Evidence
What we know, and how we know it.
Published literature for the problem. Our own cadaver and pilot data for the solution, labeled as ours. We never present one as the other.
The problem
| Figure | Claim | Basis |
|---|---|---|
| 20–25% | Humeral IO attempts that fail, attributable in substantial part to landmark and angle error | Reades 2011, Prehosp Emerg Care; Wampler 2012, Am J Emerg Med |
| Preferred route | The humeral head is the preferred adult resuscitation IO site, for medullary flow and central venous transit | Paxton 2009, J Trauma; Reades 2011 |
| Survival signal | Access route and reliability affect drug delivery timing and outcomes in out-of-hospital cardiac arrest | Vallentin 2024, N Engl J Med 391(26):2480–2490 |
| -$300 | Kit cost wasted per failed attempt | Innovo estimate from list pricing of consumables |
Ask us and we will send the citation behind any claim on this page.
Cadaver study IOP-002
Without a targeting aid, nobody hit the site. With one, everybody did.
Thirteen active EMS providers, two fresh cadaver specimens, one humeral IO attempt each on an EZ-IO driver. Placement was judged by a blinded assessor against a calibrated anatomical overlay.
Unaided
0%
first-pass success · 0 of 6
With I-Optic
100%
first-pass success · 7 of 7
13
Providers: 6 EMT-Basic, 7 Paramedic. None had used the device before.
30 sec
The entire training: one verbal briefing. No demonstration, no practice.
10 / 10
Ease of use, rated by every guided participant after their attempt.
0
Adverse events, needle fractures or soft tissue injuries across all 13 attempts.
The finding that matters
Half the unaided providers drilled through bone. In the wrong place.
Three of the six unaided providers achieved cortical penetration. They felt the click. They got the resistance change that normally signals success. And not one of them was in the target zone.
That is the whole argument for the device. These providers can run an EZ-IO. What they could not do, under the conditions of the study, was find the greater tubercle. In a real resuscitation, a provider in that position infuses into a medullary cavity that does not reach the central circulation, believes they have access, and the patient does not get the drug.
More repetitions on a mannequin do not fix this. The I-Optic does, by turning a recall task into a perception task: not "where is the greater tubercle on this patient," but "where is the laser."
Study at a glance
IOP-002 is an internal proof-of-concept cadaver study run by Innovo, not a controlled clinical trial, and not a claim of clinical efficacy. Thirteen participants and two specimens is too small for powered efficacy estimates, and both specimens were elderly with similar BMI and no shoulder pathology. A larger multicenter trial is the next step. The I-Optic Trainer is intended for skills training.
Results in full
| Outcome | Unaided (n=6) | I-Optic (n=7) |
|---|---|---|
| First-pass success | 0 / 6 (0%) | 7 / 7 (100%) |
| Cortex penetrated | 3 / 6 (50%) | 7 / 7 (100%) |
| Needle tip on target | 0 / 6 (0%) | 7 / 7 (100%) |
| Pre-attempt confidence | 1.67 / 10 | 3.57 / 10 |
| Ease of use | — | 10.00 / 10 |
| Adverse events | 0 | 0 |
Results held at both provider levels: 2 of 2 EMT-Basic and 5 of 5 Paramedic guided participants placed correctly; 0 of 4 and 0 of 2 unaided did.
Earlier pilot, IOP-001
100%
First-attempt placement
Our first pilot: fifteen participants, ten with medical training and five without, with no prior exposure to the device. Every one placed on the first attempt.
The finding worth noting is not the percentage. It is that the lay cohort performed the same as the trained cohort, which is what you would expect if the reticle and the locked vector are doing the work experience normally does. IOP-002 was designed to test that on real tissue, against a control.
Next step