vida: research session 2026-04-13 — 10 sources archived

Pentagon-Agent: Vida <HEADLESS>
2026-04-13 04:16:33 +00:00 · 2026-04-13 04:16:33 +00:00 · 54f37e36ee
commit 54f37e36ee
parent e24f006773
12 changed files with 870 additions and 0 deletions
--- a/agents/vida/musings/research-2026-04-13.md
+++ b/agents/vida/musings/research-2026-04-13.md
@ -0,0 +1,189 @@
 ---
 type: musing
 domain: health
 session: 23
 date: 2026-04-13
 status: active
 ---
 # Research Session 23 — USPSTF GLP-1 Gap + Behavioral Adherence: Breaking the Continuous-Delivery Assumption?
 ## Research Question
 What is the current USPSTF status on GLP-1 pharmacotherapy recommendations, and are behavioral adherence programs closing the gap that coverage alone can't fill — particularly for the 85.7% of commercially insured GLP-1 users who don't achieve durable metabolic benefit?
 **Why this question now:**
 Session 22 identified two active threads:
 1. The USPSTF GLP-1 pathway — potentially the most significant future offset to the access collapse (a new B recommendation would mandate ACA coverage without cost-sharing)
 2. The adherence complication: 14.3% two-year persistence even with commercial coverage means the problem isn't only financial access. Direction A was "what behavioral support programs improve adherence?"
 Session 22 also flagged "continuous-treatment model claim: READY TO EXTRACT" — but this session found evidence that complicates that extraction. The Omada post-discontinuation data is the most significant finding.
 **Note:** Tweet file was empty this session — no curated sources. All research is from original web searches.
 ## Belief Targeted for Disconfirmation
 **Primary target — Belief 1: Healthspan is civilization's binding constraint, and we are systematically failing at it in ways that compound.**
 **Specific falsification criterion:**
 If behavioral wraparound programs are demonstrably closing the adherence gap (85.7% non-adherent despite coverage), then the "continuous delivery required" thesis may overstate the pharmacological dependency. The Omada post-discontinuation claim — if real — would mean behavioral infrastructure CAN break GLP-1 dependency, converting a continuous-delivery requirement into a skill-buildable state. This would: (1) weaken the compounding failure thesis (one layer is addressable without the medication being continuous); (2) change the policy prescription (fund behavioral wraparound, not just medication access).
 **USPSTF disconfirmation criterion:**
 If USPSTF has a pending draft recommendation that would extend the B rating to GLP-1 pharmacotherapy, that would be an operational policy offset in development — challenging the "no offset mechanism" conclusion from Session 22.
 **What I expected to find:** Programs show associative improvements but with survivorship bias; no prospective RCTs of behavioral wraparound; USPSTF has no pending GLP-1 update.
 ## What I Searched For
 - USPSTF weight loss interventions draft recommendation 2026 pharmacotherapy GLP-1
 - USPSTF formal petition for GLP-1 pharmacotherapy inclusion
 - GLP-1 behavioral adherence support programs 2025-2026 (Noom, Calibrate, Omada, WW Med+, Ro Body)
 - GLP-1 access equity by state/income (the "access inversion" framing)
 - Racial/ethnic disparities in GLP-1 prescribing
 - Medical school prospective pre-AI clinical competency baselines (never-skilling detection)
 - New clinical AI deskilling evidence 2025-2026 beyond the colonoscopy ADR study
 ## Key Findings
 ### 1. DISCONFIRMATION TEST RESULT — USPSTF: No Offset in Development
 **The disconfirmation question:** Is USPSTF developing a GLP-1 pharmacotherapy recommendation that would mandate ACA coverage?
 **Answer: No — the 2018 B recommendation remains operative, with no petition or draft update for GLP-1 pharmacotherapy visible.**
 Key facts:
 - USPSTF 2018 B recommendation: intensive multicomponent behavioral interventions for BMI ≥30. Pharmacotherapy was reviewed but NOT recommended (lacked maintenance data). Medications reviewed: orlistat, liraglutide, phentermine-topiramate, naltrexone-bupropion, lorcaserin — Wegovy/semaglutide 2.4mg and tirzepatide are ABSENT.
 - USPSTF website flags adult obesity topic as "being updated" but redirect points toward cardiovascular prevention, not GLP-1 pharmacotherapy.
 - No formal USPSTF petition for GLP-1 pharmacotherapy found in any search.
 - No draft recommendation statement visible as of April 2026.
 - Policy implication: A new A/B rating covering pharmacotherapy would trigger ACA Section 2713 mandatory coverage without cost-sharing for all non-grandfathered plans. This is the most significant potential policy mechanism — and it doesn't exist yet.
 **Conclusion:** The USPSTF gap is growing in urgency as therapeutic-dose GLP-1s become standard of care. The 2018 recommendation is 8 years behind the science. No petition or update is in motion. This is an extractable claim: the policy mechanism that would most effectively address GLP-1 access doesn't exist and isn't being created.
 ### 2. MOST SURPRISING FINDING — Omada Post-Discontinuation Data Challenges the Continuous-Delivery Thesis
 **This is the session's most significant finding for belief revision.**
 Session 22 was about to flag "continuous-treatment model claim: READY TO EXTRACT" — stating that pharmacological/dietary interventions require continuous delivery for sustained effect (GLP-1 rebound, food-as-medicine reversion, antidepressant relapse pattern all confirmed this).
 Omada Health's Enhanced GLP-1 Care Track data challenges this:
 - 63% of Omada members MAINTAINED OR CONTINUED LOSING WEIGHT 12 months after stopping GLP-1s
 - Average weight change post-discontinuation: 0.8% (near-zero)
 - This is the strongest post-discontinuation data of any program found
 **Methodological caveats that limit this finding:**
 - Survivorship bias: sample includes only patients who remained in the Omada program after stopping GLP-1s — not all patients who stop GLP-1s
 - Omada-specific: the behavioral wraparound (high-touch care team, nutrition guidance, exercise specialist, muscle preservation) is more intensive than standard care
 - Internal analysis (not peer-reviewed RCT)
 **What this means if it holds:**
 The "continuous delivery required" thesis may be over-general. The more precise claim is: GLP-1s without behavioral infrastructure require continuous delivery; GLP-1s WITH comprehensive behavioral wraparound may produce durable changes in some patients even after cessation. This is a scope qualification, not a disconfirmation — but it's important.
 **Hold the "continuous-treatment model claim" extraction.** The Omada finding needs to be archived and weighed alongside the GLP-1 rebound data. The extraction should include both the rebound evidence (the rule) and the Omada data (the potential exception with behavioral wraparound). This changes the claim title from absolute to conditional.
 ### 3. Behavioral Adherence Programs Show Consistent Signal (With Caveats)
 **All programs report better persistence and weight loss with behavioral engagement:**
 Noom (January 2026 internal analysis, n=30,239):
 - Top engagement quartile: 2.2x longer persistence vs. bottom quartile (6.2 months vs. 2.8 months)
 - 25.2% more weight loss at week 40
 - Day-30 retention: 40% (claimed 10x industry average)
 - Reverse causality caveat: people doing well may engage more — not proven that engagement causes persistence
 Calibrate (n=17,475):
 - 15.7% average weight loss at 12 months; 17.9% at 24 months (sustained, not plateau)
 - Interrupted access: 13.7% at 12 months vs 17% uninterrupted — behavioral program provides a floor
 - 80% track weight weekly; 67% complete coaching sessions
 WeightWatchers Med+ (March 2026, n=3,260):
 - 61.3% more weight loss in month 1 vs. medication alone
 - 21.0% average weight loss at 12 months; 20.5% at 24 months
 - 72% reported program helped minimize side effects
 Omada (n=1,124):
 - 94% persistence at 12 weeks (vs. 42-80% industry range)
 - 84% persistence at 24 weeks (vs. 33-74% industry range)
 - 18.4% weight loss at 12 months (vs. 11.9% real-world comparators)
 - Post-discontinuation: 63% maintained/continued weight loss; 0.8% average change
 **Cross-cutting caveat:** Every program's data is company-sponsored, observational, with survivorship bias. No independent RCT of behavioral wraparound vs. medication-only with long-term primary endpoints. The signal is consistent but not proven causal.
 **Industry-level improvement:** One-year persistence for Wegovy/Zepbound improved from 40% (2023) to 63% (early 2024) — nearly doubling. This could reflect: (1) increasing availability of behavioral programs; (2) improved patient selection; (3) dose titration improvements reducing GI side effects.
 ### 4. GLP-1 Access Inversion — Now Empirically Documented
 The access inversion framing is confirmed with new data:
 Geographic/income pattern:
 - Mississippi, West Virginia, Louisiana (obesity rates 40%+) → low income states, minimal Medicaid GLP-1 coverage, 12-13% of median annual income to pay out-of-pocket for GLP-1
 - Massachusetts, Connecticut → high income states, 8% of median income for out-of-pocket
 Racial disparities — Wasden 2026 (*Obesity* journal, large tertiary care center):
 - Before MassHealth Medicaid coverage change (January 2024): Black patients 49% less likely, Hispanic patients 47% less likely to be prescribed semaglutide/tirzepatide vs. White patients
 - After coverage change: disparities narrowed substantially
 - Conclusion: insurance policy is primary driver, not just provider bias
 - Separate tirzepatide dataset: adjusted ORs vs. White — AIAN: 0.6, Asian: 0.3, Black: 0.7, Hispanic: 0.4, NHPI: 0.4
 Wealth-based treatment timing:
 - Black patients with net worth >$1M: median BMI 35.0 at GLP-1 initiation
 - Black patients with net worth <$10K: median BMI 39.4 — treatment starts 13% later in disease progression
 - Lower-income patients are sicker when they finally get access
 **This is extractable.** The access inversion claim has now been confirmed with three independent evidence types: geographic/income data, racial disparity data, and treatment-timing data. This is ready to extract as a claim: "GLP-1 access follows an access inversion pattern — highest-burden populations by disease prevalence are precisely the populations with least access by coverage and income."
 ### 5. Clinical AI Deskilling — Now Cross-Specialty Evidence Body (2025-2026)
 Session 22 had the colonoscopy ADR drop (28% → 22%) as the anchor quantitative finding. This session found 4 additional quantitative findings:
 New evidence:
 - Mammography/breast imaging: erroneous AI prompts increased false-positive recalls by up to 12% among 27 experienced radiologists (automation bias mechanism)
 - Computational pathology: 30%+ of participants reversed correct initial diagnoses when exposed to incorrect AI suggestions under time constraints (mis-skilling in real time)
 - ACL diagnosis: 45.5% of clinician errors resulted directly from following incorrect AI recommendations
 - UK GP medication management: 22.5% of prescriptions changed in response to decision support; 5.2% switched from correct to incorrect prescription after flawed advice (measurable harm rate)
 Comprehensive synthesis:
 - Natali et al. 2025 (*Artificial Intelligence Review*, Springer): mixed-method review across radiology, neurosurgery, anesthesiology, oncology, cardiology, pathology, fertility medicine, geriatrics, psychiatry, ophthalmology. Cross-specialty pattern confirmed: AI benefits performance while present; produces skill dependency visible when AI is unavailable.
 - Frontiers in Medicine 2026: neurological mechanism proposed — reduced prefrontal cortex engagement, hippocampal disengagement from memory formation, dopaminergic reinforcement of AI-reliance. Theoretical but mechanistically grounded.
 **Belief 5 status:** Significantly strengthened. The evidence base for AI-induced deskilling has moved from "one study + theoretical concern" to "5 independent quantitative findings across 5 specialties + comprehensive cross-specialty synthesis + proposed neurological mechanism." This is no longer a hypothesis.
 ### 6. Never-Skilling — Formally Named, Not Yet Empirically Proven
 The "never-skilling" concept has moved from informal framing to peer-reviewed literature:
 - NEJM (2025-2026): explicitly discusses never-skilling as distinct from deskilling
 - JEO (March 2026): "Never-skilling poses a greater long-term threat to medical education than deskilling"
 - NYU's Burk-Rafel: institutional voice using the term explicitly
 - Lancet Digital Health (2025): addresses productive struggle removal
 What still doesn't exist: any prospective study comparing AI-naive vs. AI-exposed-from-training cohorts on downstream clinical performance. No medical school has a pre-AI baseline competency assessment designed to detect never-skilling. The gap is confirmed — absence is the finding.
 ## Follow-up Directions
 ### Active Threads (continue next session)
 - **"Continuous-treatment model" claim: HOLD FOR REVISION.** Omada post-discontinuation data must be weighed. Extract the claim with explicit scope: "WITHOUT behavioral infrastructure, pharmacological/dietary interventions require continuous delivery. WITH comprehensive behavioral wraparound, some patients maintain durable effect post-discontinuation." Needs: (1) wait for Omada data to appear in peer-reviewed form; or (2) extract with explicit caveat that Omada data is internal/observational and creates a divergence. Check for Omada peer-reviewed publication of post-discontinuation data.
 - **GLP-1 access inversion claim: READY TO EXTRACT.** Three independent evidence types now converge. Draft: "GLP-1 access follows systematic inversion — the populations with highest obesity prevalence and disease burden have lowest access by coverage, income, and treatment-initiation timing." Primary evidence: KFF state coverage data, Wasden 2026 racial disparity study, geographic income analysis.
 - **USPSTF gap claim: READY TO EXTRACT.** "USPSTF's 2018 obesity B recommendation predates therapeutic-dose GLP-1s and has not been updated or petitioned, leaving the most powerful ACA coverage mandate mechanism dormant for the drug class most likely to change obesity outcomes." This is a specific, falsifiable claim — USPSTF is the institutional gap that no other mechanism compensates for.
 - **Clinical AI deskilling — divergence file update.** The body of evidence has grown from 1 to 5+ quantitative findings across 5 specialties. Session 22 archives covered colonoscopy ADR. This session's Natali et al. review is the synthesis. Consider: should the existing claim file be enriched with new evidence, or is this now ready for a divergence file between "AI deskilling is documented across specialties" and "AI up-skilling (performance improvements while AI is present)"? The Natali review makes this a genuine divergence — AI improves performance while present AND reduces performance when absent.
 - **Omada post-discontinuation: peer-reviewed publication search.** Internal company analysis is insufficient for extraction. Search for: "Omada Health GLP-1 post-discontinuation peer reviewed 2025 2026" and "behavioral support GLP-1 cessation weight maintenance RCT." If no peer-reviewed version exists, archive the finding with confidence: speculative and note what would resolve it.
 ### Dead Ends (don't re-run these)
 - **USPSTF GLP-1 pharmacotherapy petition:** No petition, no draft, no formal nomination process visible. Don't re-search until a specific trigger event (USPSTF announcement, advocacy organization petition filed). Note: USPSTF's adult obesity topic is flagged as "under revision" but redirect is cardiovascular prevention, not pharmacotherapy.
 - **Omada peer-reviewed post-discontinuation study:** Not yet published in peer-reviewed form (confirmed via search). Don't search again until Q4 2026 — that's the likely publication window if the data was presented at ObesityWeek 2025.
 - **Company-sponsored behavioral adherence RCTs:** None of the major commercial programs (Noom, Calibrate, WW Med+, Ro, Omada) have published independent RCT-level evidence for behavioral wraparound improving long-term persistence as of April 2026. The gap is real and confirmed. Don't search for this again — it doesn't exist yet.
 ### Branching Points (one finding opened multiple directions)
 - **Omada post-discontinuation finding:** Direction A — immediately refine and conditionally extract the continuous-treatment model claim with explicit scope qualification; Direction B — treat Omada data as a divergence candidate (behavioral wraparound may enable durable effect post-cessation vs. general GLP-1 rebound pattern). Direction A is more conservative and appropriate given the methodological caveats. Pursue Direction A next session after archiving the Omada finding for extractor review.
 - **Racial disparities in GLP-1 access:** Direction A — extract the Wasden 2026 finding as a standalone claim (racial disparities in GLP-1 prescribing narrow significantly with Medicaid coverage expansion → insurance policy, not provider bias, is primary driver); Direction B — combine with access inversion framing into a single compound claim. Direction A preserves specificity — the Wasden finding is clean enough to stand alone.
 - **Clinical AI deskilling body of evidence:** Direction A — enrich existing deskilling claim file with the 5 new quantitative findings and the Natali 2025 synthesis; Direction B — create a divergence file between "AI deskilling" and "AI up-skilling while present." Direction B captures the more interesting structural tension — AI simultaneously improves performance (while present) and damages performance (when absent). This is not a contradiction; it's the dependency mechanism. But it looks like a divergence from the outside.
--- a/agents/vida/research-journal.md
+++ b/agents/vida/research-journal.md
@ -1,5 +1,31 @@
 # Vida Research Journal
 ## Session 2026-04-13 — USPSTF GLP-1 Gap + Behavioral Adherence: Continuous-Delivery Thesis Complicated
 **Question:** What is the current USPSTF status on GLP-1 pharmacotherapy recommendations, and are behavioral adherence programs closing the gap that coverage alone can't fill — particularly for the 85.7% of commercially insured GLP-1 users who don't achieve durable metabolic benefit?
 **Belief targeted:** Belief 1 (healthspan as civilization's binding constraint; compounding failure thesis). Specific disconfirmation target: if USPSTF has a pending GLP-1 pharmacotherapy recommendation, that's the most powerful offsetting mechanism available. Secondary target: if behavioral wraparound programs can break the GLP-1 continuous-delivery dependency, the pharmacological failure layer is addressable without continuous access.
 **Disconfirmation result:** MIXED — two distinct findings with different valences:
 (1) USPSTF gap: NOT DISCONFIRMED. The 2018 B recommendation predates therapeutic-dose GLP-1s (Wegovy/tirzepatide absent from the evidence base). No draft update, no formal petition, no timeline for inclusion of pharmacotherapy. The most powerful ACA coverage mandate mechanism is dormant. This strengthens the "no operational offset" finding from Session 22.
 (2) Behavioral wraparound: PARTIAL COMPLICATION. Omada's post-discontinuation data (63% maintained/continued weight loss 12 months after stopping GLP-1s; 0.8% average weight change) challenges the categorical continuous-delivery framing developed in Sessions 20-22. Calibrate's interrupted access data (13.7% weight loss maintained at 12 months despite interruptions) provides a second independent signal. Both are observational and survivorship-biased. But the signal is consistent across both programs. The "continuous delivery required" claim needs scope qualification: without behavioral infrastructure → yes; with comprehensive behavioral wraparound → uncertain, possibly different.
 **Key finding:** Omada post-discontinuation data is the session's most significant finding. 63% of former GLP-1 users maintaining or continuing weight loss 12 months post-cessation with only 0.8% average weight change directly challenges the prevailing assumption of universal rebound. Sessions 20-22 were about to extract a "continuous delivery required" claim — this session's finding demands a hold on that extraction pending scope qualification. The continuous-delivery rule may be a conditional rule: true without behavioral infrastructure; potentially false with comprehensive behavioral wraparound.
 Secondary key finding: Racial disparities in GLP-1 prescribing (49% lower for Black, 47% lower for Hispanic patients pre-coverage) nearly fully close with Medicaid coverage expansion — identifying insurance policy, not provider bias, as the primary driver. This is methodologically clean (natural experiment) and extractable.
 USPSTF gap is the most actionable new finding: the policy mechanism that would mandate GLP-1 coverage under ACA is dormant and apparently no one has filed a petition to activate it.
 **Pattern update:** The compounding failure pattern is now complete (Sessions 1-22), but Session 23 introduces a complication: the behavioral wraparound data suggests one layer of the failure (the continuous-delivery layer) may be addressable without solving the access problem — if the delivery infrastructure includes behavioral support. This doesn't change the access failure finding, but it does change the policy prescription: covering medication access alone may be less effective than coverage + behavioral wraparound mandates. The Wasden 2026 finding strengthens the structural policy argument: coverage expansion directly reduces racial disparities, which directly serves the access inversion pattern.
 **Confidence shift:**
 - Belief 1 ("systematically failing in compounding ways"): **UNCHANGED BUT NUANCED** — the compounding failure is confirmed at the access layer (USPSTF dormant, state cuts accelerating). However, the behavioral wraparound data introduces a partial offset mechanism that wasn't visible in Sessions 20-22. The "compounding" remains true for the access infrastructure; but the "unaddressable without continuous medication" claim may be overstated. Belief 1 holds, but the implications for intervention design have shifted.
 - Belief 5 (clinical AI novel safety risks): **STRENGTHENED** — deskilling evidence base expanded from 1 (colonoscopy) to 5 quantitative findings across 5 specialties. Natali et al. 2025 provides the cross-specialty synthesis. Never-skilling concept is now formally named in NEJM, JEO, and Lancet Digital Health. This is no longer preliminary.
 ---
 ## Session 2026-04-12 — GLP-1 Access Infrastructure: Compounding Failure Confirmed, No Operational Offset
 **Question:** Is the compounding failure in GLP-1 access infrastructure (state coverage cuts + SNAP cuts + continuous-delivery requirement) being offset by federal programs (BALANCE model, Medicare Bridge), or is the "systematic compounding failure" thesis confirmed with no effective counterweight?
--- a/inbox/queue/2026-04-13-calibrate-glp1-behavioral-interrupted-access-outcomes.md
+++ b/inbox/queue/2026-04-13-calibrate-glp1-behavioral-interrupted-access-outcomes.md
@ -0,0 +1,70 @@
 ---
 type: source
 title: "Calibrate GLP-1 + Behavioral Coaching: Interrupted Access Data Shows Behavioral Floor Effect (Endocrine Society 2025)"
 author: "Calibrate (Endocrine Society presentation, 2025)"
 url: https://www.joincalibrate.com
 date: 2025-01-01
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: medium
 tags: [glp1, behavioral-wraparound, adherence, interrupted-access, weight-loss, calibrate]
 ---
 ## Content
 Calibrate employer-sponsored program data (n=17,475 members; interrupted access analysis from Endocrine Society 2025 presentation):
 **Primary outcomes (uninterrupted access):**
 - 12-month weight loss: 15.7% average
 - 18-month: 17.3%
 - 24-month: 17.9% — continued loss, not plateau
 - Waist circumference reduction: avg 6 inches at 12 months
 - Engagement metrics: 80% track weight weekly; 67% complete coaching sessions
 **Interrupted access data (Endocrine Society 2025):**
 - Members with treatment interruptions: 13.7% weight loss at 12 months vs. 17% uninterrupted
 - At 24 months: 14.9% vs. 20.1% for uninterrupted
 - Delta: ~2.2 percentage points at 12 months; ~5.2 percentage points at 24 months
 **Interpretation of interrupted access data:**
 Even when GLP-1 access is interrupted, Calibrate members maintained 13.7-14.9% weight loss. In context:
 - Standard GLP-1 cessation data (STEP 4 trial): patients regained ~2/3 of lost weight within 1 year of stopping — typically implying substantial regain
 - Calibrate interrupted access: maintaining 13.7% at 12 months suggests the behavioral program provides a significant floor below which weight does not revert
 - The behavioral program appears to prevent full rebound even when medication is unavailable
 **Calibrate program components (1-year employer-sponsored):**
 - GLP-1 prescriptions
 - Coaching on food, sleep, exercise, emotional health (four pillars)
 - Regular check-ins and goal tracking
 **Methodological notes:**
 - n=17,475 is a substantial sample
 - "Treatment interruptions" is company-defined — criteria not specified in available data
 - Endocrine Society presentation (not yet peer-reviewed as standalone paper)
 - Financial conflict: Calibrate is presenting its own program data
 ## Agent Notes
 **Why this matters:** The interrupted access data is the most mechanistically interesting finding from Calibrate. If the behavioral floor holds even when GLP-1 is interrupted — preventing the typical ~2/3 weight regain — this is more compelling evidence than the WW and Noom persistence data. It's suggesting behavioral change actually happened, not just medication effect.
 **What surprised me:** 13.7% weight loss at 12 months for members with treatment interruptions. I expected closer to the standard cessation pattern. If this is real (not just survivorship bias of healthiest members who had interruptions), it suggests behavioral coaching is producing durable lifestyle change beyond the medication window.
 **What I expected but didn't find:** A control condition — Calibrate members without behavioral coaching who had treatment interruptions. Without this, we can't isolate whether the behavioral program caused the floor effect or whether Calibrate members are just more health-motivated than average GLP-1 users.
 **KB connections:**
 - Omada post-discontinuation data (same structural question — does behavioral program prevent full rebound?)
 - GLP-1 continuous-delivery requirement debate
 - Behavioral vs. pharmacological intervention durability framing (Sessions 20-22)
 **Extraction hints:**
 - Not a standalone extraction target — use as one of 3-4 data points in a claim about behavioral wraparound providing a durability floor
 - The interrupted access finding is more interesting than the primary outcomes — specifically, that 13.7% maintenance at 12 months with interruptions is dramatically better than standard GLP-1 cessation data
 - Confidence would be: EXPERIMENTAL — promising pattern, not replicated in controlled design
 **Context:** Calibrate targets employer plans. Program cost ranges from $200-300+/month depending on employer negotiation. It's positioned as a higher-intensity, higher-cost program than standard GLP-1 prescribing. Sample is entirely employer-sponsored, which skews toward commercially insured, higher-income populations.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Behavioral wraparound durability floor; continuous-delivery requirement scope qualification
 WHY ARCHIVED: Interrupted access data is the most mechanistically compelling evidence that behavioral coaching produces durable effect beyond the medication window; pairs with Omada post-discontinuation data as converging evidence
 EXTRACTION HINT: Use the interrupted access data (not the primary outcomes) as the key finding — this is the novel contribution. The floor effect at 13.7% is the claim candidate.
--- a/inbox/queue/2026-04-13-frontiers-medicine-2026-deskilling-neurological-mechanism.md
+++ b/inbox/queue/2026-04-13-frontiers-medicine-2026-deskilling-neurological-mechanism.md
@ -0,0 +1,61 @@
 ---
 type: source
 title: "The Deskilling Dilemma: Neurological Mechanism for AI-Induced Clinical Skill Degradation (Frontiers in Medicine, 2026)"
 author: "Frontiers in Medicine (2026)"
 url: https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2026.1765692/full
 date: 2026-01-01
 domain: health
 secondary_domains: [ai-alignment]
 format: article
 status: unprocessed
 priority: medium
 tags: [clinical-ai, deskilling, neuroscience, prefrontal-cortex, automation-bias, cognitive-offloading, ai-safety]
 flagged_for_theseus: ["Neurological mechanism for human skill degradation from AI assistance — relevant to understanding irreversibility of deskilling and the automation alignment problem"]
 ---
 ## Content
 Frontiers in Medicine (2026): "Deskilling Dilemma — Brain Over Automation" (or similar title based on URL slug fmed.2026.1765692).
 **Proposed neurological mechanism for AI-induced deskilling:**
 1. **Prefrontal cortex disengagement:** When AI reliably handles complex reasoning tasks, the prefrontal cortex's analytical processing is reduced. Cognitive load offloaded to AI → less prefrontal engagement → reduced neural pathway maintenance for the offloaded skill.
 2. **Hippocampal disengagement from memory formation:** Procedural and clinical skills require active memory encoding during practice. When AI is handling the problem, the hippocampus is less engaged in forming the memory representations that underlie skilled performance. Skills require formation, not just performance.
 3. **Dopaminergic reinforcement of AI reliance:** AI assistance produces reliable, positive outcomes (performance improvement) that create dopaminergic reward signals. This reinforces the behavior pattern of relying on AI, making it habitual. The dopaminergic pathway that would reinforce independent skill practice is instead reinforcing AI-assisted practice.
 4. **Shift from flexible analysis to habit-based responses:** Over repeated AI-assisted practice, cognitive processing shifts from the flexible analytical mode (prefrontal, hippocampal) to habit-based, subcortical responses (basal ganglia). Habit-based processing is efficient but rigid — it doesn't generalize well to novel situations.
 **Clinical implication of the mechanism:**
 If this mechanism is correct, deskilling may be partially irreversible — not because skills are "lost" in a simple sense, but because the neural pathways were never adequately strengthened to begin with (supporting the never-skilling concern) or because they've been chronically underused to the point where reactivation requires sustained practice, not just removal of AI.
 **The mechanism also explains why deskilling is specialty-independent:**
 The cognitive architecture interacts with AI assistance the same way regardless of the domain — whether radiology, colonoscopy, or medication management. This predicts cross-specialty universality (consistent with Natali et al. 2025 findings).
 **Authors note this is theoretical:**
 The neurological mechanism is proposed based on established cognitive science and analogy to other cognitive offloading research. It has not been tested in a clinical AI context via neuroimaging.
 ## Agent Notes
 **Why this matters:** A proposed mechanism elevates the deskilling concern from empirical observation ("we see skill degradation in these studies") to mechanistic understanding ("here's why this happens and why it might be irreversible"). Mechanisms are more dangerous than patterns because they predict generalization and inform what interventions can and cannot work.
 **What surprised me:** The dopaminergic reinforcement element is underappreciated in the clinical AI safety literature. Most discussions focus on cognitive offloading (you stop practicing) and automation bias (you trust the AI). The dopamine loop (AI-assisted success → reward → more AI reliance) predicts behavioral entrenchment that goes beyond simple habit formation. This makes deskilling not just a training design problem but a motivational and incentive problem.
 **What I expected but didn't find:** Neuroimaging data supporting the proposed mechanism. This is theoretical reasoning by analogy from cognitive offloading research, not an empirical demonstration. That matters for confidence calibration.
 **KB connections:**
 - Natali et al. 2025 (provides the cross-specialty empirical base; this provides the mechanism)
 - Belief 5 (clinical AI creates novel safety risks)
 - Theseus domain: the mechanism is relevant to AI alignment discussions about human-AI collaboration design
 **Extraction hints:**
 - Claim: "AI assistance may produce neurologically-grounded, partially irreversible skill degradation through three concurrent mechanisms: prefrontal disengagement, hippocampal memory formation reduction, and dopaminergic reinforcement of AI reliance" — confidence SPECULATIVE (mechanism is theoretical, not empirically demonstrated via neuroimaging in clinical context)
 - The dopaminergic reinforcement argument is the most novel and extractable element — it predicts behavioral entrenchment beyond simple habit
 - Note: this is a mechanism claim, not a clinical outcomes claim; it supports the deskilling body of evidence but isn't itself an evidence claim
 **Context:** Frontiers in Medicine is an open-access peer-reviewed journal. The article may be a perspective/hypothesis piece rather than an original research study — the URL slug doesn't resolve to a specific research type. Extractor should verify article type.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Clinical AI deskilling claims in health domain; Theseus AI alignment domain
 WHY ARCHIVED: Provides mechanistic foundation for deskilling claims — moves from "we observe skill degradation" to "here's why it happens and why it might be irreversible"; the dopaminergic reinforcement loop is the most novel contribution
 EXTRACTION HINT: Extract as a SPECULATIVE mechanism claim — clearly mark as theoretical. The value is in the mechanism's explanatory power, not empirical proof. Pair with Natali et al. review which provides the empirical base.
--- a/inbox/queue/2026-04-13-jeo-2026-never-skilling-orthopaedics.md
+++ b/inbox/queue/2026-04-13-jeo-2026-never-skilling-orthopaedics.md
@ -0,0 +1,68 @@
 ---
 type: source
 title: "From De-Skilling to Up-Skilling: Never-Skilling Named as Greater Long-Term Threat in Medical Education (JEO, March 2026)"
 author: "Journal of Experimental Orthopaedics / Wiley (March 2026)"
 url: https://esskajournals.onlinelibrary.wiley.com/doi/10.1002/jeo2.70677
 date: 2026-03-01
 domain: health
 secondary_domains: [ai-alignment]
 format: article
 status: unprocessed
 priority: medium
 tags: [never-skilling, medical-education, clinical-ai, deskilling, ai-safety, orthopaedics]
 flagged_for_theseus: ["Never-skilling named formally in peer-reviewed literature as distinct risk category from deskilling; provides language and framing for the AI capability → human deskilling pathway"]
 ---
 ## Content
 Journal of Experimental Orthopaedics (March 2026, Wiley): "From De-Skilling to Up-Skilling" — a review of AI's impact on clinical skill development, specifically naming never-skilling as a formal concern.
 **Key passage (verbatim or close paraphrase):**
 "Never-skilling poses a greater long-term threat to medical education than deskilling; it occurs when trainees rely on automation so early in their development that they fail to acquire foundational clinical reasoning and procedural competencies."
 **Definition established:**
 - *Deskilling:* Loss of skills previously acquired, due to reduced practice from AI assistance
 - *Mis-skilling:* Acquisition of wrong patterns from following incorrect AI recommendations
 - *Never-skilling:* Failure to acquire foundational competencies in the first place, because AI was present during training before skills were developed
 **Why never-skilling is claimed to be worse than deskilling:**
 - Deskilling is recoverable: if AI is removed, the clinician can re-engage practice and rebuild
 - Never-skilling may be unrecoverable: the foundational representations were never formed; there is nothing to rebuild from
 - Never-skilling is detection-resistant: clinicians who never developed skills don't know what they're missing; supervisors who review AI-assisted work can't distinguish never-skilled from skilled performance
 - Never-skilling is prospective and invisible: it's happening now in trainees but won't manifest in clinical harm for 5-10 years, when current trainees become independent practitioners
 **Evidence base (from this and related sources):**
 - More than 1/3 of advanced medical students failed to identify erroneous LLM answers to clinical scenarios — calibration is already impaired
 - Significant negative correlation found between frequent AI tool use and critical thinking abilities in medical students
 - No prospective study yet comparing AI-naive vs. AI-exposed-from-training cohorts on downstream clinical performance
 **Status of the concept in literature:**
 - Formally named in NEJM (2025-2026), JEO (March 2026), Lancet Digital Health (2025)
 - Articulated by NYU's Burk-Rafel as institutional voice
 - ICE Blog commentary (August 2025): physician commentary by Raja-Elie Abdulnour MD amplifying the framing
 - Still classified as: theoretical + observational correlations; no prospective RCT
 ## Agent Notes
 **Why this matters:** Never-skilling has graduated from informal framing to peer-reviewed literature with a formal definition and explicit claim that it's a greater long-term threat than deskilling. This is the conceptual infrastructure needed to write the never-skilling claim in the health domain. The JEO source, combined with the NEJM and Lancet Digital Health citations, gives the claim a peer-reviewed foundation even though direct empirical proof is absent.
 **What surprised me:** The orthopaedics literature is where this appears most explicitly — not radiology or internal medicine. The procedural nature of orthopaedics (where manual skills are central) makes it a natural context for never-skilling concern.
 **What I expected but didn't find:** Any prospective study design attempting to test the never-skilling hypothesis. I expected at least one trial protocol. Not found. The conceptual literature is ahead of the empirical evidence, which is itself an important signal.
 **KB connections:**
 - Belief 5: Clinical AI creates novel safety risks requiring centaur design
 - Existing claim on de-skilling and automation bias (should be enriched/linked)
 - Theseus domain: AI safety, human-AI interaction risks
 - Lancet editorial from Session 22 (also addresses this)
 **Extraction hints:**
 - Primary claim: "Never-skilling — the failure to acquire foundational clinical competencies because AI was present during training — poses a detection-resistant, potentially unrecoverable threat to medical education, distinct from and arguably worse than deskilling"
 - Confidence: EXPERIMENTAL — conceptually grounded, named in peer-reviewed literature, but no prospective empirical proof
 - Note the detection-resistance argument as a key component: the risk is structurally invisible because neither the trainee nor the supervisor can detect what was never formed
 **Context:** JEO is a Wiley-published orthopaedics journal. This likely appeared as a perspective/commentary piece rather than an original research study — the framing and language suggest editorial rather than empirical. Extractor should verify article type.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Existing clinical AI safety claims (deskilling, automation bias) in health domain; Theseus AI alignment domain
 WHY ARCHIVED: Provides the peer-reviewed foundation for extracting the never-skilling claim as a distinct concept from deskilling; moves never-skilling from blog commentary to peer-reviewed literature
 EXTRACTION HINT: Extract as a conceptual claim (EXPERIMENTAL confidence) — the argument for why never-skilling is worse than deskilling (detection-resistance, unrecoverability) is the core contribution, not empirical data
--- a/inbox/queue/2026-04-13-kff-glp1-access-inversion-by-state-income.md
+++ b/inbox/queue/2026-04-13-kff-glp1-access-inversion-by-state-income.md
@ -0,0 +1,69 @@
 ---
 type: source
 title: "GLP-1 Access Inversion: Highest-Burden States Have Lowest Coverage and Highest Income-Relative Cost (KFF + Health Management Academy, 2025-2026)"
 author: "KFF + Health Management Academy"
 url: https://www.kff.org/medicaid/medicaid-coverage-of-and-spending-on-glp-1s/
 date: 2026-01-01
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: high
 tags: [glp1, access-equity, health-equity, medicaid, income-disparities, obesity-prevalence, structural-inversion]
 ---
 ## Content
 **Geographic and income access inversion pattern (KFF + Health Management Academy):**
 States with highest obesity rates (40%+ prevalence): Mississippi, West Virginia, Louisiana — these are predominantly Southern/Midwestern states with low per-capita income.
 Income-adjusted GLP-1 out-of-pocket burden by state:
 - Mississippi/West Virginia/Louisiana tier: ~12-13% of median annual income to maintain continuous GLP-1 treatment at standard injectable prices
 - Massachusetts/Connecticut tier: below 8% of median income for equivalent out-of-pocket burden
 - Standard maintenance pricing: ~$350/month (with manufacturer discount programs); up to $1,000+/month without coverage
 Medicaid coverage as of January 2026:
 - 13 state Medicaid programs cover GLP-1s for obesity under fee-for-service (down from 16 in 2025)
 - 43% of commercial plans include weight-loss coverage
 - GLP-1s = ~1% of all Medicaid prescriptions, but 8%+ of Medicaid prescription drug spending before rebates
 **Access inversion summary:**
 - States with highest obesity prevalence → lowest Medicaid GLP-1 coverage → lowest income → highest out-of-pocket burden relative to income
 - States with lowest obesity prevalence → most likely to have commercial insurance with GLP-1 coverage → higher income → lower relative cost burden
 - The populations most likely to benefit are precisely the populations least able to access
 **Survey data on perceived access:**
 - Over 70% of Americans believe GLP-1s are accessible only to wealthy people
 - Only 15% think they're available "to anyone who needs them"
 - Majority of survey respondents could afford $100/month or less; standard maintenance pricing is ~$350/month even with manufacturer discounts
 **Commercial vs. Medicaid utilization asymmetry:**
 - GLP-1 utilization is 8x higher in commercial than Medicaid on a cost-per-prescription basis
 - Commercial enrollees are on average higher income
 - This creates systematic pattern: higher-income → more likely commercial insurance → more likely covered; lower-income → more likely Medicaid → less likely covered
 ## Agent Notes
 **Why this matters:** The access inversion framing captures something structurally important that "access gap" doesn't. An access gap implies unmet need with a pathway to close it. Access inversion implies systematic misalignment — the infrastructure works against the populations who would benefit most. This is the structural argument for why free market / private insurance + voluntary Medicaid coverage creates systematically worse access for the highest-burden populations.
 **What surprised me:** The income-relative cost data is more dramatic than I expected. In Mississippi, a patient paying out-of-pocket for GLP-1s spends 12-13% of median annual income — that's comparable to what middle-income Americans spend on housing. This is structural exclusion, not price sensitivity.
 **What I expected but didn't find:** Evidence of regional cross-subsidization mechanisms or private philanthropy filling the gap in high-burden low-coverage states. Not found.
 **KB connections:**
 - GLP-1 access infrastructure claims (Sessions 20-22)
 - Medicaid coverage retreat (16→13 states)
 - Wasden 2026 racial disparities (cross-domain: race + income are correlated, so the Wasden finding and this finding are partly measuring the same underlying pattern)
 - Structural misalignment (Belief 3)
 **Extraction hints:**
 - Primary claim: "GLP-1 access follows systematic inversion — states with the highest obesity prevalence have both the lowest Medicaid coverage rates and the highest income-relative out-of-pocket costs, concentrating access failures in the populations with the highest disease burden"
 - Confidence: LIKELY — the structural pattern is clear from multiple data points (KFF coverage data, income data, prevalence data), though the precise income-relative cost calculations require methodological verification
 - Note the 70%/15% survey data as supporting evidence (public perception matches structural reality)
 **Context:** KFF (Kaiser Family Foundation) is a non-partisan health policy research organization — high-quality source. Health Management Academy analysis is industry-focused. Combined, they provide a reasonably complete picture of the commercial dynamics.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: GLP-1 access infrastructure claims and structural misalignment; access equity framing
 WHY ARCHIVED: Provides the geographic/income data to support the access inversion claim; complements the Wasden 2026 racial disparities finding (same structural pattern, different lens)
 EXTRACTION HINT: Extract with the "inversion" framing specifically — not just "access gap." The inversion framing makes a stronger structural argument: it's not that some people lack access (access gap), it's that the system systematically denies access to the highest-burden populations (access inversion).
--- a/inbox/queue/2026-04-13-natali-2025-ai-deskilling-comprehensive-review.md
+++ b/inbox/queue/2026-04-13-natali-2025-ai-deskilling-comprehensive-review.md
@ -0,0 +1,64 @@
 ---
 type: source
 title: "AI-Induced Deskilling in Medicine: Cross-Specialty Mixed-Method Review (Natali et al., Artificial Intelligence Review, 2025)"
 author: "Natali et al. (Springer Artificial Intelligence Review, 2025)"
 url: https://link.springer.com/article/10.1007/s10462-025-11352-1
 date: 2025-01-01
 domain: health
 secondary_domains: [ai-alignment]
 format: article
 status: unprocessed
 priority: high
 tags: [clinical-ai, deskilling, automation-bias, medical-education, ai-safety, cross-specialty]
 flagged_for_theseus: ["Cross-specialty deskilling evidence body directly relevant to AI safety in high-stakes domains; neurological mechanism proposed; automation bias in medical context"]
 ---
 ## Content
 Natali et al. (2025). Mixed-method systematic review of AI-induced deskilling across medical specialties. Published in Springer's *Artificial Intelligence Review*.
 **Specialties covered:** Radiology, neurosurgery, anesthesiology, oncology, cardiology, pathology, fertility medicine, geriatrics, psychiatry, ophthalmology.
 **Cross-specialty pattern (consistent across all specialties):**
 AI assistance benefits performance while present; removes opportunities for skill-building; produces dependence that becomes visible when AI is unavailable. This pattern holds across every specialty examined.
 **Quantitative findings synthesized (some from other sources, compiled here for completeness):**
 1. **Colonoscopy (RCT):** ADR dropped 28.4% → 22.4% when endoscopists reverted to non-AI procedures after extended AI use. ADR stable at 25.3% with ongoing AI. The drop occurred specifically when AI was removed — demonstrating dependency.
 2. **Mammography/breast imaging (controlled study, 27 radiologists):** Erroneous AI prompts increased false-positive recalls by up to 12%, even among experienced readers. Mechanism: automation bias — radiologists anchored on AI output rather than independent read.
 3. **Computational pathology (experimental web-based tasks):** 30%+ of participants reversed correct initial diagnoses when exposed to incorrect AI suggestions under time constraints. Mis-skilling in real time.
 4. **Musculoskeletal imaging / ACL diagnosis:** 45.5% of clinician errors resulted directly from following incorrect AI recommendations, across all experience levels.
 5. **UK general practice / medication management:** 22.5% of prescriptions changed in response to decision support; 5.2% of all cases involved switching from a correct prescription to an incorrect one after flawed system advice.
 **Key mechanism proposed:** AI assistance creates cognitive offloading — clinicians stop engaging the prefrontal cortex's analytical processes when AI handles reasoning. Over repeated exposure, hippocampal engagement in memory formation decreases, and dopaminergic reinforcement of AI-reliance strengthens. Skill degradation follows when AI is unavailable.
 **Natali et al.'s main thesis:** Deskilling is not a side effect of poor AI implementation — it is a predictable consequence of how human cognitive architecture interacts with reliable performance-enhancing tools. The same mechanism that makes expert system assistance effective (reducing cognitive load) also undermines the skill maintenance that cognitive load provides.
 ## Agent Notes
 **Why this matters:** This is the most comprehensive synthesis of clinical AI deskilling evidence found. It moves the deskilling evidence base from "a few individual studies" to "a coherent cross-specialty body of evidence with a proposed mechanism." Combined with the 5 new quantitative findings from this session, the deskilling evidence is no longer preliminary.
 **What surprised me:** The breadth — 10 specialties with consistent pattern. I expected deskilling evidence to be concentrated in specialties with AI-assisted image reading (radiology, pathology, colonoscopy). Finding it consistent in neurosurgery, anesthesiology, and geriatrics is surprising. The cross-specialty universality strengthens the "cognitive architecture problem" framing — it's not about specific AI tools but about how human cognition responds to reliable performance assistance.
 **What I expected but didn't find:** Any specialty where the pattern did NOT hold — a disconfirmation of the cross-specialty claim. Not found.
 **KB connections:**
 - Clinical AI safety claims in health domain (Belief 5, clinical AI safety risks)
 - Session 22 Lancet editorial on preserving clinical skills
 - Theseus domain: AI safety in high-stakes domains, automation bias as alignment-adjacent problem
 - Existing claim on automation bias and diagnostic safety
 **Extraction hints:**
 - Primary claim: "AI-induced deskilling follows a consistent cross-specialty pattern — AI assistance benefits performance while present, but produces cognitive dependency that reduces performance when AI is unavailable — confirmed across 10 medical specialties"
 - Rate: LIKELY (multiple studies, cross-specialty replication, mechanism proposed, but no RCTs across all specialties; some findings from non-RCT designs)
 - Flag for cross-domain link to Theseus: automation bias in medicine is the most concrete domain-specific manifestation of AI alignment risk (human over-reliance)
 **Context:** Springer's *Artificial Intelligence Review* is a peer-reviewed journal. Mixed-method review design means it synthesizes both quantitative studies and qualitative case analyses. Author affiliation and conflict of interest data not retrieved — extractor should check.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Clinical AI safety claims (existing health domain claims on automation bias and deskilling); Theseus domain AI alignment/safety
 WHY ARCHIVED: Most comprehensive cross-specialty synthesis of deskilling evidence; provides the research base for upgrading existing deskilling claim confidence from experimental to likely
 EXTRACTION HINT: Focus on the cross-specialty universality and the proposed mechanism (cognitive offloading → hippocampal disengagement → dependency). Flag for Theseus cross-domain connection.
--- a/inbox/queue/2026-04-13-noom-glp1-engagement-report-persistence-2026.md
+++ b/inbox/queue/2026-04-13-noom-glp1-engagement-report-persistence-2026.md
@ -0,0 +1,70 @@
 ---
 type: source
 title: "Noom GLP-1 Engagement Report: 2.2x Longer Persistence for High-Engagement Users (January 2026 Analysis)"
 author: "Noom (internal engagement report, published February 4, 2026)"
 url: https://www.noom.com
 date: 2026-02-04
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: medium
 tags: [glp1, adherence, behavioral-wraparound, digital-health, noom, engagement, persistence]
 ---
 ## Content
 Noom Engagement Report (January 2026 analysis, published February 4, 2026):
 **Sample:** 30,239 members for persistence analysis; 14,203 for weight loss metrics. Cohort: started GLP-1 programs December 2024–February 2025.
 **Methodology:** Members stratified into engagement quartiles by app opens (capped at 20/day).
 - Bottom quartile (Q1): 244.7 app opens
 - Top quartile (Q4): 2,162.2 app opens
 - Statistical significance confirmed (p < 0.001)
 **Persistence outcomes:**
 - Top engagement quartile persisted on GLP-1 medication 2.2x longer than bottom quartile within first 12 months
 - Q1 (lowest engagement): 2.8 months median persistence
 - Q4 (highest engagement): 6.2 months median persistence
 **Weight loss outcomes:**
 - Top quartile lost 25.2% more weight at week 40 vs. bottom quartile
 - Absolute difference: approximately 8.3 additional pounds
 **Retention signal:**
 - Day-30 engagement: 40% of December cohort returned on day 30 (claimed 10x higher than digital health app average)
 **Noom GLP-1 product suite:**
 1. GLP-1 Companion: behavioral support layer for people already prescribed GLP-1s elsewhere
 2. GLP-1Rx (Microdose program): Noom prescribes medication + behavioral program, starting at $119/month
 3. Components: AI food logging, medication tracking, side effect support, body composition scanning, glucose forecasting, muscle preservation ("Muscle Defense"), gamification
 **PDURS positioning:** Noom updated GLP-1 Companion to prepare for FDA's expected Prescription Drug Use-Related Software (PDURS) framework — attempting to position as regulated software companion to GLP-1 prescriptions.
 **Explicit limitation noted by Noom itself:**
 "These findings reflect observational analyses and report associations/correlations, not proof that engagement causes improved outcomes." Reverse causality acknowledged: people doing well on medication may engage more with app.
 ## Agent Notes
 **Why this matters:** The 2.2x persistence improvement for high-engagement vs. low-engagement users is the clearest engagement dose-response signal in the behavioral wraparound literature. Noom is unusual in explicitly noting the reverse causality caveat in their own report.
 **What surprised me:** That Noom acknowledged reverse causality in their own internal analysis. Most company reports present favorable data without explicitly flagging the confound. This is either genuine methodological integrity or savvy pre-emption of criticism.
 **What I expected but didn't find:** Any randomized comparison of high vs. low engagement (randomizing app access to test causal effect). This doesn't exist from Noom. Also no post-discontinuation data — Noom only reports persistence ON medication, not maintenance after stopping.
 **KB connections:**
 - Behavioral adherence thread (this session)
 - GLP-1 persistence data (14.3% two-year adherence baseline from Sessions 20-22)
 - Digital health intervention effectiveness claims
 **Extraction hints:**
 - The 2.2x persistence finding is extractable as an observational signal, but confidence should explicitly acknowledge the reverse causality problem
 - More useful as a data point in a broader behavioral wraparound claim than as a standalone
 - The PDURS positioning is separately interesting for the regulatory/atoms-to-bits boundary claims — Noom is explicitly trying to convert a behavioral app into regulated prescription software
 **Context:** Noom is a commercial digital health company with significant GLP-1 market aspirations. The $119/month price for their microdose program is substantially cheaper than branded GLP-1s alone. They have financial incentives to show engagement drives outcomes.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Behavioral wraparound for GLP-1 adherence; digital health intervention effectiveness
 WHY ARCHIVED: Provides engagement dose-response data for the behavioral wraparound claim; the reverse causality acknowledgment is noteworthy as methodological transparency
 EXTRACTION HINT: Use as one of 4-5 behavioral wraparound data points, noting the reverse causality caveat. The PDURS positioning detail is separately interesting for regulatory/digital health extractor.
--- a/inbox/queue/2026-04-13-omada-glp1-care-track-post-discontinuation-outcomes.md
+++ b/inbox/queue/2026-04-13-omada-glp1-care-track-post-discontinuation-outcomes.md
@ -0,0 +1,64 @@
 ---
 type: source
 title: "Omada Health Enhanced GLP-1 Care Track: Persistence, Weight Loss, and Post-Discontinuation Outcomes"
 author: "Omada Health (internal analysis)"
 url: https://www.omadahealth.com
 date: 2025-01-01
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: high
 tags: [glp1, adherence, behavioral-wraparound, post-discontinuation, weight-loss, continuous-delivery]
 ---
 ## Content
 Omada Health's Enhanced GLP-1 Care Track analysis (n=1,124 members without diabetes who self-reported GLP-1 use, confirmed via pharmacy claims):
 **Persistence outcomes:**
 - 94% at 12 weeks (vs. 42-80% industry range)
 - 84% at 24 weeks (vs. 33-74% industry range)
 **Weight loss outcomes:**
 - Persisters through 24 weeks: 12.1% body weight loss vs. 7.4% for discontinuers (64% relative increase)
 - 12-month persisters: 18.4% average weight loss vs. 11.9% in real-world evidence comparators
 - 28% greater average weight loss vs. matched non-Care Track members
 **Post-discontinuation outcomes (most significant finding):**
 - 63% of Omada members maintained or continued losing weight 12 months after stopping GLP-1s
 - Average weight change post-discontinuation: 0.8% (near-zero)
 - This is the strongest post-discontinuation data of any commercial program found
 **Program components:** High-touch care team, dose titration education, side effect management, nutrition guidance, exercise specialist support for muscle preservation, access barrier navigation.
 **Methodological caveats:**
 - Internal analysis (not peer-reviewed RCT)
 - Survivorship bias: sample includes only patients who remained in Omada after stopping GLP-1s — not population-representative
 - Self-reported GLP-1 use (though confirmed via pharmacy claims)
 - No randomized control condition
 ## Agent Notes
 **Why this matters:** This is the only data I've found suggesting that behavioral wraparound can produce durable weight maintenance AFTER GLP-1 cessation. The prevailing finding across Sessions 20-22 is that GLP-1 benefits revert within 1-2 years of cessation (continuous delivery required). If Omada's post-discontinuation finding holds in peer-reviewed replication, it would scope-qualify the continuous-delivery thesis: GLP-1s without behavioral infrastructure require continuous delivery; GLP-1s WITH comprehensive behavioral wraparound may produce durable changes.
 **What surprised me:** 63% maintaining or continuing weight loss 12 months post-GLP-1 cessation. I expected near-universal rebound based on STEP 4 trial and other cessation data. The 0.8% average weight change is dramatically better than the ~6-7% average weight regain seen in unassisted cessation. This is a genuine data surprise.
 **What I expected but didn't find:** Peer-reviewed publication of this finding. The data was apparently presented at ObesityWeek 2025 but a peer-reviewed paper has not been published as of April 2026.
 **KB connections:**
 - Directly challenges the "continuous delivery required" thesis being developed from Sessions 20-22
 - Relates to: GLP-1 rebound cessation data (STEP 4 trial pattern)
 - Relates to: food-as-medicine reversion claims from Session 17
 - Relates to: antidepressant relapse patterns from Session 21
 **Extraction hints:**
 - Primary claim candidate: "Comprehensive behavioral wraparound may enable durable weight maintenance post-GLP-1 cessation, challenging the unconditional continuous-delivery requirement" — but ONLY if the extractor notes the methodological limits (observational, survivorship bias, not peer-reviewed)
 - Secondary claim: "Industry-wide GLP-1 persistence at 12 weeks ranges from 42-80% without wraparound programs; programs with high-touch behavioral support report 84-94% — a 20-40 percentage point improvement"
 - Confidence: should be rated EXPERIMENTAL until peer-reviewed replication exists
 **Context:** Omada Health is a digital health company with employer-sponsored programs. They have financial incentives to show strong outcomes. The finding is potentially transformative but requires independent replication.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Claims about continuous-delivery requirement for GLP-1 effects (being drafted from Sessions 20-22 patterns)
 WHY ARCHIVED: Most significant post-discontinuation data found; directly challenges the categorical "continuous delivery required" claim and demands scope qualification
 EXTRACTION HINT: Extract the finding with EXPERIMENTAL confidence and explicit scope — "with comprehensive behavioral wraparound" not "with any GLP-1 program"; flag for divergence consideration against GLP-1 rebound cessation data
--- a/inbox/queue/2026-04-13-uspstf-2018-b-recommendation-glp1-pharmacotherapy-gap.md
+++ b/inbox/queue/2026-04-13-uspstf-2018-b-recommendation-glp1-pharmacotherapy-gap.md
@ -0,0 +1,63 @@
 ---
 type: source
 title: "USPSTF 2018 Adult Obesity B Recommendation Predates Therapeutic-Dose GLP-1s — No Update or Petition in Pipeline"
 author: "USPSTF (United States Preventive Services Task Force)"
 url: https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/obesity-in-adults-interventions
 date: 2018-09-18
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: high
 tags: [uspstf, glp1, policy, obesity, aca-coverage, pharmacotherapy, access-infrastructure]
 ---
 ## Content
 **The 2018 USPSTF Adult Obesity Recommendation (Grade B):**
 Clinicians should offer or refer adults with BMI ≥30 to intensive, multicomponent behavioral interventions (≥12 sessions in year 1). Grade B → ACA Section 2713 mandates coverage without cost-sharing for all non-grandfathered plans.
 **What the 2018 recommendation covered:**
 - Pharmacotherapy was reviewed: 32 pharmacotherapy trials and 3 maintenance trials
 - Medications reviewed: orlistat, liraglutide (lower dose), phentermine-topiramate, naltrexone-bupropion, lorcaserin
 - Decision not to recommend pharmacotherapy: "data were lacking about the maintenance of improvement after discontinuation"
 - Therapeutic-dose GLP-1 agonists (Wegovy/semaglutide 2.4mg, Zepbound/tirzepatide) were ENTIRELY ABSENT from the evidence base — they did not exist at scale when the recommendation was written
 **Current status (April 2026):**
 - The 2018 B recommendation remains the operative adult obesity guidance
 - USPSTF website flags the adult obesity topic as "being updated" — but the redirect points toward cardiovascular prevention (diet/physical activity), not GLP-1 pharmacotherapy
 - No formal petition or nomination for GLP-1 pharmacotherapy as a standalone obesity intervention has been publicly announced
 - No draft recommendation statement on adult obesity with pharmacotherapy scope is visible
 - Children and adolescents obesity recommendation was updated in 2024 — also behavioral-only, did not touch adult pharmacotherapy
 **Policy implication:**
 A new USPSTF A/B recommendation that covers GLP-1 pharmacotherapy would trigger ACA Section 2713 mandatory coverage without cost-sharing for all non-grandfathered insurance plans. This is the most powerful single policy lever available to mandate GLP-1 coverage — more comprehensive than any Medicaid state-by-state expansion approach. It does not exist.
 **The compounding gap:**
 As of April 2026: (1) the most clinically effective obesity pharmacotherapy (GLP-1 agonists) lacks a USPSTF recommendation; (2) the existing recommendation covers only behavioral interventions; (3) no update process is publicly announced; (4) the evidence base that could support an A/B rating (STEP trials, SURMOUNT trials, cardiovascular outcomes data) exists and is substantial. The policy infrastructure has not caught up to the clinical evidence.
 ## Agent Notes
 **Why this matters:** This is the policy gap that most directly addresses the access collapse documented in Sessions 20-22. If USPSTF issues an A/B recommendation covering GLP-1 pharmacotherapy, it would mandate ACA coverage without cost-sharing — more durable and comprehensive than Medicaid state-by-state coverage decisions. The fact that this mechanism doesn't exist and isn't being created is as significant as the Medicaid coverage retreats.
 **What surprised me:** That no formal petition has been filed. The clinical evidence base (STEP trials, SURMOUNT, SELECT cardiovascular outcomes) is now extremely strong. The USPSTF mechanism exists and is the most powerful available. And yet no advocacy organization has apparently filed a formal nomination/petition to initiate the review process. This is a striking gap — the most powerful policy lever is sitting unused.
 **What I expected but didn't find:** A pending draft recommendation or at minimum a formal nomination process. I expected that the strength of the GLP-1 evidence base would have triggered a USPSTF review initiation by 2025-2026.
 **KB connections:**
 - GLP-1 access infrastructure collapse (Sessions 20-22)
 - Medicaid coverage retreat (16→13 states)
 - ACA structural claims (mandate mechanism)
 - BALANCE model (voluntary, not operational) — USPSTF B rating would be the non-voluntary equivalent
 **Extraction hints:**
 - Primary claim: "The USPSTF's 2018 adult obesity B recommendation predates therapeutic-dose GLP-1 agonists and remains unupdated, leaving the ACA mandatory coverage mechanism dormant for the drug class most likely to change obesity outcomes — despite substantial clinical evidence supporting an A/B rating"
 - Confidence: PROVEN — this is a documented policy gap; the facts are verifiable
 - This is a structural claim about policy infrastructure, not a clinical outcomes claim
 - Note: the absence of a formal petition is the most striking gap; extractor should flag this as the policy action item
 **Context:** USPSTF is the independent body whose A/B recommendations trigger ACA Section 2713 mandatory coverage. Their process requires either a self-initiated update or a formal nomination/petition from an outside party. The topic being flagged as "under revision" on their website is encouraging but insufficient — the direction of the revision (toward cardiovascular prevention vs. pharmacotherapy) is the critical question.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: GLP-1 access infrastructure claims; ACA coverage mechanism; structural health policy gaps
 WHY ARCHIVED: Identifies the most powerful single policy lever for mandating GLP-1 coverage — the USPSTF pathway — as dormant and apparently not in motion; this is an extractable structural policy claim
 EXTRACTION HINT: This is a "policy infrastructure gap" claim — specific, falsifiable (either an update is in motion or it isn't), and consequential. Extract with PROVEN confidence (the gap is documented fact). Flag: "what would falsify this" = announcement of a formal USPSTF evidence review scoped to include GLP-1 pharmacotherapy.
--- a/inbox/queue/2026-04-13-wasden-2026-racial-disparities-glp1-prescribing.md
+++ b/inbox/queue/2026-04-13-wasden-2026-racial-disparities-glp1-prescribing.md
@ -0,0 +1,64 @@
 ---
 type: source
 title: "Racial and Ethnic Disparities in GLP-1 Prescribing Narrow With Medicaid Coverage Expansion (Wasden 2026, Obesity)"
 author: "Wasden et al. (Obesity journal, 2026)"
 url: https://onlinelibrary.wiley.com/doi/10.1002/oby.70152
 date: 2026-01-01
 domain: health
 secondary_domains: []
 format: article
 status: unprocessed
 priority: high
 tags: [glp1, racial-disparities, access-equity, medicaid, prescribing-disparities, health-equity]
 ---
 ## Content
 Retrospective pre-post study at a large tertiary care center examining GLP-1 prescribing disparities before and after a MassHealth (Massachusetts Medicaid) coverage change for obesity treatment (effective January 2024).
 **Pre-coverage (before January 2024):**
 - Black patients: 49% less likely to be prescribed semaglutide or tirzepatide vs. White patients (adjusted)
 - Hispanic patients: 47% less likely vs. White patients (adjusted)
 - Disparities were large and statistically significant
 **Post-coverage change:**
 - Disparities narrowed substantially after Medicaid expanded coverage
 - Authors conclude: insurance policy is a primary driver of racial disparities, not provider bias alone
 **Separate tirzepatide prescribing dataset (adjusted ORs vs. White patients):**
 - American Indian/Alaska Native: 0.6
 - Asian: 0.3
 - Black: 0.7
 - Hispanic: 0.4
 - Native Hawaiian/Pacific Islander: 0.4
 **Supplementary finding (wealth-based treatment timing):**
 - Black patients with net worth >$1M: median BMI 35.0 at GLP-1 initiation
 - Black patients with net worth <$10K: median BMI 39.4 (13% higher BMI at treatment start)
 - Lower-income Black patients receive treatment further into disease progression — higher disease burden at access point
 **Author conclusion:** Expanding insurance coverage (specifically Medicaid) substantially reduces racial disparities in GLP-1 prescribing. Policy change, not just provider education, is required.
 ## Agent Notes
 **Why this matters:** This is the strongest causal evidence I've found that Medicaid coverage expansion is the primary lever for reducing GLP-1 racial disparities. The pre-post design with a natural experiment (coverage change) is methodologically stronger than observational cross-sectional studies. Combined with the state coverage retreat (16→13 states covering GLP-1 for obesity), this creates a coherent story: the policy instrument that reduces disparities is being withdrawn.
 **What surprised me:** The magnitude — 49% lower likelihood for Black patients BEFORE coverage change. This is a very large disparity. And that disparities narrowed substantially AFTER coverage change suggests the disparity is primarily structural (coverage) rather than implicit bias. This is an important and somewhat counterintuitive finding — often disparities are attributed to provider behavior, but this data says coverage policy is the primary driver.
 **What I expected but didn't find:** Evidence that the disparities FULLY closed after coverage expansion. "Narrowed substantially" suggests residual disparities remain — provider access, transportation, trust, and other structural factors still matter even with coverage.
 **KB connections:**
 - GLP-1 access infrastructure claims (Sessions 20-22)
 - State Medicaid coverage retreat (16→13 states, Sessions 21-22)
 - Social determinants of health / structural racism claims in the health domain
 **Extraction hints:**
 - Primary claim: "Racial disparities in GLP-1 prescribing (Black: 49% less likely, Hispanic: 47% less likely vs. White) narrowed substantially after Medicaid coverage expansion, identifying insurance policy as the primary structural driver — more than provider bias"
 - Secondary claim: "Wealth-stratified treatment initiation timing for GLP-1s reveals an access-timing inversion: lowest-wealth Black patients present with BMI 39.4 vs. 35.0 for highest-wealth patients — receiving treatment further into disease progression"
 - Both claims are rated LIKELY — pre-post design at one institution; needs replication for PROVEN
 **Context:** This is a peer-reviewed study in Obesity, a major specialty journal. MassHealth's GLP-1 coverage expansion provides a natural experiment. Important caveat: this is a single tertiary care center in Massachusetts — may not generalize to other states or care settings.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: GLP-1 access equity claims; structural racism in healthcare access
 WHY ARCHIVED: Strongest methodological evidence found for the claim that insurance policy (not provider bias) is the primary driver of racial GLP-1 prescribing disparities; natural experiment design gives this causal leverage that cross-sectional studies lack
 EXTRACTION HINT: Extract two separate claims — (1) the racial disparity magnitude and natural experiment result; (2) the wealth-stratified treatment timing finding. Keep them separate for atomic claim structure.
--- a/inbox/queue/2026-04-13-ww-med-plus-glp1-success-program-march-2026.md
+++ b/inbox/queue/2026-04-13-ww-med-plus-glp1-success-program-march-2026.md
@ -0,0 +1,62 @@
 ---
 type: source
 title: "WeightWatchers Med+ GLP-1 Success Program: 61% More Weight Loss Month 1, 21% at 12 Months with Behavioral Integration (March 2026)"
 author: "WeightWatchers (internal analysis, March 2026)"
 url: https://www.weightwatchers.com
 date: 2026-03-01
 domain: health
 secondary_domains: []
 format: report
 status: unprocessed
 priority: medium
 tags: [glp1, behavioral-wraparound, adherence, weight-loss, digital-health, ww-med-plus]
 ---
 ## Content
 WeightWatchers Med+ GLP-1 Success Program internal analysis (March 2026, n=3,260 Med+ members prescribed GLP-1):
 **Weight loss outcomes (medication + behavioral program vs. medication alone):**
 - Month 1: 61.3% more body weight loss with behavioral program vs. medication alone
 - 12-month average: 21.0% body weight loss
 - 24-month average: 20.5% — sustained without significant regain
 **Behavioral program components:**
 - GLP-1 prescriptions via WW telehealth
 - Behavioral platform: nutrition coaching, community, dietitian access, workshops, app tracking
 - Side effect support: 72% of program members reported GLP-1 Success Program helped minimize side effects
 **Persistence comparison:** Not explicitly reported in this analysis (see Omada and Noom for persistence data).
 **Key finding for continuous-delivery debate:**
 - 24-month average (20.5%) shows sustained weight loss, not plateau or regain
 - Duration of data coverage (2 years) partially addresses the continuous-delivery question — though all members are presumably still on GLP-1 at 24 months (no post-discontinuation data from WW)
 **Methodological caveats:**
 - Internal analysis by WeightWatchers — financial incentive to show positive outcomes
 - No sample size, control group details, or statistical methodology disclosed in press release
 - "Medication alone" comparator group: unclear if this is historical data, concurrent comparison, or matched controls — this matters significantly for interpreting the 61.3% month-1 advantage
 ## Agent Notes
 **Why this matters:** The 61% more weight loss in month 1 with behavioral integration is a large effect size and the 24-month sustained data (20.5% without regain) is important for the continuous-delivery vs. durable effect debate. However, WW's data is the least methodologically transparent of the major programs — no sample size or statistical methods disclosed.
 **What surprised me:** The 24-month figure (20.5%) being nearly identical to the 12-month figure (21.0%). This suggests plateau, not continued loss — but importantly, no regain either. Plateau with GLP-1 + behavior is better than the typical cessation pattern (significant regain).
 **What I expected but didn't find:** Post-discontinuation data. WW Med+ doesn't report what happens when members stop GLP-1s — they only report outcomes for current program members. The Omada post-discontinuation data remains the only finding on this.
 **KB connections:**
 - GLP-1 behavioral adherence thread (this session)
 - Omada post-discontinuation data (comparable program type, different finding emphasis)
 - Continuous-delivery requirement debate
 **Extraction hints:**
 - Not a strong standalone extraction target due to methodological opacity
 - Better used as one data point in a broader "behavioral wraparound improves GLP-1 outcomes" claim alongside Omada, Calibrate, Noom data
 - The "72% found program helped minimize side effects" is potentially extractable as a behavioral factor in adherence
 **Context:** WeightWatchers rebranded to WW and launched a telehealth/GLP-1 platform (Med+) to compete with Noom, Calibrate, Omada, Ro. They have significant brand recognition and an existing community platform but are newer to the GLP-1 space than some competitors.
 ## Curator Notes (structured handoff for extractor)
 PRIMARY CONNECTION: Behavioral wraparound for GLP-1 adherence thread
 WHY ARCHIVED: 24-month sustained weight loss data (20.5%) adds to the body of evidence that behavioral programs can extend GLP-1 benefit duration; complements Omada post-discontinuation finding
 EXTRACTION HINT: Use as supporting evidence for a compound claim about behavioral wraparound + GLP-1 outcomes, not as a standalone primary source. Flag methodological opacity.