C-SATS

The product name states the mechanism honestly and the grade follows from it. Crowd Sourced Assessment of Technical Skills means the assessment is performed by people: vetted expert surgeons and a crowd of trained reviewers scoring video against validated instruments. Machine learning does the surrounding work, removing patient identifying data from the video and segmenting the procedure into steps, with AI assisted insights layered on the dashboard.

That is real and useful engineering, and it is not the thing being bought. What a health system purchases is expert human judgment delivered at a scale and speed that internal peer review cannot match. Graded C, and the candour is credited: this vendor could easily have marketed the human review as an AI scoring engine and does not.

There is no autonomous decision anywhere in this product, and that is a design choice rather than an absence of ambition. Scores are produced by humans against published, validated instruments, feedback returns confidentially to the surgeon, and no algorithm rates anyone. In a domain where an automated judgment about a named surgeon's technique could affect privileges, credentialing and litigation exposure, keeping the human as the assessor is the correct architecture.

The confidential return path is the second deliberate choice, since it separates development from evaluation and is what makes surgeons willing to upload video at all. The residual question a buyer should still ask is what happens when an institution rather than an individual holds the subscription, and whether aggregate scores can be surfaced to leadership.

Transparency here sits in an unusual place: the scoring method is published and the model is not. Assessment runs on validated published instruments, notably GEARS, the Global Evaluative Assessment of Robotic Skills, which means a buyer can read exactly what is being measured and how it has been validated in the literature. That is more methodological transparency than any peer in this category offers.

The unpublished half is the machine learning, specifically the automated removal of patient identifying data from surgical video and the segmentation into procedural steps. No accuracy or error rate is published for the de identification step, and that is the number that matters most, because a de identification failure discloses PHI from an operating room. Publishing the de identification error rate would move this to A.

The de identification is engineered into the pipeline rather than left to reviewer discretion, and the reason that matters is the reviewers. Machine learning removes patient identifying data from surgical video before review, and the review is performed by an external crowd rather than by institutional staff, so the control sits at exactly the point where the content leaves the institution's own governance.

A structural control ahead of an external step is a materially better design than a policy instructing reviewers to disregard what they see. An audited third party certification stands behind the wider posture rather than a self declaration, and compliance is stated against multiple regimes rather than the domestic one alone.

The input is the most sensitive material any vendor in this category handles: video from inside a patient's body, tied to a named surgeon on a known date, which identifies two people at once. Held below the top grade on two counts. The de identification accuracy is unpublished, so the control is credible in design without being measurable from outside, and it is the single point on which the whole architecture rests. And the reviewer chain itself is unnamed, with nothing stating who the external reviewers are, where they sit, what governs their access or what they may retain. Ask for both.

The strongest evidence base in this category, resting on academic origin rather than marketing. The platform came out of the University of Washington in 2014 with surgeons, engineers and biostatisticians, the assessment instruments are independently validated in the surgical literature, and the vendor reports a cross sectional analysis in which surgeons in the top GEARS quartile outperformed other groups.

Technical skill is described as correlated with downstream outcomes including blood loss, complication rates and 30 day readmissions, and deployments are named and varied, spanning Northwell, Providence St. Joseph, AdventHealth, Hackensack, Valley Health, UC Irvine urology, the University of Washington and a US Air Force evaluation.

What holds it at B is that the causal claim is not established: correlating skill scores with outcomes shows the instrument measures something real, not that using the platform improves outcomes. No controlled study of the intervention itself was located.

The best PHI posture located in this category, and it needs to be, because the input is the most sensitive data any vendor here handles: video from inside a patient's body, tied to a named surgeon on a known date. Three things earn the grade. The platform is HITRUST CSF Certified, which is an audited third party certification rather than a self declaration. Compliance is stated against HIPAA, HITECH and GDPR rather than HIPAA alone.

And de identification is engineered into the pipeline, with machine learning removing patient data before review rather than relying on reviewer discretion, which matters because the reviewers are an external crowd. The gap noted in the transparency row still applies: the de identification accuracy is unpublished, so the control is credible in design without being measurable from outside.

HIPAA and HITECH compliance are stated explicitly, GDPR is addressed for international use, and the underlying platform is HITRUST CSF Certified, which is a substantive foundation rather than a compliance claim. What was not located is the contractual layer: no BAA template, scope description or execution process is published, and that gap is more consequential here than for most vendors in this category because surgical video is transmitted to reviewers outside the institution.

A buyer needs to know whether crowd reviewers sit inside the BAA, whether they are subcontractors of the vendor, and what happens to video after review. Publishing the BAA scope and the reviewer data handling terms would move this to A.

HITRUST CSF Certified, stated plainly and repeatedly across the product pages, with HIPAA, HITECH and GDPR compliance named alongside it. HITRUST is the certification the index has treated as the top tier elsewhere, and holding it puts C-SATS in a very small group across the whole roster. Against the broader pattern the index has documented, where workforce vendors and even healthcare cybersecurity vendors publish no attestation at all, this is the clearest counterexample in the category.

Corporate ownership plausibly explains it, since a Johnson and Johnson subsidiary inherits enterprise compliance infrastructure a startup would struggle to fund, which is worth noting as a genuine benefit of the acquisition alongside the independence cost recorded elsewhere in this grid.

C-SATS is an educational and performance management platform rather than a medical device, so no FDA pathway applies and the missing clearance is not the reason for this grade.

The C reflects the exposure that does apply and is unaddressed publicly: discoverability, and it is the sharpest regulatory question in this vein. Video of a specific operation, scored against a validated instrument and tied to a named surgeon, is exactly the artefact a plaintiff would want in a malpractice action. Whether it is shielded depends on state peer review privilege statutes, which vary considerably and were not written with externally reviewed video in mind, and the involvement of outside crowd reviewers may complicate a privilege claim that assumes an internal committee.

Confidential return of feedback to the individual surgeon is a real architectural mitigation and it is credited. It is not a published position on privilege, which is what would move this grade. Buyers should obtain counsel's view on privilege in their own state before institutional rollout, not after.

The bias argument is made and then not evidenced. C-SATS positions crowd sourced review as removing the bias present in traditional internal peer review, which is a real and defensible argument, since an internal committee assessing a colleague carries obvious conflicts that an anonymous distributed panel does not.

But no inter rater reliability data, reviewer demographic composition, or analysis of score variation by surgeon characteristics was located, and crowd rating introduces its own bias questions rather than eliminating bias as a category. The unaddressed structural issue is ownership.

The platform belongs to Johnson and Johnson MedTech, which sells surgical robotics and instruments, so an assessment system that shapes how surgeons are judged sits inside a company with a commercial interest in which platforms they operate on. The index raised the same neutrality question when PathAI was acquired by Roche and treated Proscia's independence as a competitive advantage. Publishing inter rater reliability and a statement on assessment neutrality across device platforms would move this to B.

The scoring instrument is published and externally validated, which is unusual and matters more here than in most records because of who is being assessed. Skill evaluation runs on validated published rating scales described in the peer reviewed literature, so a surgeon can read exactly what criteria they are being judged against and how those criteria were themselves validated, and can contest a score on the instrument's own terms rather than arguing with an opaque number.

That is more methodological transparency than any peer in this category offers, and it places the standard outside the vendor's control. The affected party is a named surgeon whose intraoperative video is assessed and whose results feed professional development and potentially credentialing conversations, so contestability is not academic.

What is unpublished is the machine learning half, specifically the automated removal of patient identifying information from surgical video and the segmentation into procedural steps. No accuracy or error rate is published for the de identification step, and that is the number that matters most on this record, because a de identification failure discloses patient information from inside an operating theatre to an external review panel.

No warranty, indemnity or remediation commitment was located. Ask for the de identification error rate, what a surgeon can see and dispute about their own assessment, and who has access to their scores.

No EHR integration exists and none is required, since the data source is the operating room video stack rather than the clinical record, so the axis is assessed on the interoperability surface that applies. On that surface the design is deliberately open: video capture is described as device agnostic from any minimally invasive surgery platform, which is a meaningful commitment given that the owner sells its own surgical systems and could have restricted capture to them.

That openness is the practical answer to part of the neutrality concern raised in the governance row. What is not published is the integration mechanics, including which capture hardware is supported, whether ingestion is automated or manual, and how video reaches the platform from an OR without disrupting the case.

The model is a private, HITRUST CSF Certified cloud case library per institution, with video uploaded, de identified, segmented and then distributed to reviewers. GDPR compliance is stated and the reviewer network extends to Asia Pacific, with an expert reviewer named at Severance Hospital in the Yonsei University Health System, which implies cross border handling of surgical video.

That is the disclosure gap: no residency statement, no named regions, and no description of where video sits while a distributed international reviewer panel works on it. For an EU or UK institution that is the determining question, and it is not answered publicly. Naming the storage regions and the reviewer jurisdiction rules would move this to A.

Better than the category norm and still short of a price. C-SATS states its charging mechanism openly, describing a subscription model available to individual surgeons and to entire hospital systems, which is more than most vendors in this category disclose and is credited here, since knowing the unit lets a buyer reason about scaling. No rate, tier or figure is published at either level.

The two commercial questions a buyer should ask are whether pricing is per surgeon or per case reviewed, since those scale very differently for a high volume robotic programme, and whether continuing education credit issuance is included or separately charged. Publishing a per surgeon or per case rate would move this to B.

Coverage is broad within surgery and confined to it. The platform spans robotic, laparoscopic and open procedures across multiple specialties, with named use in urology, arthroscopy and general minimally invasive surgery, and it serves both ends of the career: residency programmes at the University of Washington and Valley Health System, applicant technical skill evaluation before resident selection at UC Irvine urology, and practising surgeons at Northwell, Providence St. Joseph, AdventHealth and Hackensack.

Reach extends internationally through the Asia Pacific reviewer network and through a Florida Chapter of the American College of Surgeons collaboration. The limit is that this is a surgeon product, so a health system buying it addresses one high value professional group rather than the frontline workforce that dominates the rest of this category, and the ownership by a surgical device manufacturer is a standing consideration recorded in the governance row.