GiveBetter x Atlanta

Georgia Harm Reduction Coalition

Harm reduction and clinical care, with a conditional naloxone-rescue pathway

Research time: ~9 min on GPT-6 Astra Light
  • Research — reviewed programs, finances and impact evidence.
  • Source audit — checked claims, assumptions and calculations.

Updated: 2026-09-29

Donate

Summary

What do they do? GHRC provides naloxone, syringe services, infectious-disease and substance-use care, linkage and basic-needs assistance. Its five direct-service sites are in metro Atlanta, while mail and telehealth services extend statewide. The modeled pathway is limited to conditional survival benefit from reported collaborative naloxone reversals.

Why we’re interested in this organization:

  • The original 2024 report records 4,000 Narcan reversals in a collaborative statewide effort.

  • Original Form 990 and Schedule D identify the recipient and reconcile $9.323 million of annual expenses.

  • Community naloxone distribution has mortality evidence, although it does not identify this recipient's incremental effect.

Our main reservations:

  • Reported reversals are neither unique people nor deaths prevented; no county breakdown or deduplication method was found.

  • Partner attribution, metro share, incremental rescue, repeat-event adjustment and gift response are judgmental.

  • Funding room is unknown. This is a conditional initial report, not a donor recommendation.

What do you get for your dollar? $11.0M per DALY in Atlanta. DALY is our one-year comparison unit. QALY-based health gains and income-welfare equivalents retain their original assumptions; they are not measured clinical DALYs. Conditional whole-recipient cost per 10 Atlanta-metro QALYs; weak confidence, donor HOLD; not an expected gift price.

4,000 reported Narcan reversal events in the 2024 collaborative statewide report; not independently verified unique people or prevented deaths. 32,000 Narcan units are distribution outputs, not rescues.

1. What do they do?

Georgia Harm Reduction Coalition, Inc., EIN 58-2227958, is the current legal recipient; the original 2024 return names its current website. Its own site lists two Atlanta locations plus Stone Mountain, Woodstock and Norcross, and statewide mail-based naloxone. These establish local service presence, not the fraction of rescued people residing in the pinned 29-county Atlanta-Sandy Springs-Roswell MSA. Current programs Original tax return

2. Monitoring and information sharing

The original 2024 annual report, pages 5 and 12, records 32,000 Narcan units and 4,000 reversals, jointly crediting GHRC, Georgia DBHDD, Georgia Overdose Prevention and NEXT Distro. Page 12 also reports 25,000 individuals received and 6,355 engagements; the CEO narrative describes distribution to 6,355 persons. These denominators are not reconciled, so neither is used as a unique-person count. The model starts with reported reversal events only and gives no underreporting uplift. The report's lives-saved language is not adopted. Page 13 maps service presence across Georgia without county-level reversal or distribution volumes. Original 2024 report

3. Qualitative assessment

A Massachusetts interrupted time-series study found lower overdose mortality where overdose education and naloxone distribution were implemented; adjusted rate ratios were 0.73 and 0.54 for lower and higher implementation. This nonrandomized, older study supports the mechanism, not a GHRC effect size or a death-per-kit conversion. Primary study

A linked Massachusetts cohort found 5.5% one-year mortality among 11,557 people discharged after nonfatal opioid overdose. This supports retaining competing mortality and a finite horizon; it does not calibrate present-day Georgia survival. Primary survival study

The model assigns only one year, a 0.9 survival-area factor and 0.6 health utility after a genuinely incremental rescue. All three are judgments, not observed GHRC outcomes. A 0.5 effective unique-benefit factor reduces duplicate and overlapping follow-up from repeat rescues. A separate 0.05 incremental-rescue factor allows spontaneous recovery, rescue breathing, EMS and other naloxone to avert death anyway. Neither factor is empirically calibrated. Acute withdrawal, delayed care and other adverse health enter the adverse sensitivity; no such net harm is asserted as observed. Infectious-disease care, meals, counseling, screenings, trainings and syringes receive no additional modeled health credit.

4. What do you get for your dollar?

The latest original return located is calendar 2024: $9,322,976 total expenses, including $7,578,168 programs and $1,744,808 management/general; revenue was $9,124,535. Schedule D reconciles audited and tax expenses at the same $9,322,976 with no reported donated-service or other expense reconciling additions. Schedule M reports $542,712 of donated drugs/medical supplies as noncash contributions; this is not added again as though necessarily absent from expenses. Unrecognized volunteer time, partner resources and downstream health-care costs remain unpriced. Return Schedule D Schedule M

The annual report's page 23 expense-like table totals $9,368,880.18, $45,904.18 above the filed/audited reconciliation; its adjacent label says year-end balance. This discrepancy is unresolved. The model uses the original return and separately tests the annual-report amount. It charges the entire legal recipient, not just naloxone or substance-use program expenses. The illustrative central result is 0.84375 metro QALYs, approximately $11.1M per DALY, conditional on the assumptions below. This is neither the measured cost of rescuing a person nor a price for an additional donation.

Model, assumptions and sensitivity

Whole FY2024 legal-recipient expenses reconciled to audited expense in Schedule D; no netting of grants, public revenue or nonlocal services. Unrecognized volunteer, partner and downstream medical resources unpriced; not a societal-cost estimate.

g=0.25 is an analyst scenario prior, anchored only by five metro sites and confirmed statewide operation across 150 counties. It is not an observed share, population allocation or county-count ratio. Sensitivity spans 0–100%; only pinned 29-county residents count locally.

Q_gross=R*d*a*r*T*s*u*f; Q_all=Q_gross-H; Q_metro=Q_gross*g-H; H is a metro harm and is included once in both net population totals; price_per_10=10*C/Q_metro only if Q_metro>0

C: annual recipient costs
9322976 USD (observed). Observed in original return and Schedule D; annual-report discrepancy tested separately. [return24] [d24]
R: reported reversal events
4000 events (observed). Observed as a claim in original annual report, not independently verified; joint effort, no reporting uplift. [annual24]
d: effective unique-benefit fraction
0.5 fraction (judgment). Avoid full independent one-year credit for repeated rescues of the same person; actual recurrence unknown.
a: GHRC causal contribution
0.25 fraction (judgment). Joint delivery includes three named partners; equal names do not imply equal causal contributions. [annual24]
r: incremental survival rescue fraction
0.05 fraction (judgment). Allows alternative EMS/naloxone/rescue breathing and nonfatal events; study supports mechanism, not this prior. [walley]
T: finite follow-up
1 year (judgment). No benefit beyond one year; no lifetime extrapolation.
s: survival-area factor
0.9 fraction (judgment). Competing mortality remains substantial; primary cohort's 5.5% annual mortality is contextual, not direct calibration. [survival]
u: utility while alive
0.60 QALY/year (judgment). Illustrative health utility for survival; no local preference-based measurement.
f: funding response
0.25 fraction (judgment). Unverified additional unrestricted gift effect; includes funding substitution.
g: Atlanta-metro share
0.25 fraction (judgment). Five metro sites plus statewide services establish plausibility only, not volume; tested 0–1. [programs] [annual24]
H: net other adverse health
0.00 metro QALYs (judgment). Central convention only, not proof of no harm.
Verified additional gift capacity
null services/USD (unknown). No current quantified funding gap or gift-created service response verified. [donate]

Conditional alpha central; donor HOLD: Cost: $9.3M; Atlanta health QALYs: 0.84; all-population health QALYs: 3.38. 4000*d .5*a .25*r .05*T 1*s .9*u .6*f .25; metro g .25; H 0. All bridge factors judgmental.

Weak pathway combination: Cost: $9.3M; Atlanta health QALYs: 0.00; all-population health QALYs: 0.02. R4000,d .25,a .1,r .01,T .5,s .8,u .5,f .1,g .05,H0. Not a confidence bound.

Favorable pathway combination: Cost: $9.3M; Atlanta health QALYs: 78.75; all-population health QALYs: 105.00. R4000,d .75,a .5,r .2,T1,s1,u .7,f .5,g .75,H0. Not a confidence bound.

No incremental gift benefit: Cost: $9.3M; Atlanta health QALYs: 0.00; all-population health QALYs: 0.00. f=0; similarly r=0 or a=0 removes positive benefit.

Net adverse health stress test: Cost: $9.3M; Atlanta health QALYs: -1.00; all-population health QALYs: -1.00. f=0,H=1 metro QALY. Hypothetical stress test, not observed harm.

Zero metro rescue share: Cost: $9.3M; Atlanta health QALYs: 0.00; all-population health QALYs: 3.38. Central except g=0; presence of sites does not prove residence of rescued people.

100% metro share sensitivity: Cost: $9.3M; Atlanta health QALYs: 3.38; all-population health QALYs: 3.38. Central except g=1; deliberately generous bound, not an observation.

Alternate published cost total: Cost: $9.4M; Atlanta health QALYs: 0.84; all-population health QALYs: 3.38. Uses unreconciled annual report page 23 cost-like total; other factors central.

Counterfactual: r accounts for survival through alternative rescue/spontaneous recovery; f accounts for substitution or no service response to an additional unrestricted gift. No lifetime survival extrapolation.

Attribution: a discounts joint GHRC/DBHDD/GOP/NEXT outcomes; d discounts repeat-event overlap. Four partners do not establish equal causal credit: a=0.25 is judgment, not division as an attribution estimator.

Weak confidence. Every bridge beyond the reported reversal count and legal-recipient cost is judgmental. Scenarios are conditional, not probability-weighted expected values or confidence intervals. Zero benefit remains plausible.

Sensitivity

  • Halving any positive bridge factor doubles the conditional price; zero removes finite price.
  • Metro share spans 0–100% explicitly; current 25% is not measured.
  • Repeat-event factor 1 doubles central benefit; no independent person-level deduplication is available.
  • No health credit for kits, trainings, meals, screenings or clinical care beyond the modeled rescue pathway.
  • Annual-report cost total increases price by about 0.49%; bridge uncertainty is much larger.

Unresolved inputs

  • Deduplicated reversal dates, people and county of residence
  • GHRC versus partner causal contribution and external resources
  • Counterfactual survival without GHRC-delivered naloxone
  • Local repeat-overdose and one-year survival follow-up
  • Preference-based utility after rescue
  • Current unrestricted funding gap and staffed additional capacity
  • Reconciliation of 6,355 versus 25,000 recipient denominators and annual-report expense discrepancy

5. Funding and previous grants

GHRC's current own-site Donate Now link routes to Stripe. Payment was not executed, and script-dependent checkout completion and any enforceable local restriction were not verified. No quantified current unfunded capacity, marginal service budget or additional gift-created rescues were found. The 0.25 funding-response factor is an explicit scenario prior, with zero allowed; historical expense or a deficit does not establish funding room. Keep donor HOLD until current need, recipient-specific deduplicated outcomes, geographic attribution and additionality are verified. Official route context

Annual expenses

Organization-level spending, including programs, administration and fundraising. The research list averages three comparable, consecutive full fiscal years when available.

  • FY 2024: $9.3M; Georgia Harm Reduction Coalition, Inc., 12-month period, Original Form 990 full functional expense, matching audited expense per Schedule D; unrecognized resources unpriced. Source

6. Sources

  1. Original 2024 annual report, visually inspected pages 5, 12, 13, 23. Georgia Harm Reduction Coalition. Published: not stated; retrieved: 2026-09-29.
  2. Original FY2024 Form 990, IRS reconstruction. GHRC / IRS. Published: not stated; retrieved: 2026-09-29.
  3. Original FY2024 Schedule D. GHRC / IRS. Published: not stated; retrieved: 2026-09-29.
  4. Original FY2024 Schedule M. GHRC / IRS. Published: not stated; retrieved: 2026-09-29.
  5. Current programs and five sites. GHRC. Published: not stated; retrieved: 2026-09-29.
  6. Official page linking Donate Now and annual reports. GHRC. Published: not stated; retrieved: 2026-09-29.
  7. Opioid overdose rates and implementation of overdose education and nasal naloxone distribution in Massachusetts. Walley et al., BMJ. Published: 2013-01-31; retrieved: 2026-09-29.
  8. One-Year Mortality of Patients After Emergency Department Treatment for Nonfatal Opioid Overdose. Weiner et al., Annals of Emergency Medicine. Published: not stated; retrieved: 2026-09-29.