GiveBetter x New York City

OnPoint NYC

Integrated harm reduction, overdose prevention and low-barrier health support

Research time: 6 min on GPT-6 Astra Light
  • Research — reviewed programs, finances and impact evidence.

Updated: 2026-09-14

Donate

Summary

What do they do? OnPoint NYC combines overdose prevention with low-barrier clinical, harm-reduction and basic-needs services. We assess the merged recipient rather than treating its overdose rooms as the whole organization. Our weak-confidence estimate is about $38.5 million per 10 MSA-resident QALYs for a finite overdose-survival component, with other health pathways unpriced.

Why we’re interested in this organization:

  • Immediate monitored care addresses an acute and preventable risk.

  • Current city records confirm continued local delivery and public support.

  • Original returns expose whole-recipient costs and donated medical property.

Our main reservations:

  • Intervention counts are not observed lives saved.

  • The unique-person distribution, counterfactual lethality and marginal funding response are judgment-based.

  • Substantial public funding and other services make a single-component price an incomplete portfolio assessment.

What do you get for your dollar? $38.5M per better life: ten additional quality-adjusted life years in New York City. Whole On Point NYC recognized annual expense charged to a conditional, finite overdose-survival component for MSA residents. Clinical, infection and other health effects remain unpriced; recognized donated medical goods are included and unrecognized partner resources are stress-tested. Not a measured whole-portfolio return or proven lower bound..

Reported intervention events -> unique-person annual hazard difference -> finite discounted survival difference, not intervention events equated with deaths.

1. What do they do?

OnPoint NYC formed from New York Harm Reduction Educators and Washington Heights Corner Project in 2022. This report uses On Point NYC Inc., EIN 20-8672015, and its entire reported resource envelope. It includes overdose prevention, clinical and mental-health care, syringe services, outreach and basic-needs support. The current service page explicitly includes Bronx outreach alongside the two Manhattan centers. Families or predecessor names do not receive separate overlapping credit. History Services

The ordinary donation route does not promise an exclusive OPC purchase. The model therefore retains all recipient costs and quantifies only a clinically meaningful overdose-survival component. Donate

2. Monitoring and information sharing

The third-year report covers November 30, 2023 to November 29, 2024, despite its 2025 navigation label. It lists 434 purported fatal overdoses prevented; this review treats these as reported intervention events, not independently demonstrated deaths averted. The model assumes that annual event scale remains a usable planning anchor with FY2025 costs; a cumulative 2,000-event milestone in May 2026 is not annualized. Third-year report Current milestone

The original first-two-month evaluation distinguishes naloxone, oxygen, monitoring and stimulant-related supportive interventions, and has no untreated comparison group. These categories have different untreated lethality. Ask for distinct intervention recipients, repeat events, severity, alternative rescue, out-of-site deaths and longitudinal survival. Initial evaluation

3. Qualitative assessment

A Vancouver model estimated 8.4–50.9 potentially averted deaths among 1,004 recorded overdose events, depending strongly on assumed community nonfatal overdose incidence. Of those events, 453 met its potentially fatal definition, based on naloxone, emergency calling or ambulance use. The 8.4–50.9 estimate is modeled deaths across the 1,004-event program, not measured deaths or the fatality rate of its 453 severe events. That historical, pre-fentanyl model is context for our 2% untreated-lethality judgment, not a modern New York causal rate. A further 50% transfer/alternative-rescue factor reflects contemporary community naloxone, emergency response and differences in severity. Original mortality model

A later Vancouver cohort found lower adjusted all-cause mortality among frequent users, but residual confounding and nonrandom selection limit causal transfer. Ontario's controlled interrupted time-series did not identify a statistically significant aggregate mortality improvement. Neither result warrants assigning all city overdose trends to OnPoint. Local difference-in-differences crime research provides reassurance against a simple crime-increase claim, not QALYs or proof that every possible adverse effect is absent. Cohort Ontario study NYC crime study

4. What do you get for your dollar?

The conditional best estimate is $38.5 million per 10 official-MSA QALYs: about 4.204 QALYs against $16,203,740 annual recipient cost. This is an overdose-survival price with other health pathways left unpriced, not a measured whole-portfolio return or a proven lower bound.

The 434 annual intervention events are assigned to 217 distinct people by judgment, or two events each. A 2% lethal-risk input and 50% contemporary transfer factor imply a 0.02 annual hazard reduction during one funded year. Counterfactual mortality hazard is 0.08 and intervention hazard 0.06 during that year. Both groups subsequently face 0.08, so later overdose and competing mortality erode the survival difference. We integrate their survival difference for at most ten years at utility 0.7 and a 3% discount rate. Each person has one survival trajectory: repeated rescue cannot create multiple lifetime awards.

Only 25% of this proportional continuation-scale benefit is attributed to an ordinary gift after public support, reserves and replacement donors; that is an explicit funding-response prior, not an observed budget elasticity. A 98% MSA-resident share is also judgment, acknowledging nonlocal users without assuming city-site location proves residency. The boundary is the official 22-county MSA, not the wider CSA.

Broad alternatives change severity, unique-person clustering, funding response, competing mortality and horizon. Zero and harmful outcomes remain possible. A 10% extra-resource stress tests unrecognized volunteer and partner resources; it is not a measured add-on. Recognized donated medication is already in the financial envelope and is not added twice. Future medical spending and broader social opportunity costs are unpriced.

Model, assumptions and sensitivity

Whole merged recipient annual recognized cost; overdose-survival component only. Unpriced clinical, infection, psychosocial and public effects prevent a whole-portfolio claim.

g=.98 is a subjective share of modeled unique participants who reside in the official22-county MSA. All sites/known outreach are within the MSA but residence records were unavailable.

d=ln(1.03); delta=(E/U)*l*a; lambda1=m-delta>=0; F(k,t)=(1-exp(-k*t))/k. H=U*u*[F(lambda1+d,1)-F(m+d,1)+(exp(-lambda1)-exp(-m))*exp(-d)*F(m+d,T-1)]; Qall=b*H-h; QNYC=g*Qall; Price10=10*C/QNYC if positive. The event-level risk input is converted to an annual hazard decrement by the explicit linear approximation delta=(E/U)*l*a, not asserted to be an exact probability-to-hazard identity.

E
434 reported interventions per historical year (observed). 434 intervention claim in a specified historical 12-month report, transferred to current planning by judgment. [year3]
U
217 distinct intervention recipients assumed (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
l
0.02 fraction as formula defines (judgment). 2% is judgment informed by an old model's 8.4–50.9/1004 range, not modern observed lethality. [fatalityModel]
a
0.5 fraction as formula defines (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
b
0.25 fraction as formula defines (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
m
0.08 annual all-cause hazard (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
u
0.7 fraction as formula defines (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
T
10 maximum follow-up years (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
g
0.98 fraction as formula defines (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]
C
16203740 USD annual recipient cost (observed). Latest original Form 990 and Schedule D agree. [return2025] [schedule2025]
h
0 all-population QALYs lost (judgment). Explicit uncalibrated modeling prior; no claim this is measured OnPoint evidence. [initial] [public26]

Conditional overdose-survival component: Cost: $16.2M; New York City QALYs: 4.204074158539265; all-population QALYs: 4.289871590346189. {"E":434,"U":217,"l":0.02,"a":0.5,"b":0.25,"m":0.08,"u":0.7,"T":10,"g":0.98,"C":16203740,"h":0}

Low lethality and strong replacement: Cost: $16.2M; New York City QALYs: 0.07451143578503404; all-population QALYs: 0.08279048420559337. {"E":434,"U":109,"l":0.005,"a":0.25,"b":0.1,"m":0.12,"u":0.5,"T":5,"g":0.9,"C":16203740,"h":0}

More consequential rescue and longer survival: Cost: $16.2M; New York City QALYs: 50.05576169437563; all-population QALYs: 50.561375448864275. {"E":434,"U":434,"l":0.05,"a":0.75,"b":0.5,"m":0.08,"u":0.8,"T":20,"g":0.99,"C":16203740,"h":0}

Complete funding replacement: Cost: $16.2M; New York City QALYs: 0; all-population QALYs: 0. {"E":434,"U":217,"l":0.02,"a":0.5,"b":0,"m":0.08,"u":0.7,"T":10,"g":0.98,"C":16203740,"h":0}

No survival benefit with induced harm: Cost: $16.2M; New York City QALYs: -0.245; all-population QALYs: -0.25. {"E":434,"U":217,"l":0,"a":0.5,"b":0.25,"m":0.08,"u":0.7,"T":10,"g":0.98,"C":16203740,"h":0.25}

Unrecognized partner-resource stress: Cost: $17.8M; New York City QALYs: 4.204074158539265; all-population QALYs: 4.289871590346189. {"E":434,"U":217,"l":0.02,"a":0.5,"b":0.25,"m":0.08,"u":0.7,"T":10,"g":0.98,"C":17824114,"h":0}

Counterfactual: Public grants, opioid settlement support, ordinary community naloxone/EMS and existing staffing continue. The intervention changes one year's mortality hazard, not all subsequent risk; equal post-year-one hazards impose recurrence and competing mortality.

Attribution: l is a historical-context-informed lethality prior, a a contemporary transfer/rescue factor, b financial additionality. No separate lives-saved multiplier, rescue-lifetime sum, or additional portfolio mortality claim.

Weak-confidence partial-health best estimate. The event count is observed as an organizational claim; causal lethality, unique-person count, hazards, utility and gift response are subjective. The adverse case prevents interpreting omissions as a proven benefit floor.

Sensitivity

  • Unique-person clustering, untreated lethality and funding response dominate the survival estimate.
  • The intervention changes only first-year hazard; all later competing and recurrent mortality returns to the baseline hazard.
  • Zero, adverse and added partner-resource scenarios do not establish a full-portfolio lower bound.

Unresolved inputs

  • Distinct intervention recipients and repeated-event severity, with alternative rescue evidence.
  • Current donation-sensitive capacity beyond public appropriations and other donors.
  • Longitudinal mortality, health utility and MSA residency.
  • Comparable FY2023 merger perimeter and unpriced non-overdose benefits or harms.

5. Funding and previous grants

Original June-ending accrual returns and Schedule D reconciliations report $11,980,572 in FY2023, $15,432,498 in FY2024 and $16,203,740 in FY2025. The FY2023 amended return supersedes its earlier filing and explicitly aligns with completed audited statements. The three returns report no separately netted fundraising, gaming or inventory expense, and no donated-service addback in Schedule D. FY2025 Schedule M reports $336,016 in donated drugs/medical supplies, already recognized as property rather than an additional service addback. We flag FY2023 as noncomparable pending confirmation of the exact merger-period perimeter rather than display an artificial three-year mean. FY2025 return FY2023 amendment

FY2025 revenue was $19,268,609, including $18,181,362 government grants, and ending net assets were $18,520,698; cash was $6,815,648 and grants receivable $9,424,280. Net assets are not all spendable cash. Current city reporting explicitly funds sustained/expanded wraparound hours and services and names OnPoint in infrastructure investments. An ordinary donor cannot receive credit for all these appropriated services. Current settlement report

The official financial directory still emphasizes older documents, so this review uses more recent original IRS filings. No verified current unfunded opening or dollar-specific capacity promise was assumed. Financial directory

Expense appendix: FY2025 Form 990 functional costs are $14,411,565 program services, $1,664,638 management/general and $127,537 fundraising, totaling $16,203,740. The earlier amended original amount is retained, but no comparable three-year average is accepted until the merger-period perimeter is resolved.

Annual expenses

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

  • FY 2023: $12.0M; On Point NYC Inc., EIN 20-8672015, 12-month period, June-ending accrual recognized recipient expense; no reported event/COGS or donated-service reconciliation addback.. Source
  • FY 2024: $15.4M; On Point NYC Inc., EIN 20-8672015, 12-month period, June-ending accrual recognized recipient expense; no reported event/COGS or donated-service reconciliation addback.. Source
  • FY 2025: $16.2M; On Point NYC Inc., EIN 20-8672015, 12-month period, June-ending accrual recognized recipient expense; no reported event/COGS or donated-service reconciliation addback.. Source

6. Sources

  1. FY2023 original Form 990. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  2. FY2023 original Schedule D. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  3. FY2024 original Form 990. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  4. FY2024 original Schedule D. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  5. FY2025 original Form 990. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  6. FY2025 original Schedule D. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  7. FY2025 original noncash property schedule. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  8. FY2023 amendment explanation. On Point NYC / IRS. Published: not stated; retrieved: 2026-09-14.
  9. Merger and recipient history. OnPoint NYC. Published: not stated; retrieved: 2026-09-14.
  10. Current service portfolio and locations. OnPoint NYC. Published: not stated; retrieved: 2026-09-14.
  11. Third-year operations, November 2023–November 2024. OnPoint NYC. Published: not stated; retrieved: 2026-09-14.
  12. 2,000 cumulative interventions milestone. OnPoint NYC. Published: 2026-05-06; retrieved: 2026-09-14.
  13. Funding and original-document directory. OnPoint NYC. Published: not stated; retrieved: 2026-09-14.
  14. Current opioid settlement expenditure and service report. NYC Department of Health and Mental Hygiene. Published: not stated; retrieved: 2026-09-14.
  15. First two months of OPC operation, uncontrolled quality-improvement study. Harocopos et al., JAMA Network Open. Published: not stated; retrieved: 2026-09-14.
  16. Vancouver model of potentially avoided overdose deaths. Milloy et al., PLOS ONE. Published: 2008-10-07; retrieved: 2026-09-14.
  17. Frequent facility use and all-cause mortality cohort. Kennedy et al., PLOS Medicine. Published: 2019-11-26; retrieved: 2026-09-14.
  18. Controlled interrupted time-series of Ontario SCS mortality. Study authors, indexed by PubMed. Published: not stated; retrieved: 2026-09-14.
  19. Difference-in-differences neighborhood crime study. Chalfin et al., JAMA Network Open. Published: not stated; retrieved: 2026-09-14.
  20. Ordinary donation route. OnPoint NYC. Published: not stated; retrieved: 2026-09-14.