Intellectual Frontiers runs its own research programs in healthcare delivery, medical devices, clinical trials, insurance and digital authorization, payments and marketplaces, and education and training. We own what those programs produce, from the published papers through to the patents and the companies built on them. And on every one of them we ask the same question about AI: which entirely new forms of work become possible?
The easy version of the AI conversation is the same everywhere: do today's work faster, or do it with fewer people. That is worth real money, and it is also where most of the discussion stops.
The question we ask across all five units is which entirely new forms of work become possible. A speedup makes an existing job cheaper. A new form of work creates a job nobody was staffed to do, and that is where the patents, the positions, and the companies come from.
A referral system that reads the record, ranks the specialists who can actually take the case, attaches the imaging, and tells the sending clinician what will be asked for next. Nobody had staff for that before, because nobody could read every chart.
A device program where the evidence for a submission assembles itself out of live telemetry as the product runs, so the clinical argument is standing before the study is written. Faster regulatory drafting saves weeks. This changes what a submission is.
A trial that runs as software: eligibility checked continuously against live records, deviations flagged the hour they happen, and the audit trail assembled while the study runs instead of reconstructed after it.
Step 1
We find inventions in the middle of ordinary work: a hospital that re-keys the same patient data into three systems before discharge, a referral that arrives without the images, a trial coordinator who keeps a paper binder because the system cannot be trusted. Somebody notices, and somebody writes it down that week.
Step 2
Then we try to kill it. An authorization idea gets run against real access rules and real record requests. A device idea gets run past the people who have to file it with the FDA. If the idea does not survive that, it stops there, and we write down why.
Step 3
What survives goes to one person who can act on it. File the patent, publish the disclosure, put money behind it, or start the company. That person puts a date on the decision. Nothing sits in a folder waiting for a committee.
Step 4
The claim language written for a hospital interoperability filing gets reused in the trials filing. The test harness built for device telemetry gets pointed at metric queues. That is what makes the second one cheaper than the first.
Hospital and clinic systems: referrals, care coordination, task and work list scheduling, patient navigation, risk management, and the interfaces between the EHR and everything else. Granted and pending filings cover this ground.
Software inside and around regulated devices: connected instruments, remote monitoring, ventilator and imaging control, and the data that has to move off the device and into a chart without breaking the device's clearance.
Digital native trials management: running a study as software instead of a binder, with eligibility, evidence, and traceability produced while the trial runs.
Self-controlled authorization to records and service delivery evidence: who is allowed to read what, for how long, and what proof survives the transaction.
Digital payment systems, marketplace networks that settle among many participants, crowdsourced credentialing and accreditation, and the rating and reputation signals that make any of it trustworthy.
Precision education and training, rewardable content delivery, and the metric queues, task lists, and traceable telemetry that tell an operator whether a system is actually working.
IP
Patents and licenses covering clinical workflow, medical device data, clinical trials operations, records authorization, payments and marketplaces, training, and system telemetry.
Capital
Money behind ideas already tested against real provider, device, insurer, and platform data.
Press
Books, fieldbooks, essays, and plain-language patent summaries for people who have to run the systems.
Studios
Operating companies built around proven assets, funded by customers with real budget lines.
Network
The intelligence layer on human capital: who matters for a specific thesis, why they might matter, and what the next useful step is.
Pending · Filed 2024-12-11
Digital Native Trials Management SystemGranted · US11880882
Computer-Controlled Marketplace Network For Digital TransactionsPending · Filed 2024-11-27
Collaborative, Continuous, And Comprehensive Risk Management SystemAn authorization backlog nobody can explain. Device data that will not make it into the chart. A trial that reconciles for six weeks after the last patient visit. A marketplace that cannot prove what it settled. Describe what is happening and how it was measured, and we will tell you within two weeks whether there is an invention in it.
Start a conversation