For how we live.
A final FDA guidance treats the on-screen output of software sent with a prescription drug as a type of drug labelling, and says how the prescribing information should describe it and when sponsors submit it for review.
In 143 people with poorly controlled type 1 diabetes, a system that prompts injections through glucose-sensor alerts was non-inferior to standard monitoring — and acting on the alerts drove the benefit.
The FDA logged the recall of the RED 62 reperfusion catheter, part of the Penumbra System used to pull clots from the brain, on 7 October. Manufacturing problems mean the device may fracture; 43 units are affected.
A final order published 9 October classifies the over-the-counter infant pulse rate and oxygen saturation monitor as class II, with special controls built around the risk that it fails to detect a pulse.
A final order published 9 October classifies systems that let people collect their own specimens for non-viral STI testing as class II, moving them out of the automatic class III that demands premarket approval.
A final order codifies Class II status for high-throughput DNA sequencing systems that detect germline mutations in cancer genes, built on Invitae's panel. It becomes a predicate other makers can clear against.
A final order codifies a new Class II category for systems that destroy soft tissue with sound rather than heat, built on HistoSonics Edison System. It becomes a predicate other makers can clear against.
A final order published 17 September denies a petition that sought to exempt four classes of computer-aided detection, diagnosis and triage devices from 510(k) clearance. Makers must keep filing before they market.
A Swiss investigator-led trial tested whether isolating the back wall of the left atrium on top of pulsed-field pulmonary vein isolation cuts arrhythmia recurrence. The primary answer was no.
The FDA logged the recall of 8,939 StatStrip Glucose and B-Ketone meters on 23 September, after a software flaw let a scanned barcode linger and file a reading to the wrong patient. It is the agency's most serious tier.
The FDA logged the recall of 2,718 Medline convenience kits on 23 September: sterile applicators in some lots could grow the mould Aspergillus penicillioides, and contaminated skin-prep risks infection or sepsis.
A single-surgeon randomised trial of 180 knees found small but statistically supported gains for robot-assisted functional alignment over conventional mechanical alignment at two years.
The FDA logged the recall of 1,640 AVID single-shot epidural trays on 23 September, because they hold sodium-chloride ampules caught in a separate Class I sterility recall that warns of possible infection.
The FDA logged the recall of 5,232 Automated Impella Controllers on 23 September, citing a possible purge-pressure sensor failure that can leave the console unable to recognise the pump's purge cassette.
Two 2026 Class I recalls cover Omnipod 5 insulin pods whose soft cannula can be damaged in manufacturing, letting insulin leak instead of reaching the user. The FDA ties under-delivery to a risk of DKA.
A final order classifies the flushing and storage solution for vascular autografts as Class II with special controls, codifying a 2023 De Novo decision on Marizyme's DuraGraft and setting the bar for future copies.
In 3,000 patients eligible for blood thinners, the device maker's trial found left atrial appendage closure noninferior to anticoagulant drugs, and superior for non-procedure-related bleeding at three years.
A new draft guidance sets out the agency's current thinking on the non-clinical and clinical data it expects in premarket submissions for robotically-assisted surgical devices. Comments are due 24 November.
Rejoyn, the first prescription digital therapeutic cleared for major depression, beat a sham app on a backup analysis. Its pre-specified primary endpoint fell just short of significance.
A January 2025 draft guidance lays out the documentation the agency expects across an AI device's whole life, including bias checks across demographic groups and postmarket monitoring. It is not yet binding.
A 40-person Harvard trial aimed accelerated magnetic stimulation using each patient's own connectivity scan. Depression scores fell further than with standard scalp measurement, though the trial was small.
The Essilor-designed HAL lenses cut two-year myopia progression by 71% versus ordinary glasses in 159 American children. The company employs the study's authors, and the result backs the device's US clearance.
A federal rule now requires certified electronic health records to publish a standard set of facts about each predictive algorithm they ship, including how it was validated and whether it was tested for fairness.
EndeavorRx is a prescription game the FDA cleared in 2020 for children with ADHD. Its pivotal trial moved a computer attention score, but the FDA says it may not help the behaviour parents notice.
Software that interprets cardiotocography in labour lifts clinicians' accuracy in reader studies. The largest randomised trial of computerised interpretation, in 47,062 women, found no benefit for mothers or babies.
RelieVRx is an authorised virtual-reality therapeutic for chronic low back pain. Its trials show real improvement — but in a 1,067-person study the edge over a placebo headset was 0.4 points on a 10-point scale.
Under Article 6, AI that is or sits inside a device already needing an independent safety check counts as high-risk. A 70-study review finds the trouble is overlap with existing device law.
The Hypotension Prediction Index warns anaesthetists before blood pressure falls. Three 2026 trials find a plain mean-pressure alarm does the same job, and the algorithm's benefit tracks how hard clinicians treat.
Abilify MyCite embeds a sensor in an antipsychotic tablet to log each dose to an app. Nearly a decade on, its own FDA label still says it has not been shown to improve whether patients take their medicine.
A Cochrane review of 51 trials found powered brushes cut plaque and gum inflammation more than manual ones, on moderate-quality evidence. An 11-year cohort suggests the small edge adds up: fewer teeth lost.
The RAPID pathway would issue a proposed national coverage decision the day a device clears the FDA, cutting a process that normally runs 9 to 12 months. The trade-off is evidence.
A final order files AI tools that flag possible cardiovascular disease as class II with special controls, letting similar triage software clear via 510(k). The controls require real-world, subgroup-tested validation.
Implanted electrodes now decode attempted speech into text for people who cannot speak. Results are striking — up to 62 words a minute — but come from single participants, need brain surgery, and remain research.
Digital adherence gadgets promise to fix forgotten doses. In a 53,480-person randomised trial, low-cost reminder devices did not raise adherence at all. Text reminders did more, but on short, self-reported trials.
A closed-loop system links a glucose sensor to an insulin pump so software doses automatically. In the pivotal 6-month trial it raised time in target range by 11 points over a standard pump — a bounded gain, not a cure.
An autonomous AI for diabetic retinopathy hit 87.2% sensitivity in primary care. A smartphone acuity test agreed with a reference chart to within 0.07 logMAR. Both are triage, not diagnosis.
Commercial software reads an X-ray for breaks about as well as a radiologist. A prospective paediatric emergency study found that accuracy did not translate into fewer missed-fracture recalls.
In the 500-man PRECISION trial, MRI-guided biopsy detected clinically significant cancer in 38% of men versus 26% with the old blind biopsy — and spared 28% a biopsy altogether.
TRIC-I-HF randomised 360 older patients with severe tricuspid regurgitation. Adding transcatheter repair to medical therapy cut a composite of death or heart-failure hospitalisation through three years.
Alpha-blockers ease symptoms within weeks; 5-alpha-reductase inhibitors shrink the gland over months. In a 3,047-man trial, combining them cut progression risk by 66%. Newer office procedures spare sexual function.
A predetermined change control plan is the FDA's answer to a model that keeps learning: pre-approve a bounded set of future changes, and the manufacturer can ship them without returning for a fresh review.
Sensors and reminder inhalers get patients to take more doses — one trial lifted adherence from 30% to 84%. That this translates into meaningfully better control is real but a smaller effect.
Systems that give a diagnosis with no clinician in the loop post high sensitivity in trials. In real-world use, one in six images was too poor to grade — and the strongest numbers come from company-run studies.
A prescribed home apnea test can diagnose clear-cut cases, and one under-mattress sensor flagged severe apnea with 86% sensitivity — but both miss milder disease and events a lab would catch.
Toilet seats and mirrors that read your heart and blood pressure are real engineering. The published validation is mostly small, single-lab or company-run — and outcome evidence is absent.
Self-testing and self-managing warfarin cut clot events by roughly 40%, and self-management also lowered deaths — one of the few at-home tests with randomised evidence of real benefit.
A deep-learning system drew tumour targets and organs at risk automatically. It saved time and improved consistency — but oncologists still reviewed and edited the contours before treatment.
A Polish study found endoscopists' unaided polyp-detection fell after AI arrived. A Norwegian trial that tracked doctors before, during and after AI use found no such deskilling.
OSPREY implanted every participant, then randomly decided whose device was switched on. That design answers a question most implant studies leave open, and it has limits the authors state plainly.
Green-light pulse sensors raise a real concern about melanin, but independent tests find the skin-tone effect small next to device and movement errors — and the validation studies barely include darker skin.
In 264 patients in China, the device maker's trial reported clinical worsening in 25.7% versus 51.5% on medical therapy alone (HR 0.49) — but no sham procedure blinded patients.
A researcher who expected the evidence base to be thin says even she was surprised by how thin. Most cleared devices never appear in a registered clinical trial at all.
A new discussion paper proposes a two-axis risk framework and a physician-training analogy for evaluating generative AI devices. It is not a rule, and the agency is asking what one should look like.
The Aletta still needs a trained supervisor in the room, but it locates the vein, inserts the needle and fills the tubes on its own — and one person can now watch three of them at once.
Emulating a trial inside Veterans Health Administration records, researchers estimated adjusted mortality risk ratios of 0.90 to 0.84 for starting continuous glucose monitoring in older-onset type 1 diabetes.
RACER-Knee masked patients with sham incisions and found a Forgotten Joint Score difference of −1.5 points at 12 months (P=0.62) — while the robotic operation cost more.
It found no meaningful difference at six weeks. It enrolled 64 people, tested one device, and ran for six weeks. That is the entire head-to-head evidence base three years after the category opened.
Where AI clot-detection was tested head-to-head, it still missed more emboli than radiology trainees, especially small peripheral ones. Its clearest win was cutting time to diagnosis.
A study of 19,539 patients echoed the randomised CONVINCE trial — but it was observational, and several of its authors work for the dialysis company that owns the data.
An umbrella review of 42 reviews found only two wearable measures — heart rate and atrial fibrillation — with comparably strong support. Getting any device into a trial is harder than buying it.
These features are verifiable and occasionally decisive. The one study of Medical ID in trauma found most unconscious patients had no usable emergency info on their phone at all.
A new structured dataset finally makes the approvals countable. Two trials published weeks apart show why counting them says little: 0.86 fewer migraine days in one, a 20-point symptom shift in the other.
The agency granted its first authorization for a software-aided adjunctive diagnostic device in wound assessment, a category built for tools that analyze a wound optically.
In the FDA study, self-fitted AirPods matched a professional fitting for perceived benefit. Independent testing finds the hearing test accurate to about 10 dB — and the amplification short of a fitted aid.
It is a wellness feature, not a cleared pulse oximeter, and it inherits pulse oximetry's known skin-tone bias. A patent fight also stripped it from US watches for about 19 months.
Most hiccups stop on their own, and the remedies everyone swears by have almost no trial evidence. The one gadget with published data was studied by its own inventor and funded through the product crowdfunding campaign.
It is a medical-device biocompatibility standard, now built largely on lab-grown skin rather than animals. It measures whether a material irritates. It says nothing about whether a product works.
A content analysis of every FDA cybersecurity safety communication from 2013 to 2025 finds a small corpus, rising over time, and 94% of the flagged vulnerabilities rated severe.
The clearance is the third in a growing family of machine-learning notification algorithms built to spot serious cardiac conditions from a routine 12-lead ECG, a test most patients already get.
A new audit of China's regulatory record finds a market concentrated in radiology, dominated by deep learning, and clustered in four cities. It also finds the approval curve flattening.
reSET, Somryst and EndeavorRx cleared a real regulatory bar, on real trial data. The pivotal ADHD game moved an attention score by 0.88 of a point. Clearance is not the obstacle the category keeps hitting — payment is.
The FDA cleared it to sort one rhythm from one other. It cannot detect a heart attack, and about one reading in eight comes back inconclusive.
An 11,018-measurement study in 24 English intensive care units found false negative rates up to 35.3 percentage points higher in patients with darker skin tones. A neonatal study published two days later did not.
A transformer model trained on 17,119 scans beat expert and non-expert examiners on every metric and cut simulated referrals by 63%. The catch is who built it and the validation the rest of the field still lacks.
One agency asked how to measure AI devices after they are deployed. Another asked how to speed adoption up. The offices that would answer the first question are losing staff.
A systematic review of every AI device authorised from 1995 to June 2024 found 97% cleared through the 510(k) pathway, which does not require independent clinical data on performance or safety.
A census of 691 AI-enabled devices cleared through 2023 found demographic data missing almost everywhere, three summaries reporting patient outcomes, and 113 recalls driven mostly by software.