Google's Gemini for Home AI, which powers camera summaries and alerts on Nest cameras, routinely hallucinates events that never happened. Documented cases include: fabricating social interactions ('Jenni and R were seen interacting with trick-or-treaters') when the person was not home, identifying a shotgun as a 'garden tool,' consistently misidentifying dogs as cats or deer, and reporting 'fake people' on live feeds when nobody was present. The AI refuses to identify weapons even when deliberately shown one, and repeatedly misidentifies pets despite user corrections. This matters because security cameras exist for one reason: to accurately tell you what is happening at your home when you are not there. When the AI fabricates events, homeowners cannot trust any alert. A false 'intruder detected' at 2 AM causes genuine panic — elevated heart rate, calling the police, or rushing home. When it happens repeatedly, users start ignoring alerts entirely, which is even more dangerous. The boy-who-cried-wolf effect means that when a real break-in happens, the homeowner dismisses the notification as another Gemini hallucination. Google is charging for these AI features through Google Home Premium (which increased from $8 to $10/month in August 2025), meaning users pay a subscription fee for a system that actively makes them less safe. This problem persists because Google shipped Gemini for Home as a marketing-driven feature rather than a safety-critical system. The AI model was not trained or validated to the standard required for security applications. There is no accuracy SLA, no false-positive rate published, and no liability when a hallucinated 'all clear' summary fails to mention an actual intruder. Google's incentive is to ship AI features fast to justify subscription revenue, not to achieve the 99.9%+ accuracy that a security system demands. Users have no way to revert to the previous, simpler motion-detection system that was less 'intelligent' but more reliable.
Real problems worth solving
Browse frustrations, pains, and gaps that founders could tackle.
On March 28, 2025, Amazon disabled the 'Do Not Send Voice Recordings' feature on Echo Dot (4th Gen), Echo Show 10, and Echo Show 15 devices. This feature had allowed users to process Alexa commands locally on the device and send only text transcripts to Amazon's cloud. After the change, all voice recordings are sent to Amazon's servers. Users who had this setting enabled were automatically switched without their active consent. This matters because people put these devices in their bedrooms, kitchens, and living rooms — the most intimate spaces in their homes. The device is always listening for its wake word, and false triggers happen regularly, capturing fragments of private conversations, arguments, medical discussions, and financial calls. Before March 28, privacy-conscious users had a meaningful technical control: their actual audio stayed on the device. Amazon took that control away unilaterally to support Alexa+ generative AI features that require cloud processing. The user's only remaining option is 'Don't Save Recordings,' which still sends audio to Amazon's servers — it just deletes it after processing. The difference between 'never sent' and 'sent but deleted' is enormous from a privacy and legal exposure standpoint. This problem persists because the smart speaker market is an oligopoly. Amazon, Google, and Apple control virtually all voice assistant hardware. When Amazon degrades privacy, users cannot switch to a meaningfully different alternative — Google collects similar data, and Apple's HomePod has a fraction of the integrations. The structural incentive is clear: voice data is training data for AI models, and AI models are the competitive battleground. Users' privacy preferences are a cost center that directly conflicts with the company's AI strategy. There is no regulatory framework in the US that prevents a hardware manufacturer from changing data collection practices post-sale via a firmware update the user cannot refuse.
On January 31, 2026, Belkin shut down cloud support for 27 Wemo smart home products, including light switches, smart plugs, coffee makers, and motion sensors. Every device that relied on Wemo's cloud servers for remote access and voice assistant integrations stopped working as a "smart" device overnight. Consumers who paid premium prices for these products — often $30-$80 per device — were left with expensive dumb switches and paperweights. This matters because smart home devices are marketed and priced as long-term investments in your home infrastructure. A traditional light switch costs $2 and lasts 30 years. A Wemo smart switch costs $50 and lasted until a corporate decision rendered it inert. Homeowners who wired dozens of Wemo switches throughout their house now face the choice of ripping them all out and rewiring (hundreds of dollars in electrician fees) or living with dead switches. The asymmetry is brutal: the consumer bears 100% of the financial loss while the manufacturer walks away. A class action lawsuit was filed by Sauder Schelkopf on behalf of affected consumers, but Belkin only offered partial refunds to devices still under warranty — meaning the vast majority of owners, whose devices were out of warranty but still perfectly functional hardware, got nothing. This is not a one-off event. Devolo killed its Home Control line in 2025. Google retired first and second-gen Nest Learning Thermostats in October 2025. Neato's cloud servers for robot vacuums were shut down three years after the company folded. The structural reason this keeps happening is that no regulation requires smart home manufacturers to open-source their server code, provide local-control fallbacks, or maintain cloud services for any minimum period after sale. The consumer has zero leverage once the purchase is made.
To design a clinical trial for a rare disease, you need to know what the disease actually does over time without treatment: how fast it progresses, what the measurable milestones of decline are, which biomarkers correlate with symptoms, and what the variability looks like across patients. This information comes from natural history studies -- systematic longitudinal observation of untreated patients. For the vast majority of the 10,000+ known rare diseases, no such study exists. There is little existing knowledge on the initial manifestations, major symptoms, day-to-day functional limitations, and progression patterns. Without this data, a drug developer cannot select appropriate endpoints, cannot power a trial correctly, and cannot construct an external control arm. The result is a Catch-22: you can't get a drug approved without clinical trial data, and you can't design a clinical trial without natural history data that nobody has collected. The FDA has explicitly recognized this gap as one of the primary bottlenecks in rare disease drug development. In response, the agency has funded natural history studies through its Orphan Products Grants Program and published guidance documents on how to conduct them. In 2024, the FDA established a Rare Disease Policy and Portfolio Council (RDPPC) and a Rare Disease Innovation Hub specifically to accelerate this work. But the scale of the problem dwarfs the response: there are thousands of conditions where the basic biology is understood (the gene is known, the mechanism is characterized, animal models exist) but there is no systematic human data on how the disease actually presents and progresses. Academic proof-of-concept studies sit in journals with no path to clinical translation. The structural barrier is that natural history studies are expensive, slow, and unglamorous. They require following patients for years, often across multiple countries (since any single center may only have a handful of patients), with no guarantee that a drug will ever be developed to treat the condition. No pharmaceutical company will fund a natural history study for a disease they're not sure they'll pursue commercially. Academic grants are competitive and short-term. Patient advocacy groups try to fill the gap but lack the epidemiological infrastructure to run rigorous longitudinal studies. The FDA's recent willingness to accept real-world data and external control arms is promising, but it requires the data to actually exist in a standardized, analyzable format -- which, as the registry fragmentation problem shows, it usually doesn't. Until someone systematically funds natural history data collection for the thousands of rare diseases that have none, the pipeline from lab discovery to patient treatment will remain blocked at the evidence generation stage.
Bluebird Bio developed two gene therapies -- Zynteglo for transfusion-dependent beta-thalassemia and Skysona (eli-cel) for cerebral adrenoleukodystrophy (CALD) -- that represented genuine one-time cures for devastating conditions. Both received European marketing authorization. Both demonstrated clinical efficacy. And both were withdrawn from the entire European market by early 2022 because Bluebird could not reach reimbursement agreements with European payers. The company's CEO called the European pricing situation "untenable" and described the market as "broken." Bluebird closed down all European operations to focus exclusively on the US. For European patients with transfusion-dependent beta-thalassemia -- who face lifelong biweekly blood transfusions and iron chelation therapy, with progressive organ damage from iron overload -- Zynteglo's withdrawal meant the loss of their only shot at a transfusion-free life. For children with CALD -- a condition where the myelin sheath of the brain is progressively destroyed, leading to loss of speech, vision, hearing, and motor function -- Skysona's withdrawal meant the only alternative was allogeneic stem cell transplant, which requires a matched donor and carries significant mortality risk. These patients didn't lose access because the drug failed. They lost access because the financial negotiation between a company and national health systems failed. The situation has only gotten worse. In the US, the FDA restricted Skysona's use in 2024 after reports of blood cancer in some treated patients, limiting it to patients without an available matched donor. Skysona recorded zero sales in the first three months of 2025. Meanwhile, Bluebird Bio itself faced severe financial difficulties, with layoffs in 2024 and ongoing cash flow struggles. The structural problem is that the gene therapy business model is fundamentally fragile: massive upfront R&D costs, tiny patient populations, uncertain long-term safety profiles, and payers who are structurally unable to absorb seven-figure per-patient costs. When the company developing your only treatment goes bankrupt or exits your market, there is no backup plan. The drug doesn't transfer to a generic manufacturer. The manufacturing process is too specialized. You simply go without.
A child with a rare metabolic disorder, skeletal dysplasia, or inherited neuromuscular condition typically receives care at a specialized pediatric center where the team has seen dozens of similar cases. The geneticist, metabolic dietitian, physical therapist, and social worker all know the disease. Then the patient turns 18 (or 21, depending on the institution), and they're discharged to adult medicine -- where their new internist has never heard of their condition, has no treatment protocol, and has no idea which specialists to refer to. A 2025 position paper in the Orphanet Journal of Rare Diseases documented 21 prioritized issues in this transition, concluding that "while they often received specialized pediatric care, the adult healthcare system provides few resources for those whose rare disease began in childhood, and many adult care providers are hesitant to take on adults with rare disease." The clinical consequences are measurable. A study of 183 sickle cell disease patients at St. Jude found that those with more than six months between their last pediatric and first adult appointment had significantly more emergency department and inpatient visits and fewer outpatient visits than those who transitioned within six months. That gap -- the months or years where a patient has no specialist following them -- is where complications go undetected, medications get discontinued, and preventive care lapses. For rare disease patients whose conditions require ongoing monitoring to prevent crises (metabolic decompensation, cardiac events, respiratory failure), losing specialist oversight isn't an inconvenience. It's a medical emergency waiting to happen. The structural cause is that adult medicine is organized around common conditions: cardiology, oncology, endocrinology. There is no "adult rare disease" specialty. Pediatric geneticists and metabolic specialists don't have adult counterparts in most health systems. The National Survey of Children's Health found that only 18.4% of adolescents with special healthcare needs received any transitional support in 2019. Building transition programs requires funding, cross-departmental coordination, and institutional willingness to create clinics for patient populations too small to generate revenue. Most hospitals won't do it because rare disease patients are, by definition, rare -- there aren't enough of them to justify a dedicated adult transition program under fee-for-service economics.
The Orphan Drug Act was designed to incentivize drug development for diseases affecting fewer than 200,000 Americans, offering 7 years of market exclusivity, tax credits, and reduced FDA fees. It has been wildly successful for conditions with thousands of patients -- the orphan drug market now exceeds $216 billion. But for ultra-rare and hyper-rare diseases affecting fewer than 100 to 1,000 patients globally, the math doesn't work even with incentives. Farber disease has about 80 known patients worldwide. Fibrodysplasia ossificans progressiva ("stone man syndrome") has roughly 800 confirmed cases globally. Hutchinson-Gilford Progeria Syndrome has about 400 known cases. For these conditions, no amount of market exclusivity makes a commercial drug development program viable because the total addressable market is too small to recoup R&D costs. The result is a two-tier system within rare diseases. Conditions with 10,000+ patients attract pharma investment and get approved therapies. Conditions with fewer than 1,000 patients remain in what researchers call "the valley of death" -- academic labs may identify the gene, characterize the mechanism, even demonstrate proof-of-concept in animal models, but no commercial entity will fund the clinical trials to bring a treatment to market. A 2025 paper in Drug Discovery Today documented that "lack of drug-discovery expertise and financial resources are key obstacles" and that "the long and difficult path to translate most academic proof-of-concept studies into approved medicines means that very few therapies ever reach patients" for these ultra-rare conditions. The structural failure is that the Orphan Drug Act treats all rare diseases as a single category when they are actually on a spectrum spanning four orders of magnitude in patient population. A disease with 150,000 patients and a disease with 80 patients both qualify for orphan status, but they have fundamentally different commercial dynamics. The non-profit sector, academic medical centers, and patient advocacy foundations try to fill the gap through mechanisms like the NIH NCATS program, but they lack the manufacturing, regulatory, and distribution infrastructure to bring drugs through Phase III trials and to market. Some companies, like Orchard Therapeutics, initially took on ultra-rare conditions only to abandon them -- Orchard discontinued its investment in rare primary immune deficiencies in 2022. For the families of children with ultra-rare diseases, the message is clear: your child's disease is too rare to be worth saving commercially.
Lenmeldy, a gene therapy for metachromatic leukodystrophy, costs $4.25 million for a single dose. Roctavian for hemophilia A costs over $2 million. Zolgensma for spinal muscular atrophy costs $2.1 million. These are one-time treatments that can functionally cure diseases that would otherwise require millions in lifetime care -- or kill the patient. The clinical value is real. The problem is that the entire insurance payment infrastructure was built for chronic treatments billed monthly, not one-time curative doses billed all at once. A single patient receiving Lenmeldy can blow through a small employer's entire annual health budget. The downstream effects ripple through the system. Insurers respond by increasing premiums, raising deductibles, excluding gene therapies from formularies, or steering patients toward manufacturer patient assistance programs -- effectively making the pharma company the insurer of last resort. Medicaid programs, which cover many rare disease patients, face even steeper barriers: state budgets aren't designed to absorb $4 million line items. The result is that gene therapies exist, are FDA-approved, and demonstrably work, but patients can't access them because the payment plumbing can't handle the transaction. Only 5% of FDA-approved gene therapies have been reimbursed across global health systems. The CMS Cell and Gene Therapy Access Model, launched with 33 states plus DC and Puerto Rico, represents the first federal attempt to solve this by negotiating pricing deals with manufacturers on behalf of state Medicaid programs, starting with sickle cell disease. But it covers only Medicaid, only specific conditions, and only treatments approved before May 2024. Private insurance remains the wild west. Proposed solutions -- installment payments, outcomes-based agreements where the manufacturer refunds money if the therapy doesn't work, annuity models -- all exist in theory but require infrastructure (patient tracking over decades, inter-payer data sharing, legal frameworks for refunds) that doesn't exist yet. The structural problem is that curative medicine broke the actuarial model that funds American healthcare. Until someone builds a new financial architecture, million-dollar cures will continue to sit on pharmacy shelves while the patients they could save go without.
There are over 10,000 known rare diseases, and for most of them, the total body of clinical knowledge exists in fragmented registries run by individual hospitals, patient advocacy groups, pharma companies, and academic consortia. These registries use different data standards, different consent frameworks, different variable definitions, and different access policies. A researcher studying Niemann-Pick disease type C cannot combine data from the NIH registry, the European registry, the pharma-sponsored registry, and the patient advocacy registry into a single dataset without years of negotiation and manual harmonization. The result is that each new clinical trial for a rare disease must re-recruit patients and re-collect natural history data from scratch, as if previous research never happened. This fragmentation directly harms patients. Families are asked to fill out the same intake questionnaires, provide the same medical records, and undergo the same assessments for multiple overlapping registries -- a burden that the Orphanet Journal of Rare Diseases has documented as increasingly unsustainable. Meanwhile, researchers can't answer basic questions about disease progression because the data exists in five places, none of which has enough patients alone to reach statistical significance, and none of which can be legally or technically combined with the others. Drug developers can't use existing registry data as external control arms for clinical trials because the data wasn't collected with consistent endpoints. The structural cause is a tragedy of the commons. Each stakeholder has rational reasons to maintain their own registry: pharma companies want proprietary competitive advantage, academic groups need to control data for publications, patient groups want to maintain trust with their community, and hospitals face HIPAA and GDPR constraints on sharing. The GA4GH (Global Alliance for Genomics and Health) has been working on FAIR data principles for rare diseases, but adoption is voluntary and slow. The European Reference Networks have attempted cross-border registry harmonization but face persistent challenges with non-interoperable systems. Until someone solves the incentive problem -- making it more valuable to share data than to hoard it -- rare disease research will continue to waste millions of dollars and years of patient time re-collecting information that already exists somewhere in a silo nobody can access.
The federal Recommended Uniform Screening Panel (RUSP) lists conditions that all states should screen newborns for, but states are not required to adopt the full panel, and many don't. The result is a geographic lottery. A baby born in New York gets screened for over 60 conditions. A baby born across the river in New Jersey may not be screened for the same set. Krabbe disease -- a devastating neurodegenerative condition where early stem cell transplant before symptom onset can preserve the ability to walk and talk -- was screened in only about 10 states before it was finally added to the RUSP in July 2024. Arthur Hutchinson, a child with Krabbe disease, was diagnosed only after he mysteriously stopped crawling. By then, it was too late for the transplant to restore full function. He was born in a state that didn't screen for Krabbe. The consequences of this inconsistency are not theoretical. Spinal muscular atrophy (SMA) was added to the RUSP in 2018, and states took years to implement screening. A 2024 study found significant variation in SMA screening practices across states, with inconsistencies in how results are reported and confirmed, and at least two states with high false-positive rates. For conditions like SMA where gene therapy (Zolgensma) must be administered before motor neurons degenerate, a delay of even weeks can mean the difference between a child who walks and one who doesn't. The structural barrier is that each state runs its own newborn screening program with its own funding, its own laboratory infrastructure, and its own political priorities. Adding a new condition requires legislative or regulatory action at the state level, and advocacy groups must fight 50 separate battles. Making matters dramatically worse, the Trump Administration disbanded the Advisory Committee on Heritable Disorders in Newborns and Children (ACHDNC) in April 2025 -- the only federal body that evaluated evidence and recommended conditions for the RUSP. Without this committee, there is currently no mechanism to add new conditions to the federal panel, meaning the pipeline of life-saving screenings has been frozen at the federal level while states remain free to do nothing.
Step therapy -- also called "fail first" -- is an insurance cost-control practice where patients must try and fail on cheaper medications before their insurer will cover the drug their doctor actually prescribed. For common conditions like hypertension, this is annoying but rarely dangerous. For rare diseases with narrow treatment windows, it is catastrophic. A patient with a progressive neurodegenerative condition forced to spend 90 days on an ineffective first-line drug isn't just wasting time -- they're losing motor neurons, muscle function, or cognitive capacity that they will never recover. A 2016 Patients Rising survey found that 77% of patients reported that step therapy undermined their healthcare by forcing them to use ineffective treatments. The cruelty is compounded by the bureaucratic loop. When a patient fails the required drug (as their specialist predicted they would), the insurer often requires documentation of that failure -- a process that can take weeks of additional delay. In some documented cases, patients are required to fail on the same medication more than once, in direct violation of state reform laws. JP Summers, a migraine patient in Wisconsin, spent four years battling his insurance company to get approved for the medication his neurologist prescribed, cycling through cheaper alternatives that caused side effects serious enough to hospitalize him six times. The structural reason this persists is that state step therapy reform laws -- now enacted in roughly 20 states -- only apply to state-regulated insurance plans. They don't cover self-insured employer plans, which cover more than half of American workers. The federal Safe Step Act, introduced in 2023 with bipartisan support (39 Senate cosponsors, 180 House cosponsors), would establish automatic patient exemptions from step therapy in employer plans, but it has not passed. Meanwhile, a 2024 New York survey found that almost half of healthcare professionals reported insurers requiring patients to fail on the same medication more than once, and the majority did not believe existing state laws had meaningfully reduced step therapy denials. For rare disease patients whose conditions are actively progressing, every week of mandated failure is tissue damage that no subsequent treatment can reverse.
Hypermobile Ehlers-Danlos Syndrome (hEDS) causes chronic pain, joint instability, fatigue, and autonomic dysfunction that mimics a half-dozen more common conditions. The average hEDS patient receives 10.45 alternative diagnoses -- including fibromyalgia, anxiety, depression, migraines, and POTS -- before anyone considers the actual connective tissue disorder. The average time to correct diagnosis is over 10 years, with some studies reporting delays of 15 to 20 years from symptom onset. During those lost years, patients are treated for conditions they don't have with medications that don't work, while the underlying joint hypermobility causes progressive damage that proper physical therapy and bracing could have mitigated. The real harm isn't just the delay. It's what happens during the delay. Patients with undiagnosed hEDS undergo surgeries that fail because the surgeon didn't account for fragile tissue. They're prescribed SSRIs for "anxiety" that's actually POTS-related tachycardia. They're told their pain is psychosomatic. A 2023 study in Genetics in Medicine Open found that 56% of hEDS patients received at least one misdiagnosis, and those who were misdiagnosed took an average of 19 years to reach the correct diagnosis compared to 8 years for those who weren't. That's 11 additional years of wrong treatment, unnecessary procedures, and being told nothing is wrong. This problem persists for a structural reason: hEDS has no genetic test. Unlike other forms of Ehlers-Danlos (vascular, classical), where a specific gene mutation confirms the diagnosis, hEDS is diagnosed by clinical criteria alone -- the 2017 international criteria that require documenting joint hypermobility, systemic features, and family history. Most primary care physicians and even many rheumatologists have never learned these criteria. Medical school curricula devote minimal time to connective tissue disorders. The Beighton score for hypermobility is rarely taught. And because hEDS symptoms overlap heavily with fibromyalgia, chronic fatigue syndrome, and functional neurological disorder, the path of least resistance for a busy clinician is to assign one of those labels and move on. Until medical education catches up, hEDS patients will continue to be gaslit by a system that doesn't know their disease exists.
When a child presents with unexplained developmental delays, seizures, or multi-system symptoms that don't match any common condition, the single most informative diagnostic test is whole exome sequencing (WES), which reads all 20,000 protein-coding genes at once. It costs between $5,000 and $12,000. Yet insurance companies deny prior authorization for WES approximately 18% of the time, often classifying it as "experimental" or "investigational" despite it being in clinical use since 2011 and now standard practice at every major academic medical center. The downstream cost of this denial is staggering. When researchers at UCLA and other institutions studied 147 patients whose WES was denied by insurance, 53.7% had at least one diagnostic or candidate finding and 10.9% had a clinically diagnostic result that immediately changed medical management. Those patients would have gotten their answer in weeks. Instead, they spent years cycling through specialist after specialist, racking up MRIs, biopsies, metabolic panels, and hospital admissions that individually cost more than the denied test. The average rare disease patient sees 8 or more doctors over 4.7 years before diagnosis. Each misdiagnosis isn't just a billing event -- it's a wrong treatment, a side effect, a lost year of a child's development. This problem persists because insurance utilization management criteria lag behind clinical practice by a decade. The coverage determination committees that write these policies rely on outdated evidence standards and don't update them until forced by regulation or litigation. Private insurers are significantly more likely to deny WES than public insurance programs, because private plans face less regulatory pressure to cover advanced diagnostics. Meanwhile, the patients who suffer most are those with the rarest conditions -- the ones whose symptoms don't match any textbook, who fall through every diagnostic algorithm, and whose only realistic path to an answer is broad genomic testing that their insurer refuses to pay for.
When transit agencies redesign their bus networks to improve frequency — running buses every 15 minutes instead of every 45 — they face an brutal tradeoff with fixed budgets: every dollar spent on higher frequency on core routes is a dollar taken from coverage on low-ridership routes. Atlanta's MARTA made this tradeoff explicitly in its bus network redesign: it cut the system from 113 fixed bus routes down to 79, concentrating resources so that 18 corridors could offer 15-minute service (up from just 5). The redesign received no additional funding. For riders on the 34 eliminated routes, the "improvement" meant their bus service disappeared entirely. The human impact of this tradeoff falls on a specific population: riders in low-density neighborhoods, often in the southern and western parts of the metro area, who had infrequent but existing bus service. A housekeeper who took the Route 83 to her job in Buckhead now has no bus at all. She can't drive to the nearest remaining route because she doesn't have a car — that's why she took the bus. The replacement is usually a vague promise of "microtransit" or on-demand shuttles that require a smartphone app, often don't materialize for months after the route cut, and when they do operate, have limited hours and capacity. The MARTA redesign held public meetings where riders in affected areas voiced exactly these concerns, but the math was immovable: with a fixed operating budget, you can have frequency or coverage, not both. This tradeoff persists because US transit agencies are funded through a patchwork of federal formula grants, state allocations, and local sales taxes that almost never grow with demand. MARTA's operating budget is constrained by a 1971 sales tax allocation that hasn't been meaningfully updated. Without new revenue, the only way to improve service quality (frequency) is to reduce service quantity (coverage). Transit planning consultants like Jarrett Walker have written extensively about this tradeoff being inherent to fixed-budget systems, but the policy response — dedicated transit funding that grows with the economy — requires voter approval of tax measures that have failed repeatedly in Atlanta's suburban counties. The riders who lose service in a redesign are usually the least politically powerful, least likely to attend public meetings, and least able to organize opposition.
The San Francisco Bay Area has 27 separate transit agencies, each with its own fare structure, payment system, and transfer policy. A rider commuting from Oakland to a job in San Francisco's Mission District might take AC Transit to BART to Muni — three separate agencies, three separate fares, no automatic transfer credit. A trip that costs $2.50 in a city with unified transit can cost $8-12 in the Bay Area because each agency charges independently. The Clipper card provides a single payment medium, but it doesn't provide fare integration: you still pay full price to each agency. This isn't a minor irritation — it's a structural tax on multi-modal trips that makes transit economically irrational for many journeys compared to driving. The real pain shows up in the math riders do every morning. A round-trip commute crossing two agency boundaries can cost $16-24/day, or $350-500/month. At that price, driving — even with gas, insurance, and parking — is often cheaper, especially for households with more than one commuter. This price penalty specifically punishes riders who live in affordable housing (further from job centers, requiring more transfers) while subsidizing riders who can afford to live within a single agency's service area. It's a regressive geography tax. The Metropolitan Transportation Commission (MTC) has studied this for years and found that inconsistent fare structures and transfer policies across Bay Area agencies actively discourage trips involving multiple operators. The reason 27 agencies can't coordinate fares is governance fragmentation that dates back decades. Each agency has its own board of directors, its own budget, its own union contracts, and its own political constituency. BART is governed by an elected board; Muni answers to the SFMTA board appointed by the mayor; AC Transit has its own elected board. Fare integration means revenue sharing, which means every agency fears losing money to the others. MTC has been working on a transfer discount pilot — no-cost and reduced-cost transfers for riders using two or more agencies — but even a pilot program has taken years of negotiation because each agency's finance department models the revenue impact differently and none want to be the one that loses. The technical infrastructure for seamless transfers exists (the Clipper card could handle it); the barrier is entirely political and financial.
In most US transit systems, the cheapest way to ride is to buy a weekly or monthly unlimited pass upfront. A daily commuter in many cities who buys a monthly pass pays effectively $1.50-2.00 per trip, while someone paying per ride pays $2.50-2.90. The catch: monthly passes require $70-130 in cash upfront, which many low-income riders simply don't have on the first of the month. So the riders who can least afford to pay more — hourly workers, people between paychecks, those without bank accounts — end up paying the highest per-trip cost. This is a textbook poverty premium: being poor is more expensive because you can't access bulk pricing. The magnitude of this penalty is staggering when annualized. A rider taking 2 trips per day, 22 days per month, at $2.75 per ride pays $1,452 per year. The same rider with a monthly pass at $90/month pays $1,080 per year. That's a $372/year penalty for not having $90 available on the first of the month — effectively a 34% surcharge on being cash-poor. For a worker earning $15/hour, that $372 represents 25 hours of pre-tax labor. And the riders paying this penalty are precisely the ones for whom $372 matters most: they're choosing between a transit fare and groceries, between getting to work and keeping the lights on. Fare capping — where the system automatically stops charging after a rider hits the equivalent of a pass price — is the known solution, and some cities have implemented it. NYC's OMNY system now caps at $35/week (equivalent to an unlimited MetroCard), and LA Metro has a similar program. But the majority of US transit agencies still don't offer fare capping, and even where it exists, it only works for riders using contactless payment cards. Cash riders — who are disproportionately unbanked, undocumented, elderly, or otherwise excluded from the digital payment system — can't benefit from fare capping at all. The structural barrier is that fare capping requires account-based payment infrastructure that costs millions to implement, and smaller agencies serving the most transit-dependent populations are the least able to afford it.
Of New York City's 472 subway stations, only about 28% are fully ADA-accessible — meaning they have elevators or ramps that allow wheelchair users, people with strollers, and those with mobility impairments to reach the platform. The remaining 72% have only stairs. For the stations that do have elevators, reliability is a persistent crisis: disability rights advocates documented an average of 25 elevator breakdowns per day across the system. When an elevator breaks at one of the few accessible stations, a wheelchair user's trip doesn't just get delayed — it gets canceled entirely, because the next accessible station might be a mile away in a different direction. The cascading impact of a single elevator outage reveals how fragile the accessible transit network really is. A wheelchair user plans a trip from the Bronx to Midtown. Their home station has an elevator. Their destination station has an elevator. But if either elevator is broken — and with 25 daily breakdowns, the odds are not trivial — they must reroute to a different accessible station, potentially adding 30-60 minutes and requiring a bus transfer to cover the gap. There is no real-time rerouting that accounts for elevator status; the MTA app shows elevator outages, but trip planners don't automatically suggest accessible alternatives. The rider has to figure it out themselves, often while standing at the top of stairs they cannot descend. Riders report missed job interviews, late arrivals to medical treatments, and the accumulated psychological toll of never being able to trust that their planned route will work. This problem is structural and decades in the making. The NYC subway opened in 1904, long before the ADA was passed in 1990. Retrofitting century-old underground stations with elevators requires excavation, structural reinforcement, and utility relocation that can cost $30-100 million per station. The MTA's 2025-2029 capital plan includes modernization of 45 elevators and accessibility upgrades at additional stations, but at the current pace, full system accessibility is estimated to take until the 2050s. The ADA itself contains loopholes: stations only need to be made accessible when they undergo "major renovations," which allows agencies to defer compliance indefinitely by classifying work as maintenance rather than renovation. A class-action lawsuit by Disability Rights Advocates has forced some progress, but litigation is slow and the physical constraints of stations built 120 years ago remain real.
Bus bunching is the phenomenon where buses scheduled to arrive every 10 minutes instead arrive in clusters of 2-3 followed by a 25-30 minute gap. It's caused by a self-reinforcing feedback loop: a bus runs slightly late (traffic, a wheelchair boarding, a long dwell at a busy stop), which means it picks up extra passengers who would have caught the next bus. More passengers mean longer boarding times, which makes it later still. Meanwhile, the bus behind it has fewer passengers to pick up, so it runs ahead of schedule and catches up to the late bus. The result: two buses arrive together, then nothing for 20+ minutes. The NYC Comptroller's office found that more than one in ten MTA buses fail to maintain even spacing, with riders on the worst routes waiting an average of 3.62 minutes longer than scheduled. The rider experience of bunching is far worse than raw delay numbers suggest. A rider who just missed the first bus in a bunch now waits 25-30 minutes for the next service, even though the schedule promised 10-minute frequency. They have no way to know this in advance — the schedule says 10 minutes, the app might say 10 minutes, but the reality is half an hour. This unpredictability is more damaging to transit usability than a reliably slow schedule, because riders can't plan around it. A bus that's reliably every 20 minutes is more useful than one that's "every 10 minutes" but actually arrives at 0-0-30 minute intervals. The bunching pattern also means the trailing buses in a bunch are nearly empty while the leading bus is dangerously overcrowded. Bunching persists because fixing it requires real-time operational control that most agencies lack. The technical solutions are well-known: hold the early bus at a timepoint for 60-90 seconds, or instruct the late bus to skip a low-ridership stop. But implementing this requires dispatchers who can see all vehicles in real time and communicate instructions to drivers, plus drivers who follow those instructions. Most agencies' dispatch operations are stretched thin (one dispatcher monitoring 50+ vehicles), and driver culture resists being told to wait when they're trying to stay on schedule. King County Metro reported progress reducing bunching in 2025 through automated holding algorithms, but this required a complete overhaul of their CAD/AVL system — an investment most agencies haven't made.
Most US transit systems stop running between midnight and 1 AM and don't resume until 4-5 AM. This creates a complete transportation blackout for the roughly 6 million Americans who work overnight shifts — hospital workers, warehouse employees, janitorial staff, hotel workers, airport ground crews. A Streetsblog analysis found that workers commuting between 10 PM and 3 AM had public transit commutes as long as 53 minutes, while the same trip by car would take half that. More critically, late-shift workers are 40% less likely than 9-to-5 workers to use mass transit at all, not because they don't want to, but because service literally doesn't exist during their commute hours. The downstream consequence is that overnight workers — who skew lower-income, disproportionately Black and Latino, and less likely to own cars — must spend money they don't have on alternatives. A night-shift hospital aide earning $16/hour who can't take the bus spends $15-25 per night on rideshare or $200+/month on a car they wouldn't otherwise need. That's a 10-15% effective pay cut imposed by the transit schedule. Some workers turn down higher-paying night shifts entirely because the transportation cost erases the wage premium. Others cobble together unsafe arrangements — walking miles at 2 AM through neighborhoods with no sidewalks, sleeping in break rooms until morning service starts, or carpooling with coworkers on different schedules. The structural reason transit shuts down at night is a combination of economics and maintenance. Agencies argue that ridership between midnight and 5 AM is too low to justify the operating cost of running full routes. This is partially true — but it's also circular: ridership is low because service is absent, and service is absent because ridership is low. The maintenance argument is stronger: rail systems in particular need overnight windows to inspect track, signals, and infrastructure, which is why even NYC's 24-hour subway has increasingly limited overnight service on specific lines. Bus service has no such constraint but still shuts down because agencies schedule driver shifts around peak service and don't want to pay overnight premiums. The few cities experimenting with solutions — Boston's MBTA tested extended Friday/Saturday service in fall 2025, offering free rides after 9 PM — treat it as a novelty pilot rather than essential infrastructure.
ADA-mandated paratransit service — the shared-ride, door-to-door service for people who can't use fixed-route transit due to disability — is extraordinarily time-inefficient. A 2024 peer-reviewed study using 2.5 years of data from Denver's Access-a-Ride system found that paratransit trips take nearly twice as long as the equivalent car trip for the same origin-destination pair. Worse, the standard deviation of paratransit trip time is four times that of car trip time, meaning riders can't even plan around the delay because it's wildly unpredictable. A 10-mile trip at 7-8 AM takes about 16 minutes by car with almost no variability, but averages 25 minutes by paratransit with 5% of trips falling completely outside statistical prediction. The real pain isn't abstract. Paratransit riders report missed medical appointments, lost jobs, surgeries that had to be rescheduled, and doctor's office no-show fees — all because the ride was late, went in circles picking up other passengers, or never arrived. Riders must book 1-2 days in advance, and the actual pickup time can be an hour earlier or later than requested. A rider with a 9 AM doctor's appointment might be told their window is 7:30-8:30 AM, then get picked up at 8:25, routed through three other pickups, and arrive at 9:40. The study found this time penalty is worse for women, older adults, cash-paying customers, and those making shorter trips — meaning the people most dependent on the service are most punished by it. This persists because paratransit's shared-ride model creates an inherent routing optimization problem that agencies solve poorly. Each vehicle serves multiple riders going to different destinations, and the routing software must balance vehicle capacity, time windows, and geographic spread. Most agencies use legacy routing systems that optimize for vehicle utilization (filling seats) rather than rider experience (minimizing detours). Modern on-demand algorithms from companies like Via have shown 30-40% improvements in trip time, but agencies are locked into multi-year contracts with incumbent vendors, and switching routing systems requires retraining dispatchers, recertifying vehicles, and navigating procurement rules that favor the lowest bidder over the best technology.
In Houston, Texas, the majority of bus stops are a bare pole with a sign. No roof, no shade, no bench. As of 2025, 64% of Metro bus stops have no shelter at all. Riders — disproportionately low-income, elderly, and disabled — stand in direct sun while waiting for buses that may be 15-30 minutes apart. A UTHealth Houston study measured wet-bulb globe temperatures (WBGT) averaging 92.5°F at unsheltered stops during summer, a level the CDC classifies as posing significant heat stress risk. Houston Public Media's "Hot Stops" investigation found at least 16 calls to 911 from people at bus stops for temperature-related emergencies in just June and July of one summer, plus roughly 200 calls classified as "unconscious person" or "sick person" at bus stops during the same period. The reason this matters beyond the immediate health risk is that it makes the bus a physically dangerous mode of transportation for a significant portion of the year. Houston's summers regularly exceed 100°F for weeks. A rider waiting 20 minutes for a bus in direct sun at 95°F ambient temperature isn't just uncomfortable — they're at medical risk for heat exhaustion and heat stroke. This effectively makes bus transit seasonal for vulnerable populations: elderly riders, people with chronic conditions, and parents with small children avoid the bus in summer because the wait is dangerous. That drives them toward cars they can't afford or keeps them home, cutting them off from jobs, medical appointments, and social participation. The problem persists for structural reasons tied to land use and property rights. Many bus stops sit on narrow sidewalks where there's physically no room for a shelter. Others are on private property frontage where the transit agency needs easements that property owners won't grant. The cost of a single bus shelter installation ranges from $10,000 to $50,000 depending on design and permitting, and Metro has roughly 9,000 stops. Even at scale pricing, sheltering every stop would cost hundreds of millions of dollars — money that competes with vehicle procurement, driver salaries, and route expansion. Metro has added about 1,000 shelters recently and plans 2,000 more by 2026, but at 9,000 total stops, full coverage is decades away at current pace and funding.
New York City's local buses crawl at an average of about 8 miles per hour systemwide. The worst offenders are dramatically slower: the M42 crosstown bus averaged 5.25 mph in 2025, earning the annual "Pokey Award" from transit advocates. The Bx35 managed 5.59 mph. At these speeds, a person on a bicycle — or in some cases, a brisk jogger — moves faster than a bus carrying 40 people. The core mechanical cause is simple: buses sit in the same traffic as private cars, stop at the same red lights, and have no physical infrastructure separating them from congestion. The downstream pain is enormous. An 8 mph average means a 4-mile crosstown trip takes 30 minutes by bus versus 12 minutes by car or 15 minutes by bike. This isn't a minor inconvenience — it makes the bus an irrational choice for anyone with alternatives, which hollows out ridership to only those who have no other option. Lower ridership weakens the political case for bus investment, which means fewer bus lanes, which means slower buses — a death spiral. The NYC Comptroller's "Life in the Slow Lane" report documented how this speed penalty falls disproportionately on low-income riders and communities of color who are most transit-dependent. The structural reason this persists is political, not technical. Physically separated bus lanes (concrete barriers, not just paint) demonstrably increase speeds by 15-33%. NYC's congestion pricing, implemented in January 2025, delivered immediate bus speed improvements in Manhattan. But installing bus lanes means removing car lanes or parking, which triggers fierce opposition from local businesses, elected officials responsive to car-owning constituents, and community boards. Each mile of protected bus lane requires years of environmental review, community engagement, and political capital. The result: the technical solution exists and is proven, but the political cost of deploying it at scale is one that few city leaders are willing to pay.
A "ghost bus" is a transit vehicle that appears on real-time tracking apps and station displays as scheduled or en route but never actually arrives. The root cause is a chain of technical failures: a driver doesn't log into the CAD/AVL system, or GPS hardware drops signal, or the agency cancels a trip but never pushes the cancellation to the GTFS-realtime feed. The app still shows the bus as coming. The rider waits. The bus never comes. Then the next bus shows up already full because it absorbed the ghost bus's passengers. This matters because it doesn't just waste 15-20 minutes of a single wait — it destroys the foundational trust that makes transit usable. If a rider can't trust the app's arrival time, they have to build in a buffer for every trip. A 30-minute commute becomes a 50-minute commute "just in case." Over weeks, that uncertainty tax adds up to hours of lost time, and riders who have any alternative — a car, a rideshare — switch permanently. The Chicago Transit Authority found ghost buses were so pervasive that it took until August 2025 to even begin sharing cancellation data with third-party apps, meaning for years riders were making decisions based on fiction. This problem persists because transit agencies' internal dispatch systems and their public-facing data feeds are separate stacks maintained by separate teams with separate incentives. Dispatch cares about moving vehicles; the data team cares about feed uptime. Nobody owns the gap between "we cancelled run 4472" and "Google Maps still shows run 4472 arriving in 3 minutes." The GTFS-realtime spec supports trip cancellation messages, but most agencies' legacy CAD/AVL systems can't generate them automatically. Fixing it requires either expensive hardware upgrades (dual GPS antennas, automated operator login detection) or manual processes where dispatchers update feeds in real time — which they don't have bandwidth to do during the exact moments (driver no-shows, breakdowns) when ghost buses are created.
Organ procurement organizations are among the last entities in U.S. healthcare funded on a pure cost-reimbursement basis. They are not paid per organ recovered. They are not paid based on transplant outcomes. They submit their costs to Medicare, and Medicare reimburses them. An OPO that recovers organs from 1 in 4 potential donors receives the same type of funding as one that recovers from 3 in 4. CMS found in its own assessments that 74% of OPOs were failing performance standards. Despite this, not a single OPO lost its contract until HHS moved to decertify the Life Alliance Organ Recovery Agency at the University of Miami in 2025 — the first decertification in the program's entire history. So what? Without financial consequences for poor performance, the lowest-performing OPOs have no external pressure to improve. As many as 28,000 viable organs fail to reach patients annually, with some of the worst OPOs recovering organs from only 25% of potential donors. These are not organs that are medically unsuitable — they are organs from patients who died in donation-eligible circumstances, in hospitals within the OPO's territory, where the OPO simply failed to identify the donor, failed to approach the family, or failed to coordinate recovery in time. So what? A Senate investigation by Senators Wyden and Grassley confirmed instances of waste, fraud, and abuse across OPOs nationally. Some OPOs gamed their performance metrics to appear compliant while underperforming on actual organ recovery. HHS investigators found that at one OPO, at least 28 patients may not have been deceased at the time of organ preparation, and 73 patients showed neurological signs incompatible with donation — suggesting that the pressure to hit numbers led to cutting corners on patient safety rather than genuinely improving recovery rates. This problem persists because each OPO holds a geographic monopoly — there is no competing OPO that a hospital or CMS can switch to. The regulatory framework was designed in 1984 and has barely evolved since. Cost-reimbursement was chosen because organ procurement was seen as a public service that should not be profit-driven, but the result is a system where the financial model actively insulates underperformers from accountability. The 2025 decertification of Life Alliance may signal a shift, but it took four decades and congressional pressure to achieve a single enforcement action.
Transplant surgery is one of the most demanding surgical specialties: surgeons are on call more nights per week than nearly any other specialty, cases arrive unpredictably at all hours, and the emotional weight of working with both dying donors and dying recipients is immense. A study of transplant surgery fellows found that 43.9% of nights involved sleep deprivation and 87.2% of days involved significant stress. The result: abdominal transplant surgery fellowships are among the least sought-after and least competitive fellowships for U.S. general surgery graduates. The field cannot attract enough trainees, and many who do complete training leave the specialty within their first few years of practice. So what? When a transplant surgeon leaves practice, they take years of specialized training with them. Each departing surgeon represents a reduction in transplant capacity — fewer organs accepted, longer wait times, more patients dying on the waitlist. The American Society of Transplant Surgeons formed a Pipeline Taskforce specifically to address the attrition crisis, acknowledging that early-career attrition is correlated with surgeon case volumes and case mix, meaning surgeons at lower-volume centers are most likely to leave. So what? The broader surgical workforce is already projected to face a shortage of 10,000 to 19,900 surgeons by 2036. Transplant surgery, as one of the least attractive subspecialties within an already-shrinking pool, will be hit disproportionately hard. As the population ages and the prevalence of kidney disease, liver disease, and heart failure increases, transplant demand is rising while the transplant surgeon workforce is contracting. The gap between need and capacity will widen. This problem persists because the fundamental working conditions of transplant surgery have not changed. Unlike elective surgery, transplant cases cannot be scheduled — they happen when organs become available, which is around the clock, seven days a week. The field has not adopted the shift-based or team-based coverage models that have reduced burnout in emergency medicine. Academic incentives favor research productivity over clinical volume, so transplant surgeons who do the most clinical work are often the least rewarded in academic promotion. The specialty is caught in a doom loop: bad working conditions drive attrition, attrition increases workload for remaining surgeons, increased workload drives more attrition.
A kidney transplant patient's waitlist clock begins when they are listed at a transplant center, not when they first develop kidney failure. The medically optimal path is preemptive referral — getting on the waitlist before starting dialysis — because every month on dialysis degrades the body and worsens post-transplant outcomes. Over 90% of Black patients are referred to transplant centers only after they have already started dialysis, compared to a significantly higher preemptive referral rate for White patients. Primary care and specialty physicians identify White patients as good candidates for transplant more often than they do Black patients, and Black patients are half as likely to be placed on the waitlist as White patients. So what? Because waitlist position is determined partly by accrued waiting time, late referral means Black patients start their waitlist clock later, even though their kidneys may have been failing for just as long. The result: Black patients face wait times that are, on average, a full year longer than White patients. This is compounded by the now-corrected but historically devastating eGFR race coefficient, which until 2021 inflated kidney function estimates for Black patients by 16-21%, masking severe kidney disease and further delaying referral and listing. So what? Each additional year on dialysis costs approximately $90,000, destroys cardiovascular health, and reduces post-transplant graft survival. Black Americans make up 13% of the U.S. population but over 33% of the kidney transplant waitlist — a disparity driven not by biology but by systematic delays at every step: later diagnosis, later referral, later listing. The OPTN has attempted to remedy the eGFR damage by backdating wait times for affected patients, but this only partially addresses a problem rooted in referral patterns, physician bias, and structural barriers to accessing transplant centers. This problem persists because the referral pipeline is decentralized and unregulated. No federal policy requires nephrologists to refer patients for transplant evaluation at a specific eGFR threshold. Each physician makes an individual judgment call, and implicit bias in those calls compounds across millions of clinical encounters. CMS's Increasing Organ Transplant Access Model attempts to address this by incentivizing transplant hospitals to identify underserved populations, but it is voluntary and does not reach the referring physicians who control the pipeline's entry point.
When a patient needs a kidney and has a willing living donor who is incompatible (wrong blood type or antibody mismatch), they can enter a kidney paired donation (KPD) program: their donor gives a kidney to a stranger whose donor is compatible with the original patient, creating a chain of swaps. In theory, KPD could dramatically increase living donor transplants. In practice, the U.S. kidney exchange market is fragmented across multiple competing registries — the OPTN/UNOS national program, the National Kidney Registry, the Alliance for Paired Kidney Donation, and individual hospital programs — that do not share patient pools. So what? A patient enrolled in one registry can only be matched against other patients in that same registry. If the perfect match exists in a different registry, neither patient knows, and neither gets transplanted. Research in operations science has shown that this fragmentation causes 30% to 63% fewer transplants than would occur with a unified national matching pool. Some patients enroll in multiple registries simultaneously, which creates a different problem: when one registry builds a chain involving that patient and the patient has already been matched elsewhere, the entire chain collapses, wasting weeks of coordination work and delaying transplants for everyone involved. So what? Each failed match means the patient stays on dialysis. Dialysis patients have a five-year survival rate of roughly 35%, compared to 85% for transplant recipients. Each month on dialysis costs Medicare approximately $7,500 per patient. The fragmented KPD system is not just an inefficiency — it is a body count. Patients are dying because registries compete instead of cooperate, and there is no regulatory mandate to unify the matching pools. This problem persists because each KPD registry is a separate organization with its own funding model, intellectual property in its matching algorithms, and institutional incentive to keep patients in its own pool. Registries compete for transplant center partnerships, and a center's choice of registry becomes a business relationship, not a medical one. There is no federal requirement to participate in a unified exchange, and OPTN's own KPD program is underfunded and underutilized relative to the private registries.
Every U.S. state has first-person authorization laws that recognize a deceased individual's documented decision to be an organ donor as legally sufficient — no family consent required. When you check 'organ donor' on your driver's license, that is supposed to be a binding legal document. Yet a survey of organ procurement organizations found that 20% would not proceed with organ procurement unless they had family consent, regardless of the deceased's registered wishes. Another 35% of OPOs had not procured organs against family objections in the previous five years. Overall, roughly 10% of potential donations from registered donors do not proceed because of family resistance. So what? This means that for a meaningful fraction of registered donors, checking the box on your license is a meaningless gesture. Your documented, legal wish to save lives after your death can be overridden by a family member's emotional objection — an objection that, in many cases, stems from misinformation about brain death, religious concerns that could have been addressed with proper communication, or simple shock and grief. The deceased person's autonomy is subordinated to the family's distress in the moment. So what? At an average of 3.5 organs per donor, a 10% override rate across thousands of potential donations means thousands of transplantable organs are lost every year — not because of medical unsuitability, not because of logistics, but because an OPO staff member chose not to enforce the law. Each lost organ is a patient who remains on dialysis, a child who does not get a heart, a person who dies waiting. This problem persists because OPOs face asymmetric consequences. If an OPO enforces the deceased's wishes over family objections, it risks bad press, complaints, and potential litigation from angry families. If an OPO defers to the family and lets organs go unrecovered, nobody knows, nobody complains, and the patients who would have received those organs never learn that a donor was available but overridden. The incentive structure rewards deference to grieving families over fidelity to the donor's wishes and the waitlisted patients' lives.
Donating a kidney requires 1-4 days of hospitalization and 4-6 weeks of recovery before returning to work. The recipient's insurance covers the donor's surgical costs, but it does not cover lost wages, travel to the transplant center, lodging, childcare, or follow-up care costs. Research shows that 80% of living kidney donors experience financial loss, with lost income accounting for 23% to 84% of total costs. Average donor-borne costs range from $900 to $19,900 in the first postoperative year alone. So what? A 2021 survey found that 80% of donors who received financial assistance from the National Living Donor Assistance Center said they would not have been able to donate without it. This means the current system filters out the majority of potential donors who cannot afford to lose six weeks of income. Construction workers, hourly retail employees, gig workers, single parents — the people most likely to know someone who needs a kidney (because kidney disease disproportionately affects lower-income communities) are the least able to donate one. So what? The United States has roughly 90,000 people on the kidney transplant waitlist right now. Each kidney transplant saves Medicare approximately $250,000 over five years compared to keeping that patient on dialysis. Living donor kidneys last longer and work better than deceased donor kidneys. Every potential living donor who is priced out of donating means another patient stays on dialysis at enormous cost to the healthcare system and enormous suffering to the patient. The financial barrier to living donation is not just a personal problem — it is a systemic bottleneck that costs lives and money. This problem persists because U.S. law (the National Organ Transplant Act of 1984) prohibits 'valuable consideration' for organ donation, and lawmakers have been hesitant to create robust reimbursement programs for fear of crossing the line into organ markets. The Living Organ Donor Tax Credit Act, which would provide a $5,000 federal tax credit, has been introduced repeatedly but never passed. The result is a system that asks people to make an extraordinary sacrifice — giving away a body part — while imposing financial punishment for doing so.
The system that allocates donor hearts to children uses just three priority tiers: Status 1A (most urgent), Status 1B, and Status 2. A child's tier is determined by their diagnosis type (congenital heart disease vs. cardiomyopathy) and the medications they receive — not by a granular assessment of how sick they actually are. A 2024 Stanford Medicine study found that these three categories show significant overlap in mortality risk, meaning a child classified as Status 2 can be sicker and closer to death than a child classified as Status 1A. So what? Within each tier, children are ranked by wait time, not by medical acuity. If two children are both Status 1A, the one who has been waiting longer gets the next available heart — even if the newer listing is far more critically ill. This means a less-sick child who was listed three months ago will receive an organ before a child who was listed yesterday and is actively dying. The system cannot distinguish between 'stable on medication but technically Status 1A' and 'will die this week without a transplant.' So what? Pediatric heart transplant has the highest waitlist mortality of virtually any organ or age group. While mortality has fallen from 21% to 13% in recent years, that improvement came from advances in medical care (better medications, better mechanical support devices), not from better allocation. The allocation system itself is still failing to route hearts to the children who need them most urgently. Children are dying on the waitlist not because there are no hearts, but because the hearts are going to less-sick children who happened to be listed first. This problem persists because pediatric transplant is a small field with limited political leverage. Adult heart allocation was reformed in 2018 with a six-tier system that more granularly stratifies medical urgency, but the pediatric system was left behind. The pediatric transplant community is small, the patient population is small, and there is insufficient advocacy pressure to force OPTN to overhaul a system that technically 'works' but distributes organs suboptimally.