science literature describes how systems drift toward failure under production pressure. What remains insufficiently specified is a formal account of the production conditions that may connect these literatures, and a candidate mechanism through which institutionally determined cognitive demands may translate, via physician decision-making, into outcome deterioration, recursive self-amplification, and observable institutional escalation. This paper offers a formal conceptual specification of that proposed mechanism. We represent clinical outcome quality as a function of the ratio of effective cognitive supply to cognitive demand: O = f(S/D), where S = I · f(WM, e, λ) and D = f(v, c, a). The notation is intended to make hypothesized relationships explicit and generate testable predictions; it is not an empirically estimated or validated cognitive production function. When effective supply becomes insufficient relative to demand, the effort-accuracy trade-off is expected to favor resource-conserving reasoning, increasing the risk of outcome deterioration. Institutions may observe indirect proxies of strain but not the proposed latent S/D condition directly or reliably in real time, creating an information asymmetry that can impede correction. A recursive feedback loop with asymmetric updating is proposed to account for nonlinear degradation and recovery. Drawing on Currie and colleagues' (2012) analysis of institutional work in healthcare and Hirschman's exit-voice-loyalty framework, we propose a four-tier escalation model (T1-T4) culminating in the agent's veto. T1-T4 is intended as a typical cost-increasing pathway rather than an invariant sequence; tiers may overlap or be skipped. A 31-case illustrative inventory across 16 jurisdictions (1999-2026) illustrates that phenomena classified as T4 are documented across diverse contexts but does not validate the causal framework or estimate prevalence. Six testable propositions are derived.