The Headstrong Project

Ordinary unrestricted recipient donation, with all accounting overhead included in session cost; no cherry-picked trauma modality. G=$10,000 is arithmetic normalization only. Donor cost is gross, with no fee-for-service leverage or avoided medical-cost deduction assumed. Paid clinicians already enter the cost. Patient time, extended-care copays and incremental downstream care are excluded from donor cost and tested separately where possible.

Reported domestic in-person network and all-50-state telehealth support predominantly 50-states/DC resident benefits. Central g=.99, favorable1 and pessimistic.95 are residence-share judgments allowing overseas/territorial service-member spillover; client military status is not residence. A verified residence extract is missing.

Episode-equivalents = G/(c×n). q = q1 × sum[t=1..T](p^(t−1)/(1.03^t)). Q_all = G/(c×n) × q × f × b × s. Q_USA = Q_all×g. Price per10 =10G/Q_USA. q1 is a net intention-to-treat first-year increment, already including trial attrition/treatment burden; no additional completer multiplier is applied.

G

10000 USD normalization (judgment). Calculation unit, not a quoted funding opportunity.

2025 expenses

13963947 USD (observed). Form990 total expense; annual report gives14029100, not silently substituted. [hs-99025] [hs-ar25]

2025 sessions

52000 sessions (observed). Rounded organizational report; more than52000 in990, use conservative52000 denominator. [hs-ar25] [hs-99025]

2025 clients

2947 clients/year (observed). Annual report page6; not completed course count. [hs-ar25]

c

275 USD/session including overhead (judgment). Rounded above13963947/52000=$268.54; sensitivity200–350 reflects capacity and delivery cost. [hs-99025] [hs-ar25]

n

30 sessions/entry-equivalent (judgment). Current first30 free and historical mean32/median20 inform full-episode prior; annual sessions/client cannot identify episode length. Sensitivity15–45. [hs-care] [hs-program-paper]

q1

0.05 net first-year QALY/therapy offer (judgment). Close to external randomized0.056 PE-vs-sertraline estimate, not copied as Headstrong efficacy; trial mostly civilian women, benefits versus no adequate care may differ. Net intention-to-treat including dropout. [ptsd-utility-trial] [ptsd-veteran-rct] [ptsd-waitlist-rct]

p

0.8 annual persistence factor (judgment). 20% annual attenuation from relapse and comparator catch-up; external economic model uses expert relapse assumptions, not measured Headstrong persistence. [ptsd-model]

T

3 years (judgment). Finite horizon; favorable5, pessimistic1. No lifetime symptom relief.

f

0.8 clinical/population transfer factor (judgment). Further adjustment for military population, treatment mix/fidelity and family services; not an observed allocation. [hs-care] [hs-ar25] [ptsd-utility-trial]

b

0.75 funding/capacity additionality (judgment). Expansion and demand support positive contribution, but reserves/replacement donors limit certainty. [hs-ar25] [hs-audit24] [hs-99025]

s

0.6 counterfactual-care health additionality (judgment). Substantial need despite VA/insurance access, without treating all alternative care as absent. [hs-nyhealth] [hs-care]

g

0.99 USA resident share (judgment). Domestic network implies mostly in-boundary; residence data not published. [hs-ar25]

discount

0.03 annual (judgment). Analyst convention; all first-year gains discounted to year end.

2023 audited expense comparator

11640268 USD (observed). Audit differs from Form99011795084; retained as reconciliation rather than mixed in990 series. [hs-audit24] [hs-99024]

Patient resource increment

1000 USD/episode-equivalent (judgment). Illustrative travel/time/copay cost addition; actual net downstream costs unknown.

Cautious treatment-access judgment: Cost: $10K; USA QALYs: 0.04990984939513714; all-population QALYs: 0.05041398928801731. c=275; n=30; q1=0.05; p=0.8; T=3; f=0.8; b=0.75; s=0.6; g=0.99; q=0.11553205878503965. Judgment scenario, not measured donor productivity.

Efficient treatment reaching otherwise untreated clients: Cost: $10K; USA QALYs: 1.2262179333194803; all-population QALYs: 1.2262179333194803. c=200; n=15; q1=0.12; p=0.9; T=5; f=0.95; b=0.95; s=0.9; g=1; q=0.45289674360830307. Judgment scenario, not measured donor productivity.

Long courses with limited additional access: Cost: $10K; USA QALYs: 0.00021081830790568656; all-population QALYs: 0.00022191400832177532. c=350; n=45; q1=0.01; p=0.5; T=1; f=0.6; b=0.3; s=0.2; g=0.95; q=0.00970873786407767. Judgment scenario, not measured donor productivity.

Equivalent treatment financed elsewhere: Cost: $10K; USA QALYs: 0; all-population QALYs: 0. b=0 or s=0; no finite favorable price.

Central with illustrative patient-resource cost: Cost: $11K; USA QALYs: 0.04990984939513714; all-population QALYs: 0.05041398928801731. Adds $1,000 per nominal entry-equivalent for time/travel/copays; not a measured societal ICER and no avoided costs credited.

Counterfactual: q1 compares an effective trauma-treatment offer with insufficient/less-effective care, not with perfect health. s discounts for an equivalent health trajectory through VA, military, insurer or other charity treatment; delayed treatment can yield partial additionality. b separately discounts for reserves, replacement funding and capacity not expanded by a gift. Annual persistence p includes relapse and untreated/comparator catch-up. No suicide mortality reduction is inferred from symptoms.

Attribution: f discounts transfer from researched modalities/populations to the actual mixture of adult trauma, family care and routine delivery. It is not the accounting program-expense share and does not remove overhead twice. Full-portfolio dollars remain in c. Clinical improvements and funding access are separate links.

The central is a deliberately cautious judgment informed by observed costs and external causal evidence, not a fitted organizational effect. First-year utility, long-term persistence, treatment episode length and counterfactual access dominate. Scenarios span plausible favorable and poor execution rather than statistical percentiles; zero health additionality remains possible.

Sensitivity

• At central inputs, halving episode length from30 to15 sessions halves the price only if net clinical efficacy is preserved; more sessions are not automatically more QALYs.
• With30 sessions at$275 and f×b×s×g=.3564, meeting$100,000 per10 requires q≥2.3148 per entry; meeting$1 million requires q≥.23148. Central q=.11553 meets neither.
• Holding central efficacy and other factors fixed, the$1 million threshold requires n≤14.97 sessions; the$100,000 threshold requires n≤1.50, which is inconsistent with a full clinical course.
• Dropping durability to one year raises central price to roughly$4.77 million per10. Benefit persistence beyond3years could improve the estimate but is not assumed without follow-up.
• No suicide deaths or family spillover are added. If validated, these could increase benefit; symptom reduction alone is insufficient evidence.
• Using2025 annual-report instead of tax expense changes the gross session benchmark by less than1%, much less than uncertainty in episode length or counterfactual care.

Unresolved inputs

• A current completed entry-cohort dataset linking total sessions across years, early exits, modality, diagnoses and health utility at6/12/24months.
• Whether marginal unrestricted dollars expand clinician capacity or replace committed fee-for-service, public/private funding or reserves.
• Client VA/insurance eligibility, previous treatment, likely alternative provider and waiting time without Headstrong.
• Current measured health utility or validated symptom-to-utility mapping, with adverse outcomes and missing follow-up included.
• Residence within50states/DC, not merely military affiliation or clinician location.
• Reconciliation of annual report, audited and Form990 financial differences; current2026 operating budget and reserve deployment plan.
• Actual patient copays/time costs and incremental downstream medical utilization, without assuming all avoided bills are net savings.