26 / 26
claims audited
4%
error rate so far (1 corrected, 0 contested)
Synthetic blood starts to affect health care around 2020
Grader A
No synthetic blood is approved in the US/EU; lab-grown RBCs reached only a first-in-human trial (RESTORE, Nov 2022, 5-10 ml). Hemopure (HBOC-201) is registered for routine use in South Africa and Russia and given under US expanded access, a niche real use far below 5% of transfusions.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: transfusion patients
Grader B
No substitute is approved in the US or EU; Hemopure is approved only in South Africa and Russia and sees niche use. The first lab-grown red cells were transfused to volunteers in the RESTORE trial in Nov 2022. Real use exists only as niche approvals and a first-in-human pilot, far below 5% of transfusions.
Test: D37 health: >=5% of transfusions use a substitute by 2022 = hit; real use below 5% (pilots, niche approvals) or 5% in 2023-2025 = partial; no real use by 2025 = miss.
Tissue regeneration starts to affect health care around 2020
Grader A
Real routine use exists: bioengineered skin substitutes for diabetic foot and venous ulcers (Medicare-covered), MACI cartilage implants (FDA 2016, >5,000 US cases). No source shows 5% of relevant procedures; MACI is a tiny share of cartilage repair. Partial; could be a hit under a broad skin-substitute reading.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with tissue defects (wounds, cartilage)
Grader B
Approved regenerative products are in routine use: MACI autologous chondrocyte implants (FDA, re-approved through 2024) and cellular skin substitutes for chronic wounds. No source found showing they reach 5% of the relevant procedures by 2022; use is real but niche.
Test: D37 health: >=5% of relevant procedures (e.g. cartilage repair, chronic wound care) use regenerative tissue products by 2022 = hit; real but smaller use = partial; none by 2025 = miss.
Artificial vascular system starts to affect health care around 2020
Grader A
The first engineered human vessel (Humacyte Symvess) was FDA-approved Dec 2024 and launched 2025, after humanitarian use in Ukraine; use is a tiny share of vascular procedures. Real use within 5 years of 2020, below 5%.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: vascular repair/bypass procedures
Grader B
The first tissue-engineered vessel (Humacyte Symvess) was FDA-approved in Dec 2024 for extremity vascular trauma, inside the 5-year window. Use in 2025 is niche (one indication, early launch). Conventional synthetic Dacron/ePTFE grafts predate the poster and are not what the label means.
Test: D37 health: >=5% of vascular grafting procedures use engineered vessels by 2022 = hit; real use below 5% or 5% in 2023-2025 = partial; no real use by 2025 = miss.
3D-printed drugs starts to affect health care around 2020
Grader A
Spritam (levetiracetam), the first 3D-printed drug, was FDA-approved 2015 and sold from 2016. It and a few later products are niche; 3D-printed drugs are a negligible share of prescriptions.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: dispensed prescriptions
Grader B
Spritam (levetiracetam), the first 3D-printed drug, was FDA-approved in 2015 and marketed; hospital pilots of printed personalised doses followed. Use is real but a tiny share of prescriptions.
Test: D37 health: >=5% of dispensed drugs or pharmacies by 2022 = hit; real but niche use = partial; none by 2025 = miss.
Stem-cell treatments starts to affect health care around 2020
Grader A
Approved novel stem-cell products are few and niche: Holoclar (EU 2015), Alofisel (EU 2018, withdrawn 2024), Ryoncil (FDA Dec 2024). Real use well under 5%. HSCT was routine long before 2012, so the poster's forward-looking reading excludes it; under that reading it would be a hit.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with the indicated conditions
Grader B
Approved regenerative stem-cell products exist (EU Holoclar, Alofisel; US Ryoncil approved Dec 2024, Omisirge, Lantidra) and are used, but in narrow indications and small numbers. Haematopoietic stem-cell transplant was routine long before 2012 and is not the poster's forward-looking reading; under that reading the claim would be an earlier hit, noted as a minority reading.
Test: D37 health: >=5% of patients in the target indications receive stem-cell therapy by 2022 = hit; real but niche approved use = partial; none by 2025 = miss.
Personalized medicine starts to affect health care around 2020
Grader A
Personalized medicines were 39% of FDA new molecular entities in 2020 and over a third in three of four years; biomarker-selected therapy is routine in oncology. Well above 5% by 2020.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: new therapies and patients in oncology
Grader B
Personalized medicines made up 34% of FDA new molecular entity approvals in 2022 and at least 25% for eight years running; biomarker-guided oncology treatment and companion diagnostics are routine. Well above 5% by 2022.
Test: D37 health: >=5% of relevant treatments or new therapies are biomarker-guided by 2022 = hit.
Robotic surgery starts to affect health care around 2020
Grader A
In Michigan hospitals the share of elective general surgery done robotically rose from 1.8% (2012) to 15.1% (2018). Above 5% before 2020.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: surgical procedures
Grader B
In 73 Michigan hospitals the robotic share of general surgery rose from 1.8% (2012) to 15.1% (2018), above 5% before the placed year. The remote (telesurgery) sense is not required by the label.
Test: D37 health: >=5% of relevant surgical procedures are robot-assisted by 2022 = hit.
App-driven diagnostics starts to affect health care around 2020
Grader A
Population survey: about 13% of German adults had used a symptom-checker app for self-diagnosis; Apple Watch and Kardia AF-detection apps FDA-cleared 2018. Above 5% by about 2022. Uncertain whether symptom checkers count as diagnosis.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: adult population/patients
Grader B
A July 2022 representative-style German survey found 6.5% of adults had used symptom checkers; FDA-cleared app diagnostics (Apple Watch ECG AF detection, 2018; KardiaMobile) were in wide consumer use. Just above the 5% line by 2022; margin is thin.
Test: D37 health: >=5% of relevant users (adults seeking diagnosis) used app-based diagnostics by 2022 = hit.
Natural language processing starts to affect health care around 2020
Grader A
Nuance Dragon Medical, clinical speech recognition, reported 500,000 clinicians, 90% of US hospitals and 80% of radiologists by 2020; already >10% of US physicians in 2008. Hit (already here).
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: clinicians/hospitals
Grader B
Nuance Dragon Medical clinical speech recognition was used by over 550,000 physicians and Nuance solutions by 77% of US hospitals (July 2022). Clinical NLP for documentation and coding was routine before 2022.
Test: D37 health: >=5% of clinicians or hospitals use clinical NLP/speech understanding by 2022 = hit.
AR surgery assistance starts to affect health care around 2020
Grader A
Augmedics xvision, the first FDA-cleared AR surgical headset (2019), did its first US case in 2020 and reached 5,000 then 10,000 spine cases by April 2025 across 26 states; a tiny share of the hundreds of thousands of annual US spine fusions. Real use below 5%.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: surgical procedures (e.g. spine fusion)
Grader B
Augmedics xvision was FDA-cleared Dec 2019, first US case June 2020, and reached 10,000 cumulative procedures only in April 2025; other AR systems are similarly small. Real routine use in some hospitals, far below 5% of spine surgeries (well over a million a year in the US).
Test: D37 health: >=5% of relevant (e.g. spine, navigated) procedures use AR guidance by 2022 = hit; real use below 5% = partial; none by 2025 = miss.
Virtual triage starts to affect health care around 2020
Grader A
NHS England told all GP practices to adopt remote 'total triage' via online and phone consultation in March 2020; by July 2020, 79% of UK practices were live with online consultations (5% pre-pandemic). NHS 111 online triage also national. Well above 5%.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: primary care practices / urgent care contacts
Grader B
England's NHS 111 triages about 16-20 million calls a year with the NHS Pathways clinical decision-support software (a registered medical device), a large share of urgent-care contacts for a population of ~56 million; COVID-era online and video triage expanded this. Above 5% by 2022 (and already before).
Test: D37 health: >=5% of urgent-care contacts in a health system are triaged remotely by software by 2022 = hit.
Telemetrics starts to affect health care around 2020
Grader A
Remote monitoring of pacemakers and ICDs is standard of care: nearly two-thirds of US Medicare CIED beneficiaries had remote monitoring encounters, and enrollment exceeds 85% in recent cohorts. Well above 5% before 2022.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with monitorable devices or chronic conditions
Grader B
In 211,346 Medicare beneficiaries with new cardiac implantable devices (2015-2020), roughly two-thirds had at least one remote-monitoring encounter (over one-third had none); CMS added RPM billing codes in 2019. Well above 5% by 2022, though adherence to recommended frequency was low.
Test: D37 health: >=5% of relevant patients (e.g. implanted cardiac device holders, chronic-condition patients) are remotely monitored by 2022 = hit.
Non-invasive glucose sensors starts to affect health care around 2020
Grader A
As of February 2024 the FDA had authorized no non-invasive glucose device and warned against wearables claiming it; CGMs that took the market are minimally invasive. Early CE-marked attempts (e.g. GlucoTrack) did not reach routine use. No real use by 2025.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: people with diabetes
Grader B
No non-invasive glucose monitor is FDA-authorised; in Feb 2024 FDA warned against smartwatches and rings claiming it. Diabetes care moved to minimally invasive CGMs (filament under the skin), which do not meet the label. A few CE-marked devices (e.g. GlucoTrack) never reached routine use, and GlucoTrack pivoted to an implantable sensor.
Test: D37 health: >=5% of glucose-monitoring diabetics use it by 2022 = hit; real use below 5% = partial; no real use by 2025 = miss.
Open health records starts to affect health care around 2020
Grader A
ONC: 38% of US individuals accessed their online medical record or portal in 2020, 57% in 2022; the Cures Act information-blocking rule took effect 2021. Well above 5%.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients
Grader B
About 3 in 5 US individuals were offered and accessed their online medical record or patient portal in 2022 (ONC), backed by the 21st Century Cures Act information-blocking rule (2021) and FHIR app access. Far above 5%.
Test: D37 health: >=5% of patients access their records online / via apps by 2022 = hit.
Epidermal sensors starts to affect health care around 2020
Grader A
Epidermal-electronics spinouts reached clinical products: Sibel's ANNE soft wireless sensors got FDA clearances (latest 2023) and were evaluated in NICUs; use is pilots and early adoption, far below 5% of monitored patients. A broad reading that counts adhesive ECG patches (Zio) could give a hit, but those are not the thin epidermal electronics the label names.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: monitored patients (e.g. NICU, inpatients)
Grader B
The Rogers-lab line reached clinic: Sibel Health's ANNE skin sensors gained FDA 510(k) clearances (ANNE One, May 2023, for neonates to adults; ANNE Maternal) and are deployed in some hospitals and low-income-country NICUs. Use is real but limited to pilots and early deployments, well under 5% of monitored patients. Conventional adhesive ECG patches are not the thin epidermal electronics the label names.
Test: D37 health: >=5% of relevant monitored patients (e.g. NICU, inpatients) use epidermal sensors by 2022 = hit; real clinical use below 5% = partial; none by 2025 = miss.
Question answering computing systems starts to affect health care around 2020
Grader A
Watson for Oncology was deployed at a limited number of hospitals, criticised for unsafe advice, and Watson Health was sold off in 2022. Real use below 5% by 2022; LLM-based clinical QA only spread in 2023-2025, which is also partial under D37.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: clinicians/hospitals
Grader B
Watson for Oncology had niche hospital deployments and was wound down when IBM sold Watson Health in 2022, so by 2022 use was well under 5%. LLM-based clinical QA then crossed the line: OpenEvidence was used by about 40% of US physicians by mid-2025 (60% claimed by Sept 2025). 5% reached within 3-5 years after 2020 = partial.
Test: D37 health: >=5% of clinicians use such systems by 2022 = hit; real use below 5% by 2022, or 5% reached 2023-2025 = partial; no real use by 2025 = miss.
Ingestible sensors starts to affect health care around 2020
Grader A
The forward-looking product, Proteus ingestible sensor in Abilify MyCite (FDA 2017), saw limited uptake and Proteus went bankrupt in 2020 (assets sold to Otsuka). Niche real use, well below 5%. Capsule endoscopy predates 2012 and is not what the poster forecast.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients on medication / GI diagnostics
Grader B
Sensor-in-pill adherence (Proteus, Abilify MyCite) reached market but was abandoned (Proteus bankrupt 2020; MyCite discontinued Oct 2024). Capsule endoscopy and motility capsules remain routine niche procedures (over 4 million PillCams sold) but are a small share of GI diagnostics and predate the poster. Real, niche use = partial; reading only the adherence-sensor line would give miss.
Test: D37 health: >=5% of relevant diagnostic procedures or medication-adherence patients use ingestible sensors by 2022 = hit; real but niche use = partial; none or abandoned by 2025 = miss.
Rapid gene sequencing starts to affect health care around 2020
Grader A
In Flatiron data, NGS testing in advanced NSCLC rose from 5% (2012-13) to 35% (2018-19); rapid genome sequencing also used in NICUs. Above 5% by 2020.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with indicated cancers or genetic disease
Grader B
In a US cohort of 63,294 patients, NGS after metastatic diagnosis in 2021 reached 34.5% (breast) to 62.9% (colorectal) and 61.4% in advanced NSCLC; NGS-based prenatal cfDNA screening and rapid NICU genome sequencing are also routine. Far above 5% by 2022.
Test: D37 health: >=5% of relevant patients (e.g. advanced cancer, rare disease, prenatal screening) receive sequencing-based diagnosis by 2022 = hit.
At-home sensors starts to affect health care around 2020
Grader A
55% of US adults aged 50-80 with hypertension or related conditions owned and used a home BP monitor (2022 survey); home pulse oximetry spread during COVID. Well above 5%; arguably already here in 2012.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with chronic conditions
Grader B
Self-measured home blood-pressure monitoring was used by about 61% of US adults with hypertension in 2019 (20 states) and by ~37-39% monthly in NHANES 2009-2014; home glucose meters, pulse oximeters and home sleep tests are routine. Above 5% before the placed year.
Test: D37 health: >=5% of relevant patients use home diagnostic sensors by 2022 = hit (earlier counts).
Data-driven patient communities starts to affect health care around 2020
Grader A
PatientsLikeMe, the reference case, had several hundred thousand members and fed research before UnitedHealth bought it in 2019; real use exists, but no measure shows 5% of patients in health care pooling data through such communities. Partial; a broad reading counting any online patient group could argue a hit.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with chronic disease
Grader B
PatientsLikeMe, the archetype, has run since 2005 with ~830,000 members, about 2% of US MS patients; it was bought by UnitedHealth in 2019 and runs on as an Optum-backed company. Real, routine use but below 5% of patients in its conditions. General online health forums (higher reach) do not pool structured health data, so are not the label's sense.
Test: D37 health: >=5% of patients with a given condition take part in data-pooling communities by 2022 = hit; real use below 5% = partial; none by 2025 = miss.
Tissue-embedded sensors starts to affect health care around 2020
Grader A
New implanted biosensors reached routine but niche use: Eversense implantable CGM (FDA June 2018), CardioMEMS PA-pressure sensor (FDA 2014); both a small share of their populations. Implantable loop recorders are routine but predate the poster and record electrical rhythm rather than tissue biochemistry. Partial.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with diabetes or heart disease
Grader B
Fully implanted monitoring sensors are in routine but niche use: Eversense implantable CGM (FDA June 2018), CardioMEMS pulmonary-artery sensor (2014) and insertable cardiac loop recorders. Each serves a small slice of its population; most CGM users wear transcutaneous (not implanted) sensors. No evidence of 5% share by 2022.
Test: D37 health: >=5% of relevant patients (e.g. CGM users, heart-failure or arrhythmia monitoring) use implanted sensors by 2022 = hit; real use below 5% = partial; none by 2025 = miss.
Hybrid assisted limbs starts to affect health care around 2020
Grader A
HAL had 300 suits in 130 Japanese medical and care facilities by 2012, EU medical CE mark in 2013, later Japanese medical approval and FDA clearance (2017). Routine niche use, far below 5% of rehab facilities or patients.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: rehabilitation facilities/patients
Grader B
HAL received EU medical certification in 2013, Japanese medical-device approval with insurance coverage for neuromuscular diseases in 2015, and FDA clearance in 2017; it is used in a few hundred facilities. Real routine use, small share of rehab patients.
Test: D37 health: >=5% of relevant rehabilitation patients or facilities use HAL-type suits by 2022 = hit; real use below 5% = partial; none by 2025 = miss.
Myoelectric prosthesis starts to affect health care around 2020
Grader A
Myoelectric prostheses have been fitted for decades; in a national VA survey 46% of recent-conflict veterans with upper-limb loss used myoelectric devices. Well above 5%; already here in 2012.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: upper-limb amputees
Grader B
Myoelectric prostheses have been fitted routinely for decades; in a national study of US veterans with upper-limb amputation, 46% of recent combat veterans and 22% of Vietnam veterans used myoelectric devices. Far above 5%, already at the poster's date.
Test: D37 health: >=5% of upper-limb amputees fitted with prostheses use myoelectric devices by 2022 = hit (earlier counts).
Exoskeletons starts to affect health care around 2020
Grader A
Ekso, ReWalk and others are FDA-cleared; EksoNR is in about 375 rehabilitation centres worldwide (300+ units in 30 countries by 2021). Routine niche use; no evidence of 5% of rehab facilities or patients globally. US inpatient rehab share might approach 5%, so this is uncertain.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: rehabilitation facilities/patients
Grader B
Medical exoskeletons (ReWalk FDA 2014, Ekso, Indego, HAL) are in routine use in rehab centres, but volumes are small: Ekso, the leading clinical vendor, booked only 100 devices in all of 2022 on $12.9M revenue. Real niche use; no evidence of 5% patient share.
Test: D37 health: >=5% of relevant rehabilitation facilities or patients (SCI, stroke) use exoskeletons by 2022 = hit; real use below 5% = partial; none by 2025 = miss.
Augmented olfaction starts to affect health care around 2020
Grader A
As of 2024-25 no olfactory prosthesis or implant exists; work is proof-of-concept (electrical bulb stimulation, a trigeminal substitution device in Science Advances 2025). No real use by 2025.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with smell loss
Grader B
Olfactory augmentation remains lab research: olfactory displays and an implant-style smell-restoration prosthesis (VCU) are at prototype or early human-study stage. No product in routine clinical or consumer health use by 2025. Electronic noses that diagnose disease do not extend the human sense and are not counted.
Test: D37 health: >=5% of relevant users by 2022 = hit; real routine use below 5% = partial; no real use in care by 2025 = miss.
Neuroprosthetics starts to affect health care around 2020
Grader A
About 736,900 cochlear implants worldwide by Dec 2019 (183,000 in the US), plus routine deep brain stimulation. Cochlear implantation is standard care for profoundly deaf children; above 5% of the relevant population. Already here; brain-controlled prostheses alone would be partial.
Test: D37 health: share of relevant patients, providers or procedures using it by 2022 (hit >=5%; partial <5% real use by 2022 or 5% in 2023-2025; miss no real use by 2025); relevant: patients with severe-to-profound hearing loss and similar indications
Grader B
The label explicitly includes cochlear implants: about 736,900 implanted worldwide by Dec 2019 (118,100 US adults, 65,000 US children), standard care for severe-to-profound deafness, with deep-brain stimulation also routine. Above 5% of the eligible pediatric population in high-income countries well before 2022. Brain-controlled prostheses alone would be partial (trials, Neuralink first implant 2024).
Test: D37 health: >=5% of relevant patients (e.g. severe-profound deafness) receive neuroprostheses by 2022 = hit (earlier counts).