GiveBetter x Los Angeles

Didi Hirsch Mental Health Services

Mental health treatment, crisis lines, post-discharge follow-up and substance-use care

Research time: 13 min on GPT-6 Astra Light
  • Research — reviewed programs, finances and impact evidence.

Updated: 2026-09-14

Donate

Summary

What do they do? Didi Hirsch provides outpatient, residential, substance-use and crisis mental-health services, including 988 and Teen Line. Its Los Angeles and Orange County clinical services coexist with national and international crisis-line activity. Its treatment teams combine therapy, medication support and practical help for people managing mental illness.

Why we’re interested in this organization:

  • A current service report identifies a concrete outpatient population and diagnosis mix.

  • The organization delivers clinical therapy, medication support and post-discharge coordination, not only awareness.

  • Original returns and Schedule D reconcile donated services and separately netted event costs.

Our main reservations:

  • The depression illustration omits much of the recipient's potentially important crisis, severe-illness and child-care value.

  • Eligibility, treatment intensity, funding response and LA residence are subjective scenario inputs rather than observed marginal outcomes.

  • Crisis-contact improvement is not a demonstrated mortality effect, and other funded centers may substitute.

What do you get for your dollar? $230.8M per better life: ten additional quality-adjusted life years in Los Angeles. Conditional clinical-component estimate charging full recipient expenses; other program effects are not estimated.

5498 outpatient clients -> 3298.8 adult-compatible client-equivalents -> 1055.616 depression client-equivalents -> 527.808 eligible offers -> 65.976 reference-effect-equivalent offers after transfer and funding -> 3.46854 discounted LA QALYs. These are model expectations, not observed additional patients or completions.

1. What do they do?

This report covers Didi Hirsch Psychiatric Service, EIN 95-1816023, including outpatient and residential treatment, substance-use care, 988, Teen Line, suicide support, training and disaster care. The current treatment page describes individual and group therapy, medication support and case management. It also identifies publicly funded ARMUNITY care and local school partnerships.

The 2025 annual report distinguishes 5,498 outpatient mental-health clients from 246,098 988 contacts. Its adult-services narrative reports depression in 32% of outpatient clients, schizophrenia in 25% and bipolar disorder in 15%. The placement does not establish the exact adult denominator within the 5,498 total; the model therefore makes adult eligibility explicit instead of assigning a depression-trial effect to everyone.

The location directory lists clinics in LA County and Santa Ana, inside this edition. National crisis routing, remote services and Teen Line's international contacts require separate beneficiary attribution. The modeled clinical share of 95% LA/Orange residents is a judgment, not a measured residence fraction.

2. Monitoring and information sharing

The annual report's symptom-improvement percentages are uncontrolled service outcomes. They do not separate natural recovery, concurrent treatment, selective follow-up or the difference made by new funding. Its 6,708 mental-health, substance-use and residential total is not added to the 5,498 outpatient total. Nor are repeated crisis contacts treated as unique patients. 2025 report

A direct UCLA–Didi Hirsch post-discharge study describes outreach within 24–72 hours and identifies referral and communication barriers. It is a qualitative implementation study, not a controlled clinical-effect estimate. The SOSA study is likewise an open-label pre/post evaluation.

The decisive next data are unique adult depression clients, baseline care, clinical eligibility, additional therapy courses enabled by unrestricted funding, treatment completion and follow-up utility. For crisis services, collect unique callers, LA residence, successful follow-up, backup-center response and subsequent outcomes. The new gift's effect is improved care relative to those alternatives, not all care delivered.

3. Qualitative assessment

The CoBalT randomized trial enrolled 469 UK adults with depression despite antidepressants. It compared an offer of 12–18 CBT sessions plus usual care with usual care, yielding an estimated 0.057 additional QALY over 12 months. This is an intention-to-treat offer effect, not an effect per completed session. The economic methods use baseline-adjusted EQ-5D area under the curve; the model does not multiply it by another year or transplant the UK treatment price.

Transport is limited: the trial excluded psychosis, bipolar disorder, major substance misuse and patients already receiving psychotherapy or secondary depression care. Didi Hirsch treats a more complex population. A separate eligibility factor and 50% effect-transfer factor are explicit judgments, not empirically fitted adjustments. The clinical bridge is useful for a subset, but cannot establish a portfolio average.

For suicide prevention, ED-SAFE supports a bundled emergency-department and follow-up intervention, with a roughly five-percentage-point difference in one-year attempt risk. It does not identify deaths prevented by Didi Hirsch alone. Hotline relief, severe-illness care, child development, substance-use treatment and avoided coercive responses may have substantial value, but no unsupported lifetime or per-call credit is added.

4. What do you get for your dollar?

Full recognized recipient resources for FY2025 are $80,043,834: $79,340,574 Form 990 expense, plus $482,866 of donated services excluded from that return total, plus $220,394 of event costs netted against revenue. The corresponding totals are $90,068,699 for FY2024 and $78,394,883 for FY2023; their mean is $82,835,805.33. Original Schedule D reconciles the annual report's $79,823,440 FY2025 expense to Form 990 before restoring event costs. The audit download was blocked; this is a filed reconciliation, not an independently reopened audit. FY2025 return and schedules

Excluding donated services, gross Form 990 expenses are $79,560,968, $89,400,520 and $77,256,644 respectively. These are accrual expenses, not cash flows. No rental, gaming or inventory expense addback was shown in the original returns. FY2024 FY2023

The conditional depression component produces about 3.47 LA QALYs at this full annual cost, or $230.77 million per 10. Positive joint stress tests span roughly $17.13 million to $68.51 billion; they are not confidence limits. This is a conditional clinical-component comparison, not an estimate of the entire portfolio. A whole-portfolio price remains unidentified, not zero.

At $1 million per 10 LA QALYs, the entire recipient would need 800.44 net additional LA QALYs per annual resource equivalent; at $100,000 it would need 8,004.38. The priced subset cannot close that gap under the chosen assumptions. Crisis, severe-illness or other benefits would need independent delivery and outcome models. External medical care and patient time are additional social resources, not negative donor costs.

Model, assumptions and sensitivity

Full annual recipient recognized resources, including donated services and restored event costs. The health numerator is only a depression-treatment component, not the complete portfolio. No program-only denominator or public-contract subtraction.

g=.95 is a judgment for clinical outpatient residents in Los Angeles and Orange Counties, supported by local sites but not observed patient residence. Do not apply it to national 988, Teen Line or other remote activity.

Q_component_LA = N*s*p*e*q*t*b*g/(1+d)-H. N=5498 annual outpatient clients; s=.6 adult-compatible share; p=.32 depression share; e=.5 clinical offer eligibility; q=.057 external 12-month intention-to-treat QALY effect; t=.5 clinical transfer; b=.25 proportional funding responsiveness; g=.95 LA residence; d=.03 next-year delivery discount; H=0 in the conditional positive case. Q=3.46854407767. Price10=10*C/Q, C=80043834. Complete portfolio Q = Q_component_LA + Q_other_LA; Q_other_LA remains unknown, not zero.

C
80043834 USD annual full recognized recipient resources (observed). FY2025 Part IX 79,340,574 + donated services 482,866 + event costs 220,394. [99025] [schedule25]
N
5498 reported annual outpatient clients (observed). 2025 annual report, printed page 9; not crisis contacts. [annual25]
s
0.6 adult-compatible share of outpatient total (judgment). Adult denominator not separately supplied. .6 central, .3–.8 joint stress tests; not an observed age distribution. [annual25]
p
0.32 depression diagnosis share (observed). Reported in adult outpatient narrative; transport to eligible count conditional on s. [annual25]
e
0.5 eligible depression-treatment offer fraction (judgment). Excludes clients unlike CoBalT, including major psychosis/bipolar/substance-use overlap and already-equivalent therapy; test .1–.8. [cobalt]
q
0.057 additional QALYs per randomized CBT offer over 12 months (observed). External CoBalT intention-to-treat estimate; not per completed course or per year indefinitely. [cobalt-economic]
t
0.5 clinical effect transfer fraction (judgment). Different intensity and clinical system; .1–1 joint tests. Not a local empirical rate. [treatment] [cobalt]
b
0.25 proportional funding responsiveness (judgment). New ordinary support changes one-quarter of proportional clinical delivery after replacement funding; .05–.75 stress tests. [99025] [case-support]
g
0.95 LA+Orange resident share of modeled clinical component (judgment). Local sites support concentration; residence not observed. .8–1 stress tests. [locations]
d
0.03 one-year delivery discount (judgment). q already measures one-year health area; discount once for future delivery, not again for duration.
H
0 component harm QALYs in positive illustration (judgment). Zero only in positive illustrative scenarios; negative stress test retained.
Q_other_LA
null additional non-overlapping LA portfolio QALYs (unknown). 988, severe mental illness, youth, residential and substance-use benefits and harms require separate models.

Conditional depression component, not whole portfolio: Cost: $80.0M; Los Angeles QALYs: 3.468544077669903; all-population QALYs: 3.6510990291262138. N=5498,s=.6,p=.32,e=.5,q=.057,t=.5,b=.25,g=.95,d=.03,H=0. A continuation-scale proportional funding model, not a claim that all existing patients depend on a new gift.

Weak component delivery and funding response: Cost: $80.0M; Los Angeles QALYs: 0.011683516893203884; all-population QALYs: 0.014604396116504853. s=.3,e=.1,t=.1,b=.05,g=.8; other positive-case inputs unchanged.

Favorable component assumptions: Cost: $80.0M; Los Angeles QALYs: 46.73406757281555; all-population QALYs: 46.73406757281555. s=.8,e=.8,t=1,b=.75,g=1; not a confidence bound.

Component excluding donated-service valuation: Cost: $79.6M; Los Angeles QALYs: 3.468544077669903; all-population QALYs: 3.6510990291262138. Same clinical assumptions; use gross Form 990 accrual expense, not cash flows.

No additional component health: Cost: $80.0M; Los Angeles QALYs: 0; all-population QALYs: 0. b=0 or t=0,H=0; other portfolio value is not implied zero.

Illustrative adverse component: Cost: $80.0M; Los Angeles QALYs: -1; all-population QALYs: -1. No benefit and H=1 LA QALY; magnitude is a stress test, not an empirical harm estimate.

Whole-portfolio health remains unestimated: Cost: $80.0M; Los Angeles QALYs: unknown; all-population QALYs: unknown. Q_other_LA is unknown; do not substitute the component price for a full-recipient ranking.

Counterfactual: Usual medication and clinical care continue; public contracts, alternative providers and reserves replace some capacity. CoBalT already compares with usual care, so no extra generic usual-care discount is applied. b only concerns whether added recipient funding changes clinical capacity.

Attribution: The clinical transfer t covers differences in delivered therapy, setting and case mix after eligibility e. No separate coalition multiplier is added. Count each client once in this component, and do not add overlapping symptom gains from crisis or substance-use services. q is per offer, so trial noncompletion must not be discounted a second time.

This is a conditional component illustration, not an expected whole-portfolio return, lower bound or confidence interval. Adult share, eligibility, transfer, funding and residence are explicit subjective inputs. Major other pathways are unpriced; zero or adverse net effect is possible.

Sensitivity

  • Central component price is $230,770,698.62 per 10 LA QALYs; favorable $17,127,512.79; weak $68,510,051,152.97. Joint cases are subjective stress tests.
  • At the full-recipient $1 million/10 target, required annual LA health is 800.43834 QALYs; $100,000/10 requires 8004.3834.
  • Halving the adult share, clinical eligibility, transfer or funding responsiveness doubles the positive component price.
  • The 0.057 QALY includes the trial's entire 12-month utility trajectory. No lifetime extrapolation, per-session multiplier or duplicated completion discount is applied.
  • Applying the depression effect to schizophrenia, bipolar disorder, all children or crisis callers would materially overstate the evidence.
  • Other portfolio value may dominate the depression component; omission is not proof of low overall effectiveness.

Unresolved inputs

  • Unique adult depression denominator and clinical eligibility, including existing psychotherapy and comorbidity.
  • Delivered treatment intensity, local utility trajectories and causal comparison with current alternatives.
  • Current unrestricted funding gap after the June 2026 fundraising deadline and public renewals.
  • Patient residence and outcome overlap across clinics, crisis services and substance-use treatment.
  • Independent cost-to-clinical-outcome models for the unpriced majority of services.

5. Funding and previous grants

FY2025 Form 990 revenue includes $67,042,157 in government grants and $8,829,632 in contract fees. Year-end net assets were $40,203,725, including $3,180,468 restricted; cash was $7,735,622. These figures do not identify freely available cash today or a current unrestricted service gap. Original return

The case for support bears an August 8, 2025 internal date and seeks $4.8 million by June 30, 2026. That deadline has passed; the document does not verify an unfilled gap today. Its service-growth projections and broad economic-return claims are not causal QALY estimates. The annual report identifies specific existing disaster-care grants, including a two-year $1 million Red Cross award, which belong in the baseline.

The model's funding responsiveness of 25% means a proportional increment of recipient resources is assumed to produce one-quarter of the corresponding clinical activity change after public reimbursement, reserves and replacement donors. It is an explicit prior, not a measured funding elasticity. A documented uncovered clinician position or continuity-of-care commitment could substantially change it. Public crisis appropriations and national backup capacity also make 'otherwise no answer' a weak default counterfactual.

Annual expenses

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

  • FY 2023: $78.4M; Didi Hirsch Psychiatric Service, EIN 95-1816023, 12-month period, June-ending accrual full recognized recipient expense: Form 990 Part IX plus Schedule D donated services and Part VIII direct event costs.. Source
  • FY 2024: $90.1M; Didi Hirsch Psychiatric Service, EIN 95-1816023, 12-month period, June-ending accrual full recognized recipient expense: Form 990 Part IX plus Schedule D donated services and Part VIII direct event costs.. Source
  • FY 2025: $80.0M; Didi Hirsch Psychiatric Service, EIN 95-1816023, 12-month period, June-ending accrual full recognized recipient expense: Form 990 Part IX plus Schedule D donated services and Part VIII direct event costs.. Source

6. Sources

  1. 2025 Annual Report, including current services and financial summary. Didi Hirsch. Published: not stated; retrieved: 2026-09-13.
  2. Original Form 990 for fiscal year ending June 2025. Didi Hirsch / IRS via ProPublica. Published: 2026-03-30; retrieved: 2026-09-13.
  3. Original Form 990 for fiscal year ending June 2024 with 2023 comparative totals. Didi Hirsch / IRS via ProPublica. Published: 2025-03-24; retrieved: 2026-09-13.
  4. Qualitative evaluation of UCLA–Didi Hirsch post-discharge follow-up partnership. Soderlund et al., BMC Health Services Research. Published: 2023-11-17; retrieved: 2026-09-13.
  5. Open-label evaluation of Survivors of Suicide Attempts groups. Hom et al., Psychological Services. Published: not stated; retrieved: 2026-09-13.
  6. ED-SAFE sequential-phase suicide prevention trial. Miller et al., JAMA Psychiatry. Published: 2017-06-01; retrieved: 2026-09-13.
  7. Original FY2023 Form 990 and Schedule D. Didi Hirsch / IRS via ProPublica. Published: 2024-05-15; retrieved: 2026-09-14.
  8. Original FY2025 Schedule D expense reconciliation. Didi Hirsch / IRS via ProPublica. Published: 2026-03-30; retrieved: 2026-09-14.
  9. Original FY2024 Schedule D expense reconciliation. Didi Hirsch / IRS via ProPublica. Published: 2025-03-24; retrieved: 2026-09-14.
  10. Original FY2023 Schedule D expense reconciliation. Didi Hirsch / IRS via ProPublica. Published: 2024-05-15; retrieved: 2026-09-14.
  11. Current treatment portfolio. Didi Hirsch. Published: not stated; retrieved: 2026-09-14.
  12. Current clinic geography. Didi Hirsch. Published: not stated; retrieved: 2026-09-14.
  13. CoBalT original randomized trial. Wiles et al.; NIHR Health Technology Assessment. Published: not stated; retrieved: 2026-09-14.
  14. CoBalT economic evaluation methods and results. Wiles et al.; NIHR Journals Library. Published: not stated; retrieved: 2026-09-14.
  15. Case for support; internal date August 8, 2025. Didi Hirsch. Published: not stated; retrieved: 2026-09-14.
  16. 2026–27 clinical internship handbook. Didi Hirsch. Published: not stated; retrieved: 2026-09-14.