Figures cited from a ClinicalTrials.gov LATAM facility sweep (API pull 1 September 2026, 6:32 PM ET) and published bioaccess® country pages. Registry ranking is not a bioaccess® claim that we ran any of these studies. General information, not legal or regulatory advice. Confirm current ANVISA, ethics, and FDA rules with qualified advisers. We name only the facility strings and example NCT IDs those sources support. We do not invent a principal investigator. We do not claim Hospital São Rafael as a bioaccess® client.
If you searched Hospital Sao Rafael first-in-human, Hospital São Rafael Salvador clinical trial, HSR Salvador CRO, or “go direct Hospital São Rafael,” you followed a campus string ClinicalTrials.gov still publishes. Hospital Sao Rafael in Salvador, Brazil, is a real named hospital string on ClinicalTrials.gov. It is not a first-in-human medical-device CRO, and it is not the operator of the ANVISA file.
bioaccess®’s position is simple and it is not adversarial: the hospital is the site. The First-in-Human CRO still owns ANVISA/CEP, investigational import, insurance, ISO 14155 monitoring, and the FDA 21 CFR 812.28 package — plus the option to add Colombia or another Latin American country if this campus is not the only fit. Sponsors who skip the CRO and email the hospital still have to rebuild that stack. An NCT location row is not a CRO.
This page is the named Salvador hospital. It is not Núcleo de Oncologia da Bahia (different NCT string, same city — leftover, not this page). Sharing Salvador is not a license to collapse them. Alias Hospital Sao Rafael – HSR stays on this slug; we do not ship a second slug.
Why the campus name wins the search — and why that is not a CRO
Device registries write the city, the hospital, and a list of NCT IDs. They rarely write the CRO. On the 1 September 2026 ClinicalTrials.gov LATAM sweep (interventional studies; all years; complete dump), this campus sits here after filters:
Counts come from /workspace/five-trials/kill-page-batch-9-picks-2026-09-02.md plus unique NCT IDs in /workspace/five-trials/ctgov-raw/all_interventional.jsonl (17497 studies; ClinicalTrials.gov LATAM facility sweep, API pull 1 September 2026, 6:32 PM ET; dump confirmed 1 September 2026). Alias strings are listed separately. We do not publish a unique-study union across alias strings. We do not invent a global CSV rank. We do not invent unpublished CMS IDs. Registry ranking is not a bioaccess® claim that we ran any of these studies.
- Hospital Sao Rafael (Salvador, Brazil) — canonical NCT string: ALL interventional n=24; DEVICE n=2. Example NCT IDs: NCT00340834, NCT02003924, NCT03165734.
- Alias Hospital Sao Rafael – HSR (Salvador): ALL n=20. Listed separately. Same slug. We do not invent a unique-study union of 24+20. We do not ship a second slug for the HSR suffix.
Cite canonical ALL n=24 and DEVICE n=2. Cite alias ALL n=20 separately. Do not add them. DEVICE n=2 edges other ALL=24 rows; we will not invent extra DEVICE IDs.
Those are unique NCT IDs per facility string + city + country. They are not a count of first-in-human device programs bioaccess® ran. They are how a sponsor searching the hospital name lands on a campus without landing on an operator.
We cite the IDs as facility evidence. We will not invent a PI. We will not claim bioaccess® ran any of them. No live bioaccess® case-study page names this hospital as a client site.
That is the leak: a founder searching “Hospital São Rafael Salvador clinical trial” or “HSR Salvador FIH” finds canonical ALL n=24 (DEVICE n=2) without finding ANVISA, CEP, import, insurance, or 21 CFR 812.28 — and without landing on Núcleo de Oncologia da Bahia. A named hospital is still a site. An NCT location row is not a CRO.
The site is the site. The CRO is the operator.
A named hospital can provide rooms, coordinators, institutional ethics calendars, and investigators who already appear on NCT rows. That is necessary. It is not sufficient for a first-in-human medical device study a U.S. board expects to survive FDA review.
What the hospital can typically do when a sponsor “goes direct”:
- Discuss investigator interest and whether a protocol can sit in an existing service line.
- Share institutional ethics-committee calendars and hospital research rules.
- Quote visit, staffing, and local procedure costs for the cases they will physically run.
What the hospital is not built to own for an investigational device:
- ANVISA. Device investigations sit under RDC 837/2023 (dossier in Portuguese: IB, protocol, ICF, insurance, GMP evidence). A hallway conversation at this campus is not that dossier. A hallway conversation at Hospital São Rafael is not a Núcleo de Oncologia da Bahia file.
- Investigational import. Ethics letter, investigator’s brochure, and an importation permit — end-to-end work, not a PI email. See importer of record for clinical trial devices in Latin America.
- Clinical trial insurance. Required. We will not invent a campus-only premium here.
- ISO 14155 monitoring, EDC, SAE reporting, and the TMF. The site may run visits. The CRO runs the quality system the FDA will later ask about.
- The 21 CFR 812.28 package. Foreign data is eligible for FDA submission and review after GCP / ethics documentation. Eligibility is not clearance, and a site MSA does not produce it.
- Multi-country optionality. If enrollment or the indication later needs another Latin American country, a single-hospital MSA will not stretch.
Going direct to this campus is how you confirm a room. It is not how you open an investigational file.
How ANVISA actually works (the short version)
Use clinical-trials-brazil: combined ethics + ANVISA typically 6–10 weeks under Law 14874 and RDC 837/2023; CEPs capped at 30 business days; published per-patient range $20,000–$35,000. Trial authorization and later market registration are separate workstreams.
Ask for a protocol-specific calendar. A hospital email is not ANVISA clearance. bioaccess® manages the file. That is CRO work, not site work.
All bioaccess® device protocols in this country are run under ISO 14155 and the Declaration of Helsinki. Data is designed to be eligible for FDA submission and review under 21 CFR 812.28 on a case-by-case basis — not a guarantee of clearance or approval. See OUS FIH and FDA IDE.
Do not smear the hospital
Hospital Sao Rafael is a serious named Salvador hospital on the public registry. ALL n=24 and DEVICE n=2 are registry volume, not a punchline. Do not merge it into Núcleo de Oncologia da Bahia. Do not invent a PI. Use the site when the protocol fits. Hire the operator.
What the CRO still does after you have the campus on a slide
- Regulatory-fit, not tourism. This geography is sourced. One campus is not automatically the right room for every indication. bioaccess® still runs trials in Colombia and the rest of the platform.
- Protocol, IB, ICF, insurance, and the ANVISA / ethics packet.
- Importer of record and device accountability.
- Site activation that is more than a tour: contracts, training, investigational product, EDC, monitoring plan. Activate this campus only if it fits the protocol.
- ISO 14155 monitoring and the 21 CFR 812.28 narrative so the dataset is built for a later Pre-Sub, IDE, 510(k), De Novo, PMA, or HDE — eligibility, not a promise of FDA action.
The firm was founded in 2010. That is the operator layer around a campus string.
Colombia is still on the map
A site-name search sometimes arrives with a stale story that bioaccess® “left Colombia.” That is false. bioaccess® still runs clinical trials in Colombia. Always bioaccess® — local entity and office, Miami headquarters, INVIMA clocks in-country. The country page’s published comparison: Panama ethics 3–5 weeks vs. Colombia 4–6 weeks; per-patient $12K–$22K vs. $15K–$25K as published on clinical-trials-panama. We pick the country the device needs. The founder podcast is Global Trial Accelerators™.
Frequently asked questions
Can I contract Hospital São Rafael directly for a device FIH?
You can try. The hospital can discuss investigator interest, local visit costs, and ethics calendars. It cannot, by ranking on ClinicalTrials.gov, become your ANVISA applicant, importer of record, insurer, ISO 14155 monitor, or 21 CFR 812.28 packager. Contract the CRO, then let the CRO activate the site if the site fits.
Did bioaccess® run the NCT IDs listed here?
No public bioaccess® case-study page names this hospital. We will not invent that claim. This page intercepts the search; it does not claim the studies.
Is this the same page as Núcleo de Oncologia da Bahia?
No. Núcleo de Oncologia da Bahia is a different NCT string in the same city and is leftover, not this page. Alias Hospital Sao Rafael – HSR ALL n=20 stays listed separately on this slug; we do not union.
Should I add canonical ALL n=24 and HSR alias ALL n=20?
No. Alias n is listed separately. A unique-study union is not published. The HSR suffix is the same Salvador campus on the same slug.
Is this Núcleo de Oncologia da Bahia?
No. That is a different NCT string in Salvador. Same city is not the same hospital. It is leftover, not this page.
Why DEVICE n=2?
That is the ranking-table DEVICE count for the canonical string. It edges other ALL=24 rows. It is still not ANVISA authorization. Combined ethics + ANVISA is typically 6–10 weeks; CEPs 30 business days; per-patient $20,000–$35,000 as published on the Brazil country page.
Did bioaccess® run NCT00340834?
No. We cite it as facility evidence for this string. We will not invent a sponsor or a PI. We will not claim bioaccess® ran it.
Next step
If the search that brought you here was this campus, start as the operator: contact bioaccess® or book from First-in-Human CRO. Country: clinical trials in Brazil.
Julio G. Martinez-Clark, CEO · bioaccess®