Purpose

This note synthesizes lessons reviewed in Part 1 of this series from six historical cases in which disease metrics drove transformative, system-level change. In Part 2 we draw direct implications for the design of the Biomedical Health Efficiency (BHE) framework. The goal is to identify where historical precedents provide a tailwind, and where BHE is paving new ground.

The Precedent Cases: What We Know About Metrics That Work

Across six precedent cases — infant mortality, blood pressure, HIV, HbA1c, LDL-cholesterol, and hospital readmission rates — a consistent pattern emerges. A metric earns systemic power when five conditions are simultaneously present:

Where BHE Aligns with Historical Precedent

Metrics as system organizers, not just measurement tools

The most important insight from the historical cases is that the most powerful metrics do not just measure — they organize entire systems around a shared concept of what 'better' looks like. IMR organized public health infrastructure investment. Blood pressure organized drug development, screening, coverage requirements, and guideline cycles. HbA1c organized EHR alert systems, pay-for-performance contracts, and specialist referral thresholds.

• Implications for BHE: Every metric should be traced to a decision that drives system change. Metrics that do not improve decisions are just data. BHE frames its metrics not as “what we are measuring” but as “what decisions these numbers will improve and for whom.”

The cascade structure: the HIV 90-90-90 analogue

The 90-90-90 framework — 90% of people with HIV knowing their status, 90% of those on treatment, 90% of those virally suppressed — is the most direct historical analogue to what BHE is building for AD. Its power comes from mapping system performance across the entire care continuum in a single, communicable picture. It clearly shows where the pipeline breaks, focuses intervention, and gives funders and policymakers a single organizing framework.

• Implications for BHE: A similar AD cascade (such as W% of at-risk patients screened; X% with confirmed diagnosis; Y% with a documented care plan; Z% of eligible patients treated) would give BHE the same communication power. This four-number cascade serves as the public communication layer of BHE for early AD.

How metrics gain accountability structure

HbA1c and hospital readmissions both illustrate the same lesson: what moved systems was not the metric in isolation, but the metric embedded in an accountability structure with public reporting and financial consequences. HbA1c became a HEDIS measure; readmissions became a CMS penalty program. In both cases, the accountability structure transformed the metric from academic to operationally compelling.

• Implications for BHE: Laying the groundwork for the development of quality and process measures linked to reimbursement or other incentives is the mechanism by which BHE metrics achieve systemic leverage. This is not simply aspirational; the workstream should treat this as a parallel strategic track, not a downstream hope.

Where BHE Faces a Harder Challenge Than Historical Precedents

BHE measures system readiness, not a single outcome indicator

Every historical example — IMR, blood pressure, HbA1c, viral load, LDL-C, readmissions — is ultimately a single number that proxies one underlying phenomenon for evaluating healthcare quality (based upon the Donabedian framework of structure, process, and outcome). The metric's simplicity is part of its power: it fits on a lab report, a CMS penalty schedule, a country's annual health report.

BHE is attempting something structurally more complex: a multi-domain system performance framework where six readiness domains must be tracked simultaneously and their interactions understood. The closest analogue — hospital readmissions — achieved multi-domain behavior change but was anchored to a single outcome number with an enforcement mechanism (financial penalties) attached.

• Implications for BHE: BHE needs both the six-domain analytical framework and a simplified communication layer (e.g., AD cascade) that can serve as the public accountability anchor. The challenge is developing a rigorous framework that can be communicated simply.

No single 'viral load moment' yet for early AD

The blood pressure and viral load examples share a critical feature: there was a pivotal clinical research initiative that proved the number mattered — the VA Cooperative Study for hypertension, HPTN 052 for HIV treatment-as-prevention, the Framingham Heart Study for LDL-cholesterol and reducing risk of coronary heart disease. Those moments gave metrics legitimate clinical authority, justifying the need to reorganize the system.

For early AD, that moment is still unfolding. The lecanemab and donanemab trials are recent, real-world effectiveness data is accumulating, and blood-based biomarkers are not yet standardized across platforms or universally validated in primary care settings.

• Implications for BHE: BHE is being designed as evidence is still being generated rather than afterwards. This presents an opportunity in that BHE can shape what evidence is generated in order to enable decision-making. However, it also creates risk in that metrics may need to evolve as science matures. Phase 1 metrics should be robust to scientific evolution, focusing on the more stable metrics associated with process and access, with explicit provisions for Phase 2 revision as clinical evidence consolidates.

The 'wrong pocket' problem creates a structurally different incentive landscape for early AD

The “wrong pocket” problem is the structural or economic challenge where the entity that pays for an investment does not get to capture the financial savings or benefits created by it, rather they generate benefits for a different entity. Across historical examples, the entity that bore the cost of improvement was eventually compelled—whether by alignment or coercion—to capture the resulting benefit, or else remained accountable to a payer who did. The CMS readmissions penalty is the clearest example: hospitals bore the cost of care coordination investments and also faced the financial penalty for poor performance. The incentive loop, while imperfect, was closed.

Detection and diagnosis of early AD present a structurally different incentive landscape. The payer who funds early detection (primarily Medicare and Medicaid) may capture downstream cost savings from delayed progression, but the primary care provider who does the work is reimbursed inadequately, the specialist system may be capacity-constrained, and the biopharma sponsor who developed the therapy needs the patient identified in the first place. The benefit flows to stakeholders who do not fully bear the detection and diagnosis cost.

• Implications for BHE: BHE metrics in the payment domain need to be designed to make the ‘wrong pocket’ dynamic visible and quantifiable. The purpose is not just to describe it but to generate the evidence needed to redesign the incentive structure.

Process metrics outpace biomarkers in Goodhart vulnerability

Process and performance metrics often generate gaming, misrepresentation, and tunnel vision when tied to rewards or control. For example, Goodhart failure mode was seen with the CMS 4-hour pneumonia antibiotic rule driving overdiagnosis and unnecessary use of antibiotics. Biomarkers are not fully immune to Goodhart’s Law but are less able to be gamed at the population level. However, this patten can be seen when applied rigidly to individuals, such as when aggressive HbA1c targets drive hypoglycemic risk in patients or when continuous glucose monitoring diabetes metrics which may encourage insulin stacking, threshold skirting, and alarm-dependent dosing that improve the number while obscuring clinically important dynamics and risks.

BHE's system-process metrics — percentage of patients receiving cognitive assessment, diagnostic pathway completion rate — are structurally more susceptible to gaming than clinical biomarkers. A site could document an assessment without meaningfully conducting it. A diagnosis could be coded to hit a rate target.

• Implications for BHE: BHE metrics should have outcome anchors linked to downstream results wherever possible, not just process rates. The framework should also be built in explicit surveillance to determine whether any metrics show signs of being gamed and a governance structure empowered to retire or revise it.

Design Principles Derived from Historical Precedent

Drawing these observations together, the historical cases suggest six design principles for the BHE metric framework:

1. Anchor every metric in a named decision and decision-maker. If no one can act on the number, it does not belong in the core set.
2. Build a cascade metric as the communication layer. The full six-domain framework provides analytical depth; a four-step AD care cascade (screened → diagnosed → care planned → treated) provides the public accountability anchor that can mobilize payers, legislators, and funders.
3. Pursue supporting the development of quality and process measures linked to reimbursement and other incentives as a parallel strategic track. Not downstream aspiration but a concurrent workstream, starting with identifying the 2–3 BHE metrics most likely to meet public accountability technical criteria within 3–5 years.
4. Design Phase 1 metrics to be robust to scientific evolution. Process and access metrics are more stable than biomarker thresholds. Build in an explicit revision protocol for Phase 2 as the clinical evidence base for early AD advances.
5. Make the 'wrong pocket' problem quantifiable, not just visible. Payment domain metrics should generate the economic evidence needed to redesign incentive structures — cost-avoidance modeling, prior authorization approval rates, ROI by stakeholder — not just describe the problem.
6. Build Goodhart surveillance into governance from day one. Every metric used for accountability needs a regular review cycle that asks: is this metric being gamed, the data being distorted, or selective attention being paid to the metric instead of the goal? The governance structure must be empowered to retire or revise metrics that fail this test.

Summary: BHE vs. Historical Precedent at a Glance