[new-pack] obstetric-fistula-repair-access-global — separate repair backlog from prevention failure

[cschanhniem]

Summary

Discovery mode. Adds public-health/obstetric-fistula-repair-access-global, a scoped problem pack for separating untreated obstetric fistula repair backlog from continuing incident prevention failure, repair-capacity gaps, and measurement artifacts.

This is deliberately narrower than generic maternal health. The decision wedge is whether a UNFPA campaign lead, ministry maternal-newborn director, foundation chair, or fistula repair network should fund prevalence measurement, case finding, surgical repair capacity, referral transport, or EmONC prevention first.

Discovery Ranking Table

Problem	Scale (source)	Tractability	Neglectedness	Evidence gap	Why not the others
Obstetric fistula repair backlog and prevention access	UNFPA estimates about 500,000 women and girls living with fistula globally; BMJ Global Health modelling estimates about 457,000 women ages 15-64 living with obstetric fistula in 2020 across 55 LMICs (UNFPA, BMJ GH DOI).	Repair is surgically tractable, prevention depends on timely obstructed-labour care, and UNFPA reports 129,000+ directly supported repairs through the Campaign to End Fistula (UNFPA).	The problem persists despite being almost entirely preventable; survivors are isolated, stigmatized, and often invisible to policymakers (UNFPA).	Distinguish untreated backlog from incident prevention failure, repair-output limits, and symptom-survey artifacts before allocating repair, case-finding, referral, or EmONC funds.	Chosen: strongest decision wedge because measurement family confusion directly changes allocation.
Rheumatic heart disease detection and secondary prophylaxis gaps	WHO/partners estimate about 40 million people live with RHD and more than 300,000 die annually (WHO).	Prevention and secondary prophylaxis are known, but diagnosis, adherence, and valve-surgery access are bottlenecks.	Neglected relative to burden in LMICs, but there are established RHD alliances and WHO framing.	Which school-screening or pregnancy-detection pathways produce durable prophylaxis adherence rather than diagnosis-only activity?	Strong candidate, but existing cardiovascular and maternal-risk adjacency makes the initial wedge less unique than fistula backlog-versus-incidence separation.
Scabies control and MDA targeting in crowded LMIC settings	WHO estimates scabies affects more than 200 million people at any time and more than 400 million cumulatively every year (WHO).	Ivermectin/permethrin treatment and mass drug administration can be tractable in defined settings.	WHO recognizes scabies as common in LMICs, but surveillance and complication attribution are weak.	Which communities have evidence strong enough to justify MDA versus ordinary treatment access, especially where impetigo and kidney/cardiac sequelae are undercounted?	Very large scale, but severity and allocation consequence are more heterogeneous; a pack needs tighter decomposition before it beats fistula.
Cataract surgical backlog quality and access	WHO's World Report on Vision estimated at least 2.2 billion people have vision impairment or blindness, with at least 1 billion preventable or unaddressed (WHO World Report on Vision).	Cataract surgery is highly tractable and often cost-effective.	Already has stronger institutional evidence pipelines and overlaps the repo's existing blindness/vision impairment pack.	Distinguish surgical volume from effective cataract surgical coverage and post-operative quality.	Disqualified for this run because the portfolio already has public-health/blindness-vision-impairment-global; this is enrichment, not discovery.

Files Changed

• Added full canonical pack under problem-packs/public-health/obstetric-fistula-repair-access-global/.
• Regenerated tasks-available.json, agent-radar.json, and generated Wiki pages.
• Repaired one source-rot URL in problem-packs/biodiversity/wetland-loss-migratory-birds-global/evidence.json so source verification passes.
• Applied Prettier table formatting to five docs introduced on current main; validation failed on those files before this mechanical fix.

Evidence Added Or Changed

Added four initial evidence records:

• unfpa-obstetric-fistula-topic-2022
• bmjgh-global-fistula-prevalence-2025
• plosone-fistula-repair-success-2024
• unfpa-secretary-general-fistula-report-2024

Changed one existing evidence source URL:

• international-waterbird-census-coverage: moved stale Wetlands International path to https://www.wetlands.org/tools/international-waterbird-census/.

Validation Method

• pnpm build passed; generated indexes now report 106 packs and 101 scoped tasks.
• pnpm validate passed.
• pnpm reproducibility:check passed for 549 tasks.
• pnpm format:check passed.
• pnpm verify:sources passed after source-rot repair: 330 evidence source URLs verified, with only existing allowlisted warnings.
• Pre-push hook reran validate and build successfully.

Known Limitations

• The BMJ prevalence record is intentionally treated as burden framing, not direct proof of confirmed untreated country backlog.
• The pack does not claim that repair count means need is falling.
• The first scoped task is still source-inventory; country selection and allocation signals are blocked until measurement-family reconciliation exists.
• Any operational ranking remains medium/high-risk and requires domain review, field-reality review, and replication.

Kill Condition

The pack's allocation wedge should be rejected or narrowed if source inventory shows that public prevalence, repair-output, and prevention-capacity sources cannot distinguish untreated backlog from incident prevention failure in any high-burden country with enough reliability to change a real funding or program decision.

Reviewer Type Needed

Domain reviewer for maternal-newborn health, fistula repair programs, or global surgery access. Field-reality review is required before any country feasibility or allocation artifact is treated as actionable.

Protocol Notes

Current main contained five Markdown tables that failed pnpm validate through Prettier. This PR includes the mechanical format repair because otherwise a fresh rebased branch could not pass the repo's own validation gate.

Daily Run Checklist

• Read AGENTS.md, SKILL.md, SAFETY.md, all relevant schemas
• Chose mode: discovery
• Ranked candidate problems, picked the strongest, drafted full pack, included ranking table in PR
• Every claim has a kill condition: no claims.json claim object added; PR-level kill condition included
• Every evidence record has dated source, stable URL, method, limitations
• pnpm validate passes
• pnpm reproducibility:check passes
• pnpm verify:sources passes
• pnpm build passes, regenerated files committed
• PR opened with required body fields
• Protocol Notes included

Mode chosen: discovery. Pack targeted: public-health/obstetric-fistula-repair-access-global. One-line verified-knowledge addition: the portfolio now has a reviewable pack for distinguishing obstetric fistula repair backlog from prevention failure before funders allocate repair, surgical workforce, referral, or EmONC resources.