GiveBetter x California

End Overdose

Overdose response training, naloxone access and community outreach

Research time: ~20 min on GPT-6 Astra Light
  • Research — reviewed programs, finances and impact evidence.
  • Source audit — checked claims, assumptions and calculations.

Updated: 2026-09-13

Donate

Summary

What do they do? End Overdose trains people to recognize and respond to overdoses and distributes naloxone through online and community channels. Its national operations include California chapters and a California-resident kit offer. Participants learn overdose recognition and response before requesting medication.

Why we’re interested in this organization:

  • Timely naloxone can avert fatal opioid overdoses when no adequate alternative response is available.

  • California-specific fulfillment and local chapters establish a real in-state delivery pathway.

  • Current campaign reporting distinguishes donated medication from cash-supported training and operations.

Our main reservations:

  • Chapter counts and mailing eligibility do not establish California's share of additional health.

  • Free public supply, other rescuers and later overdose risks substantially reduce attribution relative to doses distributed.

  • National conditional estimates rely on wide judgments about risk, additional timely rescue and gift-sensitive expansion.

What do you get for your dollar? $13.0M per better life: ten additional quality-adjusted life years in California. Naloxone benefits only; California share modeled.

Additional two-dose kit opportunities; unique otherwise unprotected overdose situations; additional timely rescues; unique-person survival and quality-adjusted life-years, not gross doses or training completions. New geographic bridge distinguishes venue/delivery outputs from where additional patient health occurs; no observed dose or chapter is mechanically converted to a California QALY.

1. What do they do?

End Overdose combines overdose education, naloxone access, test strips and youth-oriented outreach. Its chapter directory includes California communities and campuses, while the California kit page offers a two-dose intranasal kit after training, with shipping charged. The page displayed 'Sold out' when reviewed; its stated shipping delay is not proof that more donations would resolve the constraint.

2. Monitoring and information sharing

The July–September 2024 impact report reports aggregate training and distribution. These are outputs, not independently measured unique lives saved. The August 2026 campaign is newer and documents national chapter growth and operations beyond California. No state-by-state unique-recipient, additional rescue or funding allocation dataset was located. A useful evaluation would link kit exposure, actual emergency use, alternative rescue availability and unique-person follow-up without counting repeat reversals as repeated lifetimes.

A focused geographic audit counted 16 California cards among 62 in the current directory, which is not reconciled with the campaign's 80-plus chapter claim. Four named California event highlights in the 2024 report sum to 5,280 of 71,733 reported quarterly doses (7.36%). This is a subset of historical delivery, not a floor for current California health, and the venue does not establish patient residence. These observations inform a weak scenario only; no chapter or dose ratio is mechanically substituted for QALYs.

3. Qualitative assessment

My central judgment is that broad-reach delivery probably has positive health value, but an ordinary gift is not yet a compelling high-confidence health bargain. Walley's community study associated greater overdose-education/naloxone implementation with lower mortality; it was not randomized and did not evaluate End Overdose. Zang's 2024 Rhode Island simulation highlights targeting and whether somebody actually witnesses an overdose. It is a model, not an observed charity effect. Importantly, the HEALing Communities cluster trial found no statistically significant primary mortality reduction for a broader intervention bundle: rate ratio 0.91, 95% CI 0.76–1.09. This does not negate naloxone's clinical action, but warns that increasing activities does not guarantee population benefit.

The model therefore discounts broad-audience exposure, unusable or unavailable kits, other rescuers, repeated scenes and replacement funding. Nonfatal morbidity and test-strip effects are omitted rather than invented.

A 2025 festival survey found 55.5% of 227 respondents already carried naloxone in some context and 8.4% had ever used it on someone. This small, selected cross-section cannot supply an annual End Overdose rescue rate; it supports both relevance and substantial existing protection. In a more targeted Rikers visitor cohort, 40 of 226 follow-up respondents witnessed an overdose in six months. That high-risk result is not transferable wholesale to students or festival audiences.

For California, residence at delivery is insufficient if rescue or subsequent health occurs elsewhere. California chapter share is not an outcome-weighted marginal allocation.

4. What do you get for your dollar?

Our revised conditional scenario is about $13 million per ten California QALYs. It retains the national model's naloxone-mortality calculation and all recipient costs, then makes the geographic assumptions explicit. This is a weak, judgment-based allocation—not a measured California return or a comprehensive estimate of every End Overdose program.

The central geographic prior assigns 15% of marginal exposure/person-time to California before allowing for relative effectiveness. It then assumes California's incremental yield is 75% of that elsewhere because existing supply and alternative rescue may substitute more; normalizing those weights assigns about 11.7% of national modeled health to California. Neither fraction is observed. The original 2024 report, current chapters, a California kit offer and event footprint support meaningful California presence, but not an exact health share. National online delivery and out-of-state expansion prevent assuming that an ordinary gift stays here.

Holding clinical assumptions fixed, a 1% California health share gives roughly $152 million per ten QALYs, while 50% gives $3.0 million. Combining geographic and clinical stress assumptions expands the range from roughly $17,000 to $1.46 trillion; these are not confidence bounds or donation offers. A gift could produce national benefit and no additional California benefit. The direct-delivery mechanism supports an inspectable scenario, but a state-by-channel marginal delivery and unique-patient dataset is needed to replace these priors.

The 2024 Form 990 records $970,440 total expenses: $730,276 program, $207,274 management and $32,890 fundraising in Part IX. Its Part III program total is inconsistent with those figures, so it is not used to allocate the model. Dividing total spending by reported trainees gives roughly $4.20 per training, but does not price a completed training-plus-kit package or an extra rescue.

Model, assumptions and sensitivity

Full ordinary recipient gift across all functions. National model conservatively quantifies the naloxone pathway while costing unquantified activities; C=100000 is normalization only.

An explicit subjective geographic allocation replaces the null: p=0.15 is the share of marginal exposure/person-time provisionally assigned to California before relative yield; rho=0.75 is California's incremental health yield relative to elsewhere. g_CA=p*rho/(1-p+p*rho)=0.1168831169. p accounts for patient location and later mobility once; rho accounts for relative incremental yield after local alternatives. Neither is an observed resident, chapter or dose share. Normalization calibrates the two regional yields to the fixed national mean: it reallocates, rather than reduces, national benefit. CA and USA totals must not be added. The unchanged all-population health total is divided between CA and elsewhere consistently. This is a weak prior, not a geographic finding.

Offers=G*a/c. Unique additional deaths prevented=Offers*e*d*r*f*b. L=sum(k=1..T)[u*(1-m)^k/(1.03)^k]. All-populationQALYs=deaths*L/1.03 for rescue at end ofyear1. CaliforniaQALYs=allQALYs*g_CA. DonorUSD/10CaliforniaQALYs=10G/CaliforniaQALYs. Gross resources=G+Offers*(unpaid labor+donated stock value). Central parameters: a=.75,c=50,e=.005,d=.5,r=.2,f=.2,b=.5,u=.75,m=.04,T=30. Geographic extension only: g_CA=p*rho/(1-p+p*rho), Q_CA=Q_all*g_CA. Central p=.15,rho=.75; favorable p=.5,rho=1.5; pessimistic p=.02,rho=.25. National clinical parameters and Q_all remain unchanged. p is a modeled marginal exposure/person-time allocation, not a chapter-count ratio; rho reallocates relative local effectiveness without applying a second blanket discount to national benefit.

FY 2024 recipient spending
970440 USD/year (observed). PartI andIX total agree; PartIII allocation does not and is not used. [eo-99024]
Naloxone package content
2 doses/package (observed). Current recipient-facing kit description. Quarterly dose counts are not package counts. [eo-kit]
2026 manufacturer contribution
25000 reported naloxone units (observed). Campaign statement; unit definition, remaining stock and recurring availability not verified. [eo-campaign]
CalRx procurement benchmark
19 USD/twin-pack (observed). California2026 price; not an EO invoice or automatically available national price. Central$50 cash package can be read as$19 stock plus$31 implementation, both transfer assumptions. [calrx-price]
a
0.75 fraction (judgment). Gift share supporting costed training-plus-kit pathway; other work receives zero quantified benefit. [eo-campaign]
c
50 USD/offered package (judgment). All-in marginal cash delivery assumption, not historical average cost per training.
e
0.005 fraction (judgment). 0.5% annual opportunity is a deliberately cautious broad-audience prior, not derived by annualizing lifetime festival reports. High-risk jail visitors witnessed far more events; actual EO channel mix remains unknown. [naloxone-rikers] [naloxone-festival]
d
0.5 fraction (judgment). Conservative duplicate-scene/person adjustment; needs unique-recipient follow-up.
r
0.2 fraction (judgment). Incremental timely rescue after completion, availability and existing rescuers; do not add another adherence multiplier. [naloxone-festival] [naloxone-zang]
f
0.2 fraction (judgment). Absolute episode mortality reduction from extra timely rescue; not survival after reported reversal.
b
0.5 fraction (judgment). Only half nominal delivery assumed additional to other funding. Current campaign supports plausibility, not proof. [eo-campaign]
u
0.75 fraction (judgment). Utility while alive; not measured in EO recipients.
m
0.04 fraction (judgment). Annual competing mortality, capturing continued overdose and other risks.
T
30 years (judgment). Finite modeled health horizon, not measured life expectancy.
unpaid
10 USD/offered package (judgment). Volunteer/support resource envelope, not marginal opportunity cost.
p_CA
0.15 modeled marginal exposure/person-time share before local yield (judgment). A cautious 15% prior, not derived by equating outputs to health. Named CA events account for 7.36% of all reported summer-2024 doses, but highlights omit other events/channels; current directory lists 16 of62 CA chapters, while newer campaign says80+ nationally. CA events and separate kit offer support nontrivial presence; national online delivery and expansion justify not assuming a dominant CA share. Future patient residence and marginal allocation are unmeasured. Stress2%–50% is subjective, not a bound. [impact24] [chapters] [campaign] [ca-kit] [eo-events-2026]
rho_CA
0.75 California incremental health yield relative to non-California (judgment). 25% lower local incremental yield is a cautious relative-effect prior for public naloxone availability and other rescue, not a measured state comparison; targeted gaps might instead make CA more productive. Stress0.25–1.5. National average clinical total remains unchanged rather than being discounted twice. [dhcs] [ca-kit] [eo-events-2026]
g_CA
0.11688311688311688 California fraction of unchanged national modeled QALYs (judgment). Derived .15*.75/(.85+.15*.75)=.1168831169; not a measured allocation. Normalization assigns complementary modeled benefit outside California. [chapters] [campaign] [impact24]

Conditional central — subjective California share: Cost: $100K; California QALYs: 0.0769425918759858; all-population QALYs: 0.6582866193834341. {"a":0.75,"c":50,"e":0.005,"d":0.5,"r":0.2,"f":0.2,"b":0.5,"u":0.75,"m":0.04,"T":30,"g_CA":0.11688311688311688,"unpaid":10}; calculated offers=1500; unique deaths=0.07500000000000001; finiteQALYs/death=9.040469572865828; gross resourceUSD=115000 Geographic assumptions p=0.15, rho=0.75, g_CA=0.11688311688311688. All-population QALYs unchanged; these geography values are judgments, not measured shares.

Favorable clinical and geographic stress: Cost: $100K; California QALYs: 58.604111277072015; all-population QALYs: 97.67351879512003. {"a":0.85,"c":35,"e":0.05,"d":0.8,"r":0.3,"f":0.3,"b":0.8,"u":0.85,"m":0.02,"T":40,"g_CA":0.6,"unpaid":5}; calculated offers=2428.5714285714284; unique deaths=6.9942857142857155; finiteQALYs/death=14.383702420604887; gross resourceUSD=112142.85714285714 Geographic assumptions p=0.5, rho=1.5, g_CA=0.6. All-population QALYs unchanged; these geography values are judgments, not measured shares.

Adverse clinical and geographic stress: Cost: $100K; California QALYs: 6.833939493325894e-7; all-population QALYs: 0.00013462860801852008. {"a":0.5,"c":100,"e":0.001,"d":0.25,"r":0.05,"f":0.05,"b":0.1,"u":0.65,"m":0.08,"T":15,"g_CA":0.005076142131979696,"unpaid":25}; calculated offers=500; unique deaths=0.00003125000000000001; finiteQALYs/death=4.4373589202904204; gross resourceUSD=112500 Geographic assumptions p=0.02, rho=0.25, g_CA=0.005076142131979696. All-population QALYs unchanged; these geography values are judgments, not measured shares.

Donated stock; central subjective California share: Cost: $100K; California QALYs: 0.1241009546386868; all-population QALYs: 1.0617526119087648. {"a":0.75,"c":31,"e":0.005,"d":0.5,"r":0.2,"f":0.2,"b":0.5,"u":0.75,"m":0.04,"T":30,"g_CA":0.11688311688311688,"unpaid":10,"stockResource":19}; calculated offers=2419.3548387096776; unique deaths=0.1209677419354839; finiteQALYs/death=9.040469572865828; gross resourceUSD=170161.29032258067 Geographic assumptions p=0.15, rho=0.75, g_CA=0.11688311688311688. All-population QALYs unchanged; these geography values are judgments, not measured shares.

No additional national rescue: Cost: $100K; California QALYs: 0; all-population QALYs: 0. Setb=0orr=0; no finite favorable health price, regardless of observed total reach. Geographic assumptions p=0.15, rho=0.75, g_CA=0.11688311688311688. All-population QALYs unchanged; these geography values are judgments, not measured shares.

National benefit, no California marginal benefit: Cost: $100K; California QALYs: 0; all-population QALYs: 0.6582866193834341. Hold central national clinical yield; set p=0 if marginal gift delivery and subsequent patient health are entirely outside California. No finite California price.

Central clinical yield; 1% California share: Cost: $100K; California QALYs: 0.0065828661938343415; all-population QALYs: 0.6582866193834341. Direct g_CA=0.01 sensitivity only; not a measured or probabilistic bound. All clinical assumptions held central.

Central clinical yield; 50% California share: Cost: $100K; California QALYs: 0.32914330969171707; all-population QALYs: 0.6582866193834341. Direct g_CA=0.5 sensitivity only; not a measured or probabilistic bound. All clinical assumptions held central.

Counterfactual: Other free distribution, public stock, purchase and emergency responders remain available; only changes in timely rescue and subsequent survival beyond those alternatives count. Ordinary gifts may replace funding or support national expansion outside California. Geographic p may be zero if the marginal gift supports only out-of-state work. rho distinguishes local alternatives from the already included nationwide average rescue and budget additionality; it is not another medication-stock replacement factor.

Attribution: Funding additionality is distinct from incremental rescue; donated stock is a complement with no assumed cash purchase cost. Public/partner contributions and repeated rescues cannot each receive full causal health credit.

The original California share remains empirically unidentified. This version provides a deliberately subjective conditional central rather than pretending lack of measurement requires no scenario. Historic event doses, current California chapters and a CA-specific offer establish substantial delivery presence, but no calibrated distribution for p or rho. The central is not a statistical expectation, and finite stress scenarios are not bounds. A California-only null can coexist with positive national health. This quantifies the naloxone mortality pathway while retaining all recipient costs; unmodeled morbidity/test-strip/education effects are not demonstrated zero.

Sensitivity

  • Conditional central Q_CA=0.0769425918759858 per $100,000 gives $12996702.80943714 per ten California QALYs. This is the naloxone mortality pathway with all recipient costs, not a comprehensive portfolio estimate.
  • Geography alone, holding clinical yield central, gives roughly $3.04 million at g=50% and $151.91 million at g=1%. No evidence establishes either as a bound.
  • At p=.15, changing rho from .75 to1 removes the relative CA-yield adjustment: g=.15, price about $10.13 million. Public alternatives do not automatically erase the value of last-mile outreach.
  • Combined favorable and pessimistic clinical/geographic cases inherit the national stress assumptions and should not be advertised as alternative donation offers. Zero California benefit is possible even with positive national benefit.
  • With all clinical inputs central, even g=1 costs about $1.52 million per ten QALYs. Geographic concentration alone cannot reach the $1 million benchmark; better targeting or delivery yield is required.
  • Venue, shipping address, chapter count and eventual California patient health are distinct. A state-by-channel fulfilled-kit and unique-patient follow-up dataset should replace p/rho before the point estimate is treated as decision-grade.

Unresolved inputs

  • Outcome-weighted California marginal share by channel and recipient, with migration/health-location treatment.
  • Current unrestricted funding allocation and inventory/fulfillment bottleneck, distinguishing donated units and two-dose kits.
  • Unique-person exposure, alternative rescue and finite follow-up; repeat rescue deduplication.
  • California-specific substitution relative to state, pharmacy, community and emergency-responder supply.
  • Channel shares of marginal delivery, geographic exposure and relative state-level incremental rescue; no weights were observed.
  • No probability distribution supports a statistical best estimate for geographic allocation. This model is a disclosed weak prior, to be updated rather than presented as measurement.

5. Funding and previous grants

The August 2026 campaign reports 25,000 naloxone units donated by Amneal, alongside training, chapter growth and state operations. Units are not explicitly two-dose kits, and remaining inventory is unknown; a gift must not automatically be modeled as purchasing medication. California DHCS offers naloxone to eligible entities and identifies other free community sources and a $19 consumer two-pack alternative. Individuals cannot apply directly to NDP, and NDP stock must remain in-state. We did not verify that End Overdose receives NDP stock. Public supply can complement outreach while also making replacement of another source more likely.

The original 2024 filing reports $20,042 cash and negative $16,572 net assets; these historical figures do not establish a current cash emergency. The independently checked 2022–24 expense series averages $809,773.67.

Annual expenses

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

  • FY 2022: $516K; End Overdose, EIN83-0696963, 12-month period, Form990 totalexpenses, calendar year; not donated-care sticker value. Source
  • FY 2023: $943K; End Overdose, EIN83-0696963, 12-month period, Form990 totalexpenses, calendar year; not donated-care sticker value. Source
  • FY 2024: $970K; End Overdose, EIN83-0696963, 12-month period, Form990 totalexpenses, calendar year; not donated-care sticker value. Source

6. Sources

  1. Our Solution. End Overdose. Published: not stated; retrieved: 2026-09-13.
  2. Chapter directory. End Overdose. Published: not stated; retrieved: 2026-09-13.
  3. Naloxone Kit (CA residents). End Overdose. Published: not stated; retrieved: 2026-09-13.
  4. Overdose Awareness Month August 2026. End Overdose. Published: not stated; retrieved: 2026-09-13.
  5. July–September 2024 impact report. End Overdose. Published: not stated; retrieved: 2026-09-13.
  6. Naloxone Distribution Project Frequently Asked Questions. California Department of Health Care Services. Published: not stated; retrieved: 2026-09-13.
  7. 2024 Form990, PartsI/III/IX/X. End Overdose / IRS, reproduced by ProPublica. Published: not stated; retrieved: 2026-09-13.
  8. 2023 Form990, current and prior-year expenses. End Overdose / IRS, reproduced by ProPublica. Published: not stated; retrieved: 2026-09-13.
  9. Our Solution. End Overdose. Published: not stated; retrieved: 2026-09-13.
  10. Summer2024 Impact Report, July–September. End Overdose. Published: 2024-10-01; retrieved: 2026-09-13.
  11. August 2026 Overdose Awareness Month campaign. End Overdose. Published: not stated; retrieved: 2026-09-13.
  12. Naloxone Kit. End Overdose. Published: not stated; retrieved: 2026-09-13.
  13. Official donation route linked from current impact page. End Overdose. Published: not stated; retrieved: 2026-09-13.
  14. Overdose education/naloxone implementation: interrupted time-series study. Walley et al. / BMJ. Published: 2013-01-30; retrieved: 2026-09-13.
  15. Strategies to Enhance Community-Based Naloxone Distribution. Zang et al. / JAMA Network Open. Published: 2024-05-30; retrieved: 2026-09-13.
  16. Community-Based Cluster-Randomized Trial to Reduce Opioid Overdose Deaths. HEALing Communities Study Consortium / NEJM. Published: 2024-06-16; retrieved: 2026-09-13.
  17. Naloxone Savings, price effective January2026. California CalRx. Published: not stated; retrieved: 2026-09-13.
  18. Naloxone Distribution Project FAQ, April2026 product update. California Department of Health Care Services. Published: not stated; retrieved: 2026-09-13.
  19. Naloxone Use, Carrying Practices, Prior Training, and Confidence at a Colorado Music Festival. Piercey, Pince and Karoly / Journal of Addiction Medicine. Published: 2025-12-23; retrieved: 2026-09-13.
  20. Witnessed overdoses and naloxone use among trained Rikers Island visitors. Huxley-Reicher et al. / Addictive Behaviors. Published: 2017-11-16; retrieved: 2026-09-13.
  21. Current event directory, first page. End Overdose. Published: not stated; retrieved: 2026-09-13.