This case does not predict whether PwC's projected 2027 medical cost trend actually materializes, whether AI-driven claims-denial litigation reaches a ruling, or whether employer-sponsored coverage keeps eroding at its historical pace. It scoreboards four independent, dated tracks instead. MACPAC voted May 7, 2026 to recommend that HHS direct CMS to require human expert review of AI-driven adverse medical-necessity determinations in Medicaid — a recommendation to Congress and CMS, not a binding rule, with no confirmed federal action yet.[1] Two separate lawsuits alleging improper AI-driven claims denials are in active discovery: Kisting-Leung v. Cigna, where class claims survived a partial 2025 dismissal ruling and scheduling orders continue through mid-2026, and Estate of Lokken v. UnitedHealth, where a March 2026 magistrate order compelled broad discovery into the insurer's nH Predict tool.[2][3] PwC's Health Research Institute projects 2027 medical cost trend at 9.0% (group) and 8.5% (individual) — the steepest in the report's 17-year history — a projection, not yet realized data.[4] And the employer-coverage share that has fallen from roughly 67% to 60% since 1998 has no confirmed next data point showing whether that decline accelerated, held steady, or reversed.[5] None of these four threads depends on the others. The honest position is a scoreboard, not a guess.
Four clocks are running on the same underlying question — who actually ends up absorbing the rising cost of care this cluster's other cases document — and the temptation at the end of it is to guess which resolves first: a regulatory recommendation, a court ruling, a cost-trend projection, or a decades-long coverage pattern. This case refuses that temptation for the same reason the rest of the cluster does: each thread moves on an independent clock, and naming all four precisely is more honest than picking a favorite.
The regulatory track has the least certain timeline. MACPAC's May 7, 2026 vote recommends HHS direct CMS to clarify that federal Medicaid regulations require human expert review of adverse medical-necessity determinations, and that automated tools alone cannot make those determinations.[1] It's a recommendation to Congress and CMS in MACPAC's June 2026 Report to Congress — not a rule with an effective date, and not binding until and unless HHS or CMS acts on it. Whether, and when, that happens is genuinely open.
The litigation track is the most concrete and furthest along. Kisting-Leung v. Cigna, alleging improper use of the PXDX system to deny claims, survived a partial motion-to-dismiss ruling in March 2025 with class claims intact, and scheduling orders through 2026 show active pretrial discovery — no class certification decision yet.[2] Estate of Lokken v. UnitedHealth, over the nH Predict tool, saw a March 9, 2026 magistrate order compel broad discovery into the tool's governance documents — a real, recent escalation, not a stalled case.[3] Neither has reached a ruling on the merits.
The two data tracks carry no single dramatic moment but will supply the clearest ongoing evidence. PwC's 9.0%/8.5% trend projection for 2027 is the steepest in the report's 17-year history — whether actual 2027 claims experience confirms, undershoots, or exceeds it won't be knowable until well into 2027 itself.[4] And the employer-coverage share that fell from roughly 67% to 60% since 1998 has no confirmed next data point yet showing whether STAT's documented small-business dropout pattern is accelerating the decline or remains a smaller, anecdotally-visible piece of a slower-moving trend.[5]
None of the four has resolved as of this writing. None carries a fixed near-term date. The honest answer is the scoreboard, not a prediction.[1][2][3][4][5]
The four independent clocks this cluster is watching, and their status as of July 2026.
MACPAC voted May 7, 2026 to recommend HHS/CMS require human review of AI-driven Medicaid denials. No confirmed federal action has followed as of this writing.[1]
Not FiredPwC's 9.0%/8.5% 2027 medical cost trend projection, the steepest in 17 years, won't be confirmable against actual claims data until well into 2027.[4]
PendingWhether the documented small-business coverage dropout pattern shows up in official Peterson-KFF or Census data remains unconfirmed as of this writing.[5]
Not FiredChosen to land after year-end 2026 insurer earnings, giving both litigation tracks and the regulatory-response track room to move, and edging toward the first real 2027 trend data. Review then: has any of the four triggers fired?
ReviewAutomation tools alone may not make adverse determinations. — MACPAC, Report to Congress recommendation, June 2026
| Dimension | Evidence |
|---|---|
| Revenue (D2) Origin · 82 | The unresolved question beneath all four tracks is the same: who ultimately absorbs the cost of care as trends rise, denials get automated, and coverage structures shift.[1][2][3][4][5] D2 is the origin because a regulatory recommendation, two lawsuits, and two data tracks are five different angles on one cost-absorption question.Who Absorbs the Cost |
| Regulatory (D4) L1 · 78 | MACPAC's recommendation and the two active AI-denial lawsuits are the most directly accountability-focused, differently-paced tracks in this capstone.[1][2][3] D4 amplifies from D2 as the most institutionally active dimension.Two Accountability Tracks |
| Customer (D1) L1 · 74 | Members and employers are the parties directly experiencing whichever combination of cost trend, coverage decisions, and claims-review outcomes actually materializes.[4][5] D1 amplifies alongside D4 as the population bearing the outcome. |
| Operational (D6) L2 · 62 | Whether insurers continue the operational responses documented in this cluster — automation, membership discipline, market exits — depends partly on how the regulatory and litigation tracks resolve.[1][2][3] D6 sits here as the operational dimension shaped by the other tracks. |
| Quality (D5) L2 · 56 | Whether PwC's projected trend becomes realized fact, and whether the coverage-erosion pattern accelerates as documented, are both projection-versus-reality questions this capstone can't resolve yet.[4][5] D5 sits here as that distinction. |
| Employee (D3) 32 | Deliberately the thinnest dimension. This capstone synthesizes cost, regulatory, and coverage-structure questions; no comparable workforce-level finding exists across either companion case. |
The cascade originates in D2 — Revenue — because the unresolved question underneath all four tracks is the same: who ultimately absorbs the cost of care as trends rise, denials get automated, and coverage structures shift.[1][2][3][4][5] From D2 it runs to D4 (the regulatory and litigation tracks most directly testing accountability for that cost) and D1 (members and employers whose coverage and claims experience the outcome directly). It then reaches D6 (whether insurers' operational responses — automation, membership discipline, market exits — continue or change) and D5 (whether a projection becomes realized fact), with D3 kept thin — a healthcare-economics and regulatory cascade, not a workforce one. This is the cluster capstone: it synthesizes [UC-281]'s profit-cost divergence, [UC-282]'s structural coverage erosion, and [UC-283]'s counterexample into one forward scoreboard. Confidence is deliberately low (0.43): four independent, genuinely unpredictable tracks compound into real uncertainty, and displaying false confidence here would betray the discipline the whole cluster runs on.
-- UC-284: The Trend Nobody's Paying Yet: 6D Prognostic Capstone
-- Four independent unresolved tracks on who absorbs rising health costs (synthesizes UC-281/282/283)
FORAGE trend_nobody_paying_yet
WHERE verdict_held_open = true
AND four_tracks_independently_unresolved = true
AND no_track_depends_on_another = true
ACROSS D2, D4, D1, D6, D5, D3
DEPTH 3
SURFACE trend_nobody_paying_yet
WATCH macpac_federal_action WHEN hhs_or_cms_acts_on_ai_review_recommendation = true
WATCH ai_denial_litigation WHEN cigna_or_unitedhealth_case_reaches_ruling_or_class_cert = true
WATCH cost_trend_realized WHEN 2027_actual_medical_trend_data_confirms_or_breaks_pwc_projection = true
WATCH coverage_erosion_continues WHEN next_peterson_kff_data_point_confirms_dropout_pattern = true
DRIFT trend_nobody_paying_yet
METHODOLOGY 85
PERFORMANCE 40
FETCH trend_nobody_paying_yet
THRESHOLD 1000
ON WATCH CHIRP medium 'Four independent unresolved tracks: MACPAC voted May 7 2026 recommending HHS/CMS require human review of AI Medicaid denials, no federal action yet. Cigna PXDX (Kisting-Leung) and UnitedHealth nH Predict (Lokken) lawsuits both in active discovery, no ruling on merits. PwC projects 2027 medical trend 9.0pct/8.5pct, steepest in 17yrs, not yet realized. Employer coverage share 67pct (1998) to 60pct (recent), next data point unconfirmed. None resolved as of Jul 2026'
SURFACE review ON '2027-02-15'
SURFACE analysis AS jsonRuntime: @stratiqx/cal-runtime · Spec: cal.semanticintent.dev · DOI: 10.5281/zenodo.18905193
A federal regulatory recommendation, two lawsuits, an actuarial projection, and a coverage-erosion trend share a common subject but no common cause. Any one could resolve without the others moving at all.[1][2][3][4][5]
Both AI-denial lawsuits have real, recent, verifiable escalations — but discovery orders aren't rulings, and neither case has a predictable date for reaching one.[2][3]
MACPAC voting to recommend federal action is a real, dated event. Whether HHS or CMS actually acts on it is a separate, unscheduled question this capstone holds open rather than assumes.[1]
A capstone that guessed which track resolves first would be pretending to knowledge nobody currently has. Confidence 0.43 and a February 2027 review date are the honest alternative to that guess.
Five sources, each anchoring one of the capstone's tracks: MACPAC's own recommendation, the two active AI-claims-denial lawsuits' court dockets, PwC's published cost-trend projection, and Peterson-KFF's tracked employer-coverage data.
Watch all four. When one resolves, the question sharpens. Until then, the scoreboard is the honest answer.