Date: 2026-08-11. Extends Phase 0 / protocol anchor 7 (988 volume) and adjudicates H4. Perishable: answer rates, routing, and state telecom-fee counts move on SAMHSA/FCC and state budget clocks.
PRISMA-lite search log
| Field | Logged |
|---|---|
| Strings | 988 Lifeline Vibrant SAMHSA answer rate volume; GAO-26-108114 988 capacity; GAO-25-107586 crisis response; KFF 988 two years after launch; 988 ED diversion causal evaluation multi-state; mobile crisis arrest diversion Michigan Swanson; behavioral health crisis care ED utilization walk-in stabilization |
| Date range | Prefer post-July 2022 (988 transition) for volume; pre-988 mobile-crisis / crisis-system papers retained only as continuum comparators |
| Inclusion | Vibrant/SAMHSA/GAO/KFF performance series; peer-reviewed diversion or crisis-system evaluations with methods. Exclude advocacy diversion claims without a stated comparator design. |
| Searched | 2026-08-11 |
1. Volume and answer performance — LOCKED (anchor 7 sharpened)
| Source | Construct | Figure |
|---|---|---|
| Vibrant monthly KPIs (Phase 0) | Answered network contacts ex-VCL, Jan | 161,267 (Jan 2022) → 403,544 (Jan 2024) — ~2.5× that month pair |
| GAO-26-108114 | Contacts routed to centers, Jul 2022–Sep 2025 | ~19.1M calls/texts/chats; overall contact volume ~+90%; calls +87%, texts +260%, chats +23% |
| KFF (Jul 2024; Lifeline through May 2024) | Contacts since launch (incl. VCL in headline) | ~10.8M contacts by May 2024; May 2024 monthly contacts >0.5M (~+80% vs May 2022) |
| KFF / Lifeline | National answer rate / wait | Answer rate 70% (May 2022) → 89% (May 2024); wait 2:20 → 1:31 (year-1 peak better: ~93% / 35s) |
SAMHSA capacity frame (per GAO): volume, answer rates, average time to answer — not ED or arrest diversion. Local answer share rose; ~76% of calls answered locally vs most texts/chats at national backup over the GAO window.
Anchor 7 verified cell (sharpened): Direction holds. Prefer Jan answered ~2.5× (2022→2024) or GAO ~+90% routed contacts (Jul 2022–Sep 2025) over a single official “doubled in two years” headline (none located).
Limits: Lifeline covers ~200 of ~544 crisis centers nationally (KFF); state vs administrator metric systems can diverge (GAO interviews).
2. What remains unmeasured for 988 itself
KFF and GAO both flag the outcome gap: published Lifeline metrics do not establish (a) precipitating clinical circumstances, (b) resolution / warm handoff to mobile crisis or outpatient care, or (c) whether the contact diverted an ED visit or arrest. State dashboards (AZ, UT, SD, WI cited by KFF) sometimes add referral/reason fields; they are not a national causal diversion instrument.
GAO-25-107586 (crisis-response programs) describes SAMHSA grant and continuum investments and notes evaluation status — it does not deliver a multi-state causal ED/arrest diversion estimate attributable to 988.
3. Continuum evidence ≠ 988 diversion evidence
Adjacent crisis-system studies exist; none located this pass is a multi-state causal estimate that 988 diverted ED visits or arrests:
| Study | Design | Finding | Why it does not refute H4 |
|---|---|---|---|
| Swanson et al. / Michigan multi-site mobile crisis (Psychiatric Research & Clinical Practice lineage) | IPTW; mobile vs LE-only | Mobile crisis associated with ~45% lower 11-month arrest incidence vs LE-only | Single-state; mobile response, not 988 call diversion |
| Kalb et al., Health Serv Res 2025 (PMC11911218) | Zip-level BHCC availability vs MBD ED use | Walk-in crisis stabilization associated with ~2.8% lower mean MBD ED utilization per additional facility | Facility availability, not 988 routing |
| Arizona comprehensive crisis system vs comparison states (Mark et al. / related AZ evaluations) | State crisis model vs peers | Mixed/small hospitalization effects; not a 988-only instrument | Whole-system, not national 988 causal |
Independence audit: these roots do not cross-validate a national or multi-state 988→ED/arrest causal claim. Treat them as continuum steelman material for §7 / architecture #5, not as H4 refute evidence.
4. H4 adjudication
H4 — 988 grew calls without proving diversion. Connection volume rose; measured ED/law-enforcement diversion is thin or single-root.
| Criterion | Result |
|---|---|
| National answer-volume gains documented | Met (Vibrant KPI pairs; GAO-26-108114; KFF Lifeline extract) |
| and no multi-state causal estimate of ED or arrest diversion survives an independence audit | Met — no such 988-attributable estimate located; continuum studies are different instruments |
| Refute path (≥2 independent evaluations show material diversion) | Does not fire for 988 |
| Indeterminate if only vendor-reported metrics | Vendor + GAO federal audit of SAMHSA data — volume leg is not vendor-only |
Adjudication: Supported.
Implications for scorecard
- Architecture #5 (988 + mobile crisis + stabilization) scores volume/access-to-line as achieved and diversion as unproven at national scale — fund the middle tier on continuum evidence, not on a 988 diversion proof.
- Do not treat answer-rate gains as O2 acute-capacity substitutes; they measure contact-center throughput.
- Perishable: georouting, telecom fees, specialized-line status, and SAMHSA answer-rate goals (GAO notes near-term text/chat local-answer goals).
Confidence: Strong (volume) / Moderate–weak (diversion absence)
Strong on volume and answer metrics (administrator + GAO). Strong that a multi-state 988 diversion causal series was not located; weak that none could ever exist — future state 911→988 diversion evaluations could move H4.