Regulatory comment and research study · In preparation, 2026 · Docket FDA-2026-N-9505 · Clinical lead: Puja D. Patel, MD, Maimonides Medical Center · Data lead: Kurt E. Sather, API Automations
Overview
On September 25, 2026 the U.S. Food and Drug Administration issued its first system-level draft guidance for robotically-assisted surgical devices, Robotically-Assisted Surgical Devices - Premarket Submissions (GUI01500081). It sets out what a manufacturer should test and document before a robotic surgical system reaches the market: emergency stops and emergency patient access, motion scaling and tremor filtering, software fault messaging, bedside-assistant training, team-based human-factors validation, and a fixed list of clinical endpoints that includes conversion rate. Public comments close on November 24, 2026.
Whose voice reaches a device-guidance docket is itself one of the project's questions: it measures who comments on comparable dockets rather than assuming the answer. The project's own contribution is a public comment led by an obstetrics and gynecology physician, in which every recommendation is grounded in federal adverse-event data for gynecologic procedures, followed by publication of the analysis behind it.
Why It Matters
The draft guidance identifies the right hazards. It does not yet set a benchmark for how quickly a patient can be reached in an emergency, treat the competency of the bedside team as something to validate rather than describe, or require the post-market reporting fields that would let anyone compute the clinical endpoints it names. Real-world reports for gynecologic robotic surgery can show where each of those gaps has consequences. A comment that says so by guidance line number, with a reproducible number beside each claim, is one the reviewing division can act on.
Scope
- Adverse-event surveillance: medical device reports for the four robotic surgical device product codes (NAY, QNM, SAB, SAQ), filtered to gynecologic procedures, coded to the guidance's own hazard categories.
- Clearance and recall history: the 510(k) lineage and recall records for the same product codes.
- Who comments: the composition of public comments on comparable device-guidance dockets, and on this one as it closes.
- Trials and utilization: registered robotic gynecologic trials with posted results, Medicare utilization trends, and manufacturer-reported procedure volumes as the only usable denominator.
- Industry relationships: Open Payments records for the specialty, and each author's own declaration before submission.
Outputs
Each output is led by the author whose side of the work it reports, and both authors are on every output.
- The public comment to docket FDA-2026-N-9505, to be submitted before the November 24 deadline, structured by guidance section and line number with a proposed text change for each recommendation. Led by Dr. Patel, with Kurt E. Sather as co-author and clinician co-signatories. Published here once it is a public record.
- The analysis, as a preprint posted the same week, with the code, the coding schema, the audit sample and the derived dataset released alongside. Led by Kurt E. Sather, with Dr. Patel as co-author. Linked here by DOI.
- Clinical publications: a letter in minimally invasive gynecology, a clinical-opinion piece on credentialing and post-market surveillance, and society abstracts. Led by Dr. Patel, with Kurt E. Sather as co-author. Linked here when published.
The methods page describes the pipeline and the coding schema. The results page shows what has been computed from a dated snapshot and nothing else.
Where the project stands
As of November 9, 2026. This section changes as each stage closes. Nothing below is a result.
| Stage | Status |
|---|---|
| Design: the technical workbook, the coding schema and the analysis plan | Complete (workbook v2.4.0; coding schema v1.3.0 as registered, revised to v1.4.2 on November 9 through the registered audit gate, below). The guidance's Appendix A was read on October 9, 2026; the umbrella and covered grouping follows its Table 9, and the clinical lead confirmed it against FDA's definitions on October 16 |
| Pre-registration of the analysis plan on OSF | Registered on OSF on October 24, 2026: osf.io/xsy5z. The analysis plan, hypotheses and search strategy were fixed before any data were pulled. Any change the librarian's review of the search strategy brings is filed on OSF as a dated update |
| Systematic literature search (PubMed, Ovid MEDLINE, CINAHL) | Strategy revised October 16, 2026. The Maimonides Medical Center library does not license Embase or Scopus, so the clinical lead's searches run in Ovid MEDLINE and CINAHL Complete. The PubMed search was tested on October 17 and retrieves every earlier device-report study it is meant to find. A medical librarian peer-reviews the strategy. Embase and Scopus are searched only if a collaborator with authorized access can run them; otherwise they are reported as not searched |
| Extraction and analysis code | Built and self-tested. Extraction, cohort, a blind hand-audit sheet, the registered tests and the trend models |
| Data snapshot (adverse events, dockets, trials, utilization, payments) | Taken (snapshot of October 31, 2026, completed November 3). Every source passed the completeness check, and the archive's checksum is recorded. The adverse-event pull was reconciled against FDA's own MAUDE data files on November 8; the registration named FDA's public dashboard for this, but the dashboard has no device view, so the change is filed as a deviation. openFDA's newest report was received on September 30, 2026, which sets the end of the analysis window |
| Gynecologic cohort | Built on November 8, 2026 from the snapshot. It is rebuilt under the revised coding rules before any analysis |
| Blind hand audit of the coding rules | In progress. Kurt E. Sather hand-coded a stratified sample of 300 reports on November 8 and 9, blind to the rules' codes. The rules did not reach the registered threshold (Cohen's κ ≥ 0.7 on every coded field), so, as the registration requires, they were revised (coding schema v1.4.2), the change is filed as a deviation, and a fresh, disjoint sample is coded by hand next. The clinical lead, Dr. Patel, settled the two coding definitions the audit left open on November 9, and adjudicates the disagreements. No hypothesis is tested until the fresh sample clears the threshold |
| Analysis | Next, once the audit clears, ahead of the comment draft |
| Ethics determination | To be requested from Maimonides Medical Center before the preprint posts |
| Public comment to FDA-2026-N-9505 | To be submitted by November 20, ahead of the November 24 close |
| Preprint, dataset and code | Planned for the same week as the comment |
Research Team
- Puja D. Patel, MD, clinical lead. PGY-3 Resident, Department of Obstetrics and Gynecology, Maimonides Medical Center, Brooklyn, New York. Leads the clinical interpretation, the recommendations and the clinical publications, and adjudicates coding disagreements. ORCID 0000-0002-6659-9744.
- Kurt E. Sather, data lead. Director of Research, API Automations. Leads the data, the methods and the analysis, the preprint and the dataset, and is responsible for the integrity of the data. ORCID 0009-0008-3493-2961.
Affiliations are listed for identification only. The views expressed are the authors' own and do not represent Maimonides Medical Center.
Collaborate
Gynecologic surgeons, residents, bedside teams and professional societies who want to review or co-sign the comment can reach the research team at medical-research@apiautomations.com.