What we learned from 303 reports
Research synthesis · 18 October 2026
Summary
We have published 303 reports: 191 across 11 city, state and national editions, 91 in the San Francisco Bay Area, and 21 earlier national and international comparisons. Five lessons hold across all of them.
- The cheapest local opportunities remove a specific barrier to care that already exists. Coordinating donated surgery and dentistry, free clinics for uninsured adults, harm reduction and naloxone, and vision and hearing access lead in city after city.
- Looking harder usually makes an estimate worse. When an initial estimate went through an in-depth review, the price rose in 19 of 26 cases, typically by about 1.9×, and 3 lost their estimate entirely.
- Where you draw the boundary changes the answer. The same organization can differ several-fold between editions, depending on how much of its work reaches the people in that edition.
- No local report comes close to the best global health charities. Our best current local estimate is roughly 70 times the cost of the Against Malaria Foundation estimate.
- We have not found a verified marginal-gift offer anywhere. Every price is a conditional research estimate. Before a larger gift, a donor still needs to confirm what an extra dollar would add.
Where we looked
Each edition began with about 100 candidate organizations, published 25 initial reports, took ten into in-depth review and named four featured opportunities. “Lowest” and “median” are current central estimates; they are not rankings of recommendation strength.
Early reports also exist for Seattle, Boston, Atlanta and Detroit; those editions have no in-depth reviews yet.
What keeps looking promising
The same few mechanisms surface in every city. None is proven best; each has a known way of failing that a donor should check.
Coordinating donated clinical care
In several editions, some of the lowest estimates come from organizations that match uninsured patients with surgeons, dentists or equipment that would otherwise go unused. The charity pays for coordination; clinicians and hospitals donate the expensive part. The catch is that donated time is still a real cost, and a gift only helps if more volunteers, operating-room slots or devices are actually available.
Examples: Operation Access (California, $332K per DALY); Surgery on Sunday (USA, $124K per DALY); Dental Lifeline Network (USA, $135K per DALY); The ReCARES Network (SF Bay, $26K per DALY)
Free clinics for uninsured adults
Small free clinics recur near the top in Chicago, Denver, Houston, Los Angeles and New York. They treat people with few alternatives at low cost per visit. The main uncertainty is the counterfactual: where Medicaid or a community health center would have served the same patient, the clinic adds less than its visit count suggests.
Examples: Old Irving Park Community Clinic (Chicago, $356K per DALY); Tri City Health Partnership (Chicago, $330K per DALY); Lestonnac Free Clinic (Los Angeles, $802K per DALY); Parker Family Health Center (New York City, $1.0M per DALY); Doctors Care (Denver, $785K per DALY)
Harm reduction and naloxone
Harm-reduction programs rank among the cheapest local estimates in most cities we studied. Overdose is common, often fatal and cheaply reversible. Our estimates count unique people and finite survival rather than treating every reversal as a lifetime saved, a correction that first changed our HEPPAC report and now applies everywhere.
Examples: Harm Reduction Action Center (Denver, $36K per DALY); Chicago Recovery Alliance (Chicago, $264K per DALY); Harm Reduction Services (California, $685K per DALY); People’s Harm Reduction Alliance (Seattle, $203K per DALY); HOPE Pacifica (SF Bay, $93K per DALY)
Vision and hearing access
Glasses and hearing aids are cheap relative to the years of better function they can provide. The questions are whether the device fits, whether people keep using it, and whether they would have obtained one anyway.
Examples: Vision To Learn (California, $428K per DALY); Help America Hear (USA, $307K per DALY); Hearing and Speech Center of Northern California (SF Bay, $151K per DALY)
National policy research
The lowest numbers in the whole library are national policy organizations. They are also the most assumption-dependent: a small change in the assumed chance of influencing a rule moves the estimate by orders of magnitude. Treat them as high-variance bets, not safer choices.
Examples: Institute for Progress (USA, $33K per DALY); Center for Science in the Public Interest (USA, $113K per DALY)
What usually looks expensive
Violence intervention, supportive and HIV housing, residential children’s services, medically tailored meals and local environmental-justice organizing tend to sit at the costly end in every edition. Examples: Chicago CRED (Chicago, $23.0M per DALY); Youth ALIVE! (California, $47.6M per DALY); Common Justice (New York City, $14.1M per DALY); Allies in Hope (Houston, $35.1M per DALY); Illumination Health + Home (Los Angeles, $49.4M per DALY); Tennyson Center for Children (Denver, $70.6M per DALY); Community Servings (Boston, $31.4M per DALY); WE ACT for Environmental Justice (New York City, $36.1M per DALY).
That is not a verdict that this work is low-value. It costs more per person served, its health effects are harder to attribute, and much of its value (safety, rights, dignity, community power) falls outside the health-and-income measure we use. A donor who values those outcomes directly should read these reports as partial estimates.
What deeper research changed
In-depth reviews keep the initial estimate for comparison. Of 26 reviews where both exist, 19 raised the price, 4 lowered it and 3 withdrew it. The largest rises: Champions for Health ($3.3M per DALY to $30.7M per DALY); Sanctuary for Families ($3.5M per DALY to $20.7M per DALY); The Rose ($4.2M per DALY to $23.8M per DALY).
The usual causes were the same everywhere: counting unique people instead of contacts, using finite effect durations, assigning the whole organization’s cost rather than its most favorable program, and crediting only care that would not have happened anyway. The lesson for comparing reports: an initial estimate is an optimistic first look. An organization that has not had an in-depth review should not outrank one that has merely because its number is lower.
Income cuts both ways. Of 139 reports with a nonzero household-resources estimate, 65 found a net loss from fees, time, travel or lost benefits. We keep those losses instead of setting them to zero.
Geography changes the answer
6 organizations appear in more than one edition. Their estimates differ because each edition credits only the benefits that reach its own residents, and because the cost scope can differ:
- National Health Law Program: California $2.8M per DALY · USA $886K per DALY
- Coalition for Clean Air: California $1.0M per DALY · Los Angeles $1.0M per DALY
- End Overdose: USA $271K per DALY · California $1.5M per DALY
- Dental Lifeline Network: USA $135K per DALY · Denver $3.4M per DALY
- Breathe Southern California: Los Angeles $42.8M per DALY · California $22.4M per DALY
- Didi Hirsch Mental Health Services: Los Angeles $23.1M per DALY · California $23.8M per DALY
Compare organizations within an edition, not across editions, unless the reports use the same boundary.
The global gap
The Against Malaria Foundation estimate is $378 per DALY and New Incentives is $387 per DALY. Our best current local estimate is $26K per DALY. The models differ, but the gap is too large for those differences to close.
Only 2 of 297 U.S. reports fall under our $10K per DALY screening hypothesis (Remedy Alliance / For The People, New Eyes for the Needy), and those are older whole-gift estimates. A donor who cares only about health per dollar should give globally. Our local editions answer a different question: if you want to give where you live, where does a dollar go furthest?
San Francisco: ten giving priorities
Our Bay Area work came first and is the most developed. Our first places to investigate are ReCARES, Project Homeless Connect and Pacific Hearing Connection; Pacific Vision Foundation is a fourth lead. GLIDE’s housing-prevention model shows why housing and policy stay in scope even when their health estimates are less certain.
Editorial order, not numerical rank. None has a verified marginal health offer in the reviewed public evidence. The questions below are for a prospective donor; we have not contacted these organizations.
1. The ReCARES Network
Gift scope: Whole organization; partial equipment-health estimate.
General support route and physical handling constraints are documented. Existing training/software support does not establish an unfunded safe-placement increment.
What to confirm before a larger gift
- What dated, fully costed inspection, repair, storage or access increment would an additional gift enable, with staff and space available?
- How many additional unique Bay recipients would receive useful equipment beyond current funding and realistic alternatives, and will a purpose restriction be accepted?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2
2. Project Homeless Connect
Gift scope: Whole designated sponsored project, not whole Community Initiatives; partial clinical-health estimate.
Project-designated giving, a dental waitlist and denture lottery are documented. Donation examples and a coordinator vacancy do not establish additional funded clinical slots.
What to confirm before a larger gift
- Which eligible clinical queue could the gift shorten, with confirmed partner appointments and complete coordination, device and follow-up costs?
- What current project funds and alternative coverage already support those patients, and will the fiscal sponsor accept the intended clinical designation?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
3. Pacific Hearing Connection
Gift scope: Whole accounting cost; modeled hearing-care component and unpriced donated resources.
Transport, language and device needs are described. The $129,987 Health Authority award supports training, education and mobile screening; $500 device contribution examples are not complete additional-fitting prices.
What to confirm before a larger gift
- What additional completed fittings and supported use can a dated gift fund, including clinician capacity, suitable devices, transport, testing and follow-up?
- How do patient fees, inventory, Health Authority and Ability Central commitments affect that increment, and will a fitting-specific restriction be accepted?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3 · Source 4
4. Pacific Vision Foundation
Gift scope: Whole ordinary gift; first-eye earlier-care health only. Remaining 90% of assumed allocation health unquantified; historical restricted-program comparison retained.
Patient sponsorship and a $250 monthly surgery-support tier are advertised. Neither $250 nor its annual arithmetic establishes a complete additional-procedure price or accepted Bay/first-eye restriction.
What to confirm before a larger gift
- How many additional completed eligible surgeries would the proposed gift enable, at full cost with committed surgeon, facility and follow-up capacity?
- Will patient-care and relevant geography/procedure restrictions be accepted, and what funding and treatment timing would apply without the gift?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
5. Dentists on Wheels
Gift scope: Whole clinic organization; finite symptomatic relief only.
Lottery-like access and expansion aspirations are documented. $250 patient and $500 denture donation examples lack dated extra sessions, complete costs and a funding counterfactual.
What to confirm before a larger gift
- What additional dated clinic sessions and unique completed symptomatic-treatment episodes could a gift fund, with committed clinical staff and a full budget?
- Which patients otherwise wait or obtain alternative care, and which existing funds or partner subsidies already cover those episodes?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
6. Face to Face — Sonoma County AIDS Network
Gift scope: Whole organization cost; finite unique-person overdose-survival component, with other portfolio health unquantified.
Ongoing naloxone delivery and retrospective categorical rescue reports support the mechanism. Public support and existing distribution remain baseline; reported rescues do not identify prospective hazards or cash-created coverage.
What to confirm before a larger gift
- What dated additional outreach, supply-access or response-support activity would a gift enable beyond current public and donor commitments, at full cost?
- What unique-person, repeat-event and alternative-rescue evidence supports the expected additional coverage, and can the intended use be accepted and tracked?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
7. Friends of the Urban Forest
Gift scope: Whole organization; additional tree planting/care and uncertain finite health transfer.
Current yard-tree funding extends through August 2027; this does not establish street-tree funding room. Historical contract and donation examples are not a priced backlog of additional surviving trees.
What to confirm before a larger gift
- Which approved but unfunded sites or establishment-care tasks could a dated gift activate, at full cost and with expected additional surviving tree-years?
- How do current city maintenance, planting contracts, grants and donor commitments cover the same sites, and will a non-substituting purpose restriction be accepted?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
8. Breathe California
Gift scope: Whole organization gift cost; partial clinical health from cessation, child/adult asthma and prescription CPAP access. Remaining services and outside resources are incompletely quantified.
Current services and CPAP inventory constraints do not establish cash-created additional care. Public-funded asthma delivery, free cessation alternatives, donated machines and external clinical calibration remain part of the baseline and resource boundary.
What to confirm before a larger gift
- What dated additional completed care could the gift fund—supported quitting, asthma home visits or safely calibrated CPAP episodes—with a full budget, suitable inventory and committed clinical capacity?
- Which existing public payments, patient fees, other donors and realistic alternatives already cover that care, and will the organization accept and track the proposed purpose, Bay recipients and useful outcomes?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3 · Source 4 · Source 5
9. GLIDE Foundation
Gift scope: Whole Foundation gift cost; partial clinical and housing health. Not a separate Memorial Church gift or a complete organization return.
Current Foundation services and historical consolidated accounts do not establish a Foundation-only marginal budget. Public prescribing, HSH/Abode rental support, partner resources and other donors remain in the baseline. A general gift is not an accepted clinical or rental restriction.
What to confirm before a larger gift
- Which additional completed clinical or housing services would a dated Foundation gift fund, with a complete budget and confirmed staff or partner capacity?
- How do public funding, private commitments, existing treatment and household alternatives cover the same services, and will the Foundation track the proposed allocation and unique Bay recipients?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
10. North East Medical Services
Gift scope: Foundation EIN94-3171797 is separate from operator EIN94-1722562. Ordinary-gift allocation and Bay health are unestimated; recurring HBV model is conditional only.
Current broad Foundation campaign cites flu vaccination and mold mitigation, not a priced HBV offer. Existing ReLink and routine care overlap the conditional model.
What to confirm before a larger gift
- Which additional Bay patient panel or extension is outside current ReLink and routine-care commitments, with a dated full navigation and downstream-care budget?
- What support duration, retained monitoring/treatment and alternative-care assumptions are justified, and will the Foundation accept and transfer the restricted gift to that work?
Primary sources, reviewed 11 September 2026: Source 1 · Source 2 · Source 3
Method and limits
Each estimate divides the reviewed cost by modeled health gains plus signed household-resource changes for residents of that edition. Reports share clinical evidence and assumptions, so 303 reports are not 303 independent experiments. Initial reports are single-pass; in-depth reviews were independently checked.
Public donation pages, waitlists and reserves do not prove a funding gap. A useful next step is a dated, accepted gift purpose with full delivery costs, available capacity and the expected additional care. No organization is required to guarantee outcomes.
Each linked report keeps its sources, dates, model and uncertainties. The original 25-program San Francisco ledger is historical evidence, not a current recommendation list. Read all research or browse the earlier national and international reports.