Every open enrollment season, benefits leaders say the same thing: "Get your screenings. Early detection saves lives and lowers costs."
For traditional screenings (mammograms, colonoscopies, A1C checks), that logic works fine. CPT codes exist. Plan documents are settled. The administrative pathway is smooth.
But there's a revolution happening in early detection, and it's quietly colliding with a structural weakness most employers don't see coming. I'm talking about multi-cancer early detection tests (MCEDs like Grail's Galleri) and continuous biosurveillance wearables.
The usual objections are "too expensive" or "not yet FDA approved." That second one is already changing: Grail filed its FDA premarket approval application for Galleri in January 2026, and an FDA advisory committee is scheduled to review it on September 23, 2026. Approval will not fix the deeper problem, which sits in your benefits administration system, your TPA's claims engine, your stop-loss contract, and your plan document.
Your system is built to pay for claims (events). These tests generate signals (probabilities). That's the impedance mismatch creating a fiduciary and administrative crisis no wellness vendor can fix alone.
1. The Coverage Gap: Where MCED Claims Disappear
Standard claims adjudication follows a binary logic tree:
- Preventive: Free to the member (USPSTF A/B rated).
- Diagnostic: Subject to deductible/coinsurance (requires symptoms).
- Treatment: High cost, medical necessity review.
An MCED test falls outside both buckets. The preventive bucket requires a USPSTF A or B rating, and MCED screening does not yet have one. The diagnostic bucket requires symptoms, and most plan documents define medical necessity as "signs or symptoms." An asymptomatic patient fails that test. Medicare gained authority to cover FDA-approved MCED tests in February 2026, but that law does not reach employer plans.
When the lab sends a claim for a $949 Galleri test, your TPA's engine has no bucket for it. Two outcomes are likely:
- Auto-deny: "Not a covered preventive service, no symptom to trigger diagnostic."
- Default to member liability: The employee gets a surprise bill for the "wellness test" your vendor encouraged them to take.
The fiduciary risk: If your wellness program promoted the test and your TPA denied it, the employee has a credible ERISA claim against you, the plan fiduciary. You created a system where no one owns responsibility for processing a new class of clinical input.
2. The Stop-Loss Trap: A Million-Dollar Game of "Date of Occurrence"
Self-funded employers often think: "If we pay $1,000 to find one late-stage cancer early, we save hundreds of thousands in treatment. Easy ROI."
Clinically, you're right. Financially, you may be wrong, because of how your stop-loss contract interprets the data.
Consider this scenario:
- Your plan covers an MCED test (or an employee pays out-of-pocket).
- The test returns "Signal Detected" for pancreatic cancer.
- The employee undergoes a six-month diagnostic odyssey to locate the tiny tumor.
- At month seven, cancer is confirmed. Treatment begins. The claim hits $1M, your stop-loss attachment point.
- The stop-loss carrier audits the claim and sees the signal date in the lab data feed.
The trap: Occurrence language and renewal rules can turn a screening signal into a pre-existing condition. A carrier can argue the occurrence date is when the signal appeared, not when the diagnosis was pathologically confirmed. At renewal, a positive MCED result is a documented reason to laser that employee with a higher deductible or a full exclusion.
If your policy also has an "incurred expense" clause requiring treatment to start within the policy year, you've just self-funded a million-dollar claim. Your system processed the test as a wellness screen; the carrier reclassified it as a pre-existing trigger.
Million-dollar claims are not a hypothetical edge case. Sun Life's analysis shows claims exceeding $1 million rose 29% between 2024 and 2025.
You're marketing an early detection tool clinically, but your risk transfer system is designed to reject it retroactively.
3. Continuous Monitoring vs. Episodic Care: The Design Gap
Your entire benefits system is built for episodic care:
- Enrollment event → join the plan.
- Occurrence event → get sick.
- Claim event → get treatment.
Early detection changes the paradigm to continuous monitoring.
Think about a wearable biosensor that detects atrial fibrillation before a stroke happens. The wearable generates a clinical signal. The biometric screening vendor captures it. The wellness platform records "completion." But the claims system has no trigger, because the employee isn't "sick" yet.
If the employee ignores the signal because the plan never assigns a care coordinator or a path of action, and they have a stroke two months later, your plan has failed.
The system has no concept of a "pre-morbid claimant."
- FMLA/STD: You can't trigger intermittent leave for "investigating a biomarker signal."
- Pharmacy: No formulary pathway for "prophylactic treatment of a high-probability liquid biopsy result."
- Network: No "Signal Navigator" benefit code.
You're paying for the signal but have no administrative pathway to act on it.
4. False Positives and False Negatives: The Rest of the Signal
A positive MCED signal is frequently a false alarm, and a negative result is not a guarantee. The full results of the NHS-Galleri trial, presented in May 2026, put numbers on both.
The trial reported specificity of 99.55%, which still produced a false positive rate of 0.45%, and a positive predictive value of 52%. Roughly half of "Signal Detected" results were not cancer. Each one still triggers a diagnostic workup: PET scans, endoscopies, biopsies, specialist visits. The plan has to fund and coordinate that cascade even when there is no tumor, and its budget has to account for false positives alongside the true ones.
On the negative side, episode sensitivity was 54.7% for the 12 prespecified cancer types and 30.7% across all cancers. Across all cancers, the test misses more than it catches. A clean result cannot replace the USPSTF-recommended screening employees are already supposed to get. A plan document that funds the signal but lets a negative result quietly cancel standard screening is setting up a different failure.
The fix is to write the full range of signal outcomes into the plan document and stop-loss contract, not just the true-positive path.
How to Fix the System: The Actionable Playbook
This is a benefits administration system design problem. Three fixes follow.
1. The Plan Document Amendment (The "Signal Rider")
Write explicit language into your self-funded plan document defining "Covered Early Detection Signals."
- State that a positive MCED result triggers Diagnostic Acceleration Coverage: 100% coverage of all necessary PET scans, endoscopies, and biopsies for 90 days, regardless of symptoms.
- Contractually override your TPA's default medical necessity logic for this specific use case.
2. The Stop-Loss Reset
Get a Letter of Agreement from your stop-loss carrier stating explicitly:
- The "occurrence date" for a late-stage cancer is the date of pathological confirmation, not the date of the screening signal.
- The MCED signal does not constitute a "known condition" for the purposes of a new policy year or a specific employee exclusion (lasering).
3. The Data Triad Architecture
Your HR and benefits tech stack must move from a duad to a triad:
- Duad: Enrollment data ↔ Claims data.
- Triad: Enrollment data ↔ Signal data (from wellness and lab) ↔ Claims data.
This architecture lets the system flag a high-probability signal and automatically trigger a care navigation workflow before the claim arrives. It prevents the million-dollar stroke and the ERISA lawsuit simultaneously.
What Changes When the System Reads Signals
The wellness industry is selling you a Ferrari: MCEDs, wearables, biosurveillance. Your benefits administration system is a horse-drawn carriage: claims logic, plan documents, stop-loss contracts.
Until you fix the impedance mismatch between the signal and the claim, you are not offering early detection. You are offering a sophisticated trap for your employees, your budget, and your fiduciary duty.
Better science alone won't carry the early detection revolution. It needs benefits systems that know how to handle a probability before it becomes a loss.