New York State Medicaid spent $94.6 billion in FY 2023 - more per resident than any state in the nation, 77% above the national average. A meaningful fraction of that sum funds not the treatment of women, but the failure to treat them - repeated ED visits that yield no diagnosis, specialist referrals that chase symptoms, and exploratory testing that generates billing without answers.
We can estimate the scale of this waste using published per-patient excess cost data. Endometriosis patients cost Medicaid $7,891 more per year than matched controls (JMCP, 2019). SLE patients cost $10,984 more (NIH systematic review). These are measured, peer-reviewed figures - not projections.
"Systems are only as good as the assumptions they are built on."
ORI-02 runs HCAI inpatient 2022 plus CMS Chronic Conditions plus state managed-care feeds through a deterministic pathway-misrouting filter. Output is a county-by-county map of avoidable spend tied to specific ICD-10 handoff failures.
The Failure Demand Framework
In every domain of modern operations - manufacturing, logistics, software - we have a name for the cost of getting things wrong the first time: failure demand. A field study evaluating healthcare communication systems found that 42% of all patient inquiries were classified as failure demand.
For conditions that exclusively or disproportionately affect women, failure demand is supercharged by systemic nosological biases. A Danish population-level registry study found women were diagnosed on average about four years later than men across 770 diseases. Every year of delay compounds the failure demand.
The underlying pathology remains unresolved, symptoms escalate, and the patient is forced to repeatedly re-engage the healthcare system. Because outpatient providers fail to identify the root cause, patients escalate care-seeking to the emergency department - driving costs exponentially higher. The average ED visit in New York: $1,200. A standard primary care visit: $300.
Condition-Specific Analysis
Affects ~10-11% of individuals with a uterus aged 15-44. Average diagnostic delay: 7.5 to 11 years. 75% of women experience misdiagnosis. 58% visit their GP more than ten times. 53% present to the ED when outpatient care fails.
Published excess cost: $13,670 vs. $5,779 for controls = $7,891/year excess per patient (JMCP, 2019, Medicaid population). In the 5 years surrounding diagnosis, patients spend $26,305 more than controls (Soliman et al.).
Median time to diagnosis: 7 months. Median consultations: 3. Not a biological equivalent - used to illustrate pathway efficiency, not biological parity.
Leading cause of female infertility. Up to 70% remain entirely undiagnosed (Gibson-Helm et al.). 33.6% reported delay of 2+ years. 47.1% saw 3+ distinct health professionals before diagnosis.
Women are treated in rigid clinical silos: a dermatologist treats acne; a PCP focuses on weight; a gynecologist prescribes OCs. The underlying endocrine pathology is ignored. What could be managed with inexpensive metformin transitions into lifelong T2D management.
Standard of care dictates immediate, comprehensive endocrine panel. Illustrates pathway efficiency difference, not biological equivalence.
Women account for ~80% of all autoimmune disease. Sjögren's: 9:1 to 19:1 F:M. SLE Medicaid excess cost: $10,984/year per patient above controls (NIH). Severe SLE: $68,260/year direct costs (Lupus Foundation claims study).
Delayed immunosuppressive therapy allows irreversible tissue damage. 40% of SLE patients quit jobs within 3.4 years of symptom onset (Partridge et al., 1997). Prompt diagnosis → low-cost outpatient management.
CVD: leading cause of death in NYS (31% of all deaths). Women wait 21, 62.9, and 134 days longer than men for acute, mid-chronic, and long-term chronic cardiovascular phenotypes.
Sex-differentiated troponin levels increase diagnosis of ACS in women by 42%. The failure to use these metrics generates continuous high-cost ED failure demand.
Aggregate Cost Modeling
The table below builds from published per-patient excess costs and applies them to estimated NYS Medicaid cohort sizes. Per-patient figures are from peer-reviewed studies. Cohort sizes are estimates derived from national prevalence data applied to NYS Medicaid enrollment (~7.9M total, ~3.5M female).
| Condition | Per-Patient Excess/Yr | Source | Est. NYS Cohort | Basis | Modeled Annual Excess |
|---|---|---|---|---|---|
| Endometriosis | $7,891 | JMCP, 2019 | ~65,000 | 10% of ~1.3M reproductive-age ♀ on Medicaid × 50% undiagnosed | $513M |
| Autoimmune (SLE) | $10,984 | NIH review | ~18,000 | 0.5% prevalence × 3.5M ♀ | $198M |
| CVD (women <65) | $8,300 (est.) | AHRQ ED + inpatient data | ~50,000 | Acute-phase, dual-eligible cohort | $20M (conservative) |
| Conditions with published per-patient excess data | $731M | ||||
Note: PCOS excess cost data is not available at a Medicaid-specific, per-patient level and is therefore excluded from the aggregate. The $731M figure is a modeled estimate: per-patient costs are published; cohort sizes are estimates. The actual figure could be higher (comorbidity overlap, downstream metabolic costs from undiagnosed PCOS) or lower (not all excess cost is attributable to diagnostic delay). This is not an official government projection.
The published per-patient data is unambiguous: endometriosis patients cost Medicaid $7,891 more per year than matched controls. SLE patients cost $10,984 more. These are measured excess costs from claims databases - not projections. The only modeling input is the cohort size.
ORI: Algorithmic Failure Demand Interception
While human clinicians view encounters in isolation, an AI-driven case management tool like ORI views the patient's entire longitudinal claims history instantaneously.
Example: A female Medicaid beneficiary registers three ED visits for non-specific pelvic pain, two PCP visits for fatigue, and a GI referral within 18 months. Human providers miss the pattern. ORI flags this profile as consistent with pre-diagnosis endometriosis - short-circuiting the failure demand cycle.
ORI is strictly administrative and care-coordinating, not diagnostic. Its algorithms can be continuously audited for demographic bias - ensuring that a governed AI system corrects the historical, systemic human bias.
Even a conservative 20% reduction in per-patient excess costs through faster diagnosis would yield approximately $146 million in direct annual savings for NYS Medicaid (20% × $731M modeled aggregate).
For Decision-Makers
Action: Commission SPARCS-based retrospective analysis of sex-disaggregated pre-diagnosis utilization for endometriosis, autoimmune, and CVD cohorts. Validate the per-patient excess costs in NYS-specific data. Pilot ORI in 2-3 MCO networks.
Published claims data shows $7,891-$10,984 in annual per-patient excess for undiagnosed endo and SLE patients. This excess is currently priced into your baselines. Action: Apply sex-disaggregated lens to claims analytics. Build "diagnostic fast-track" pathways for high-frequency female presenters.
The real-world evidence for sex-differentiated therapeutics sits in state Medicaid claims databases. Action: Partner with SPARCS/CMS data access programs for sex-stratified RWE generation.
SPARCS, CMS utilization data, and HCUP cost-to-charge ratios are publicly accessible. Action: Build sex-specific utilization pattern libraries. Position for DIGIT procurement as the office seeks operational proof-of-value.
Scope a Pilot ORI Workstream
Operational memo showing where Medicaid is paying for failure demand and how to intercept it safely.
Contact FemTechnology →Sources & Evidence Base
All statistics in this analysis are sourced from peer-reviewed literature, government statistical offices, or published claims datasets. Key references:
- Soliman AM et al., Adv Ther 2018 - Endometriosis: $16,573/yr vs $4,733 controls. DOI: 10.1007/s12325-018-0667-3
- Bernstein JA et al., JAMA Intern Med 2024 - Only 19% of GDM women get postpartum screening. n=47,953
- Stevens W et al., AJOG 2017 - SMM hospitalization avg $32,000. DOI: 10.1016/j.ajog.2017.04.032
- Parker ED et al., Diabetes Care 2024 - T2D: $19,736/patient/yr. DOI: 10.2337/dci22-0078
- NYS DOH SPARCS (Statewide Planning and Research Cooperative System) discharge data