Hospital and research collaborations are often the fastest route to clinical signal, clinical sites, and credibility for a MedTech startup. But under MDR, manufacturer obligations under Article 10 stay with you regardless of who runs the study, who owns the data, or whose logo is on the paper. The quality of your collaboration agreement determines whether the partnership accelerates your CE mark or creates a file your Notified Body cannot accept.

By Tibor Zechmeister and Felix Lenhard.

TL;DR

  • MDR Article 10 keeps all manufacturer obligations with the legal manufacturer, regardless of who executes the clinical work.
  • Any clinical investigation conducted to generate evidence for CE marking must follow Articles 62 to 82 and Annex XV.
  • The sponsor under Article 62 is the entity that takes responsibility for the investigation. Startups are usually the sponsor, even when the hospital runs the study.
  • Data ownership, IP allocation, and publication rights must be specified in writing before the first patient is enrolled.
  • Investigator-initiated studies still feed evidence into your CER, but only if they meet Annex XV quality standards and you can use the data under your agreement.

Why this matters

Every early-stage MedTech founder eventually gets the same offer. A clinician at a university hospital is interested in the device. She has patients, she has colleagues, she has a research ethics committee contact, and she wants to run a study. The conversation feels like a gift. It often is. But it is also the moment where the regulatory and commercial structure of your entire clinical evidence strategy gets decided in a handshake.

The mistakes are predictable. Startups let the hospital take the sponsor role to save paperwork, then discover the data cannot be used in their CER because the consent form did not authorise commercial use. Startups sign a collaboration agreement that lets the university keep all IP on any algorithmic improvement made during the study, then cannot commercialise the improved version. Startups treat an investigator-initiated study as free clinical data, then find the methodology does not meet Annex XV and their Notified Body refuses the evidence.

Open innovation works in MedTech. It has launched plenty of successful products. But it works when the startup understands what MDR requires of the manufacturer and writes collaboration agreements that protect those requirements.

What MDR actually says

Article 10 — manufacturer obligations. The manufacturer is responsible for the design, manufacture, packaging, labelling, clinical evaluation, technical documentation, post-market surveillance, and conformity assessment of the device. These obligations cannot be transferred to a hospital, a university, or a research partner. If you are the legal manufacturer placing the device on the EU market, Article 10 applies to you and only you.

Articles 62 to 82 and Annex XV — clinical investigations. Any clinical investigation conducted to demonstrate conformity with the GSPR must comply with Article 62 and the general requirements of Annex XV Chapter I. The investigation plan requirements are set out in Annex XV Chapter II. Ethics committee review, informed consent under Article 63, adverse event reporting under Article 80, and sponsor obligations under Article 71 and Article 72 all apply.

Article 62(2) — the sponsor. The sponsor is the individual, company, institution, or organisation that takes responsibility for the initiation, management, and setting up of financing of the clinical investigation. In a startup-hospital collaboration intended to generate CE-mark evidence, the startup is almost always the sponsor because the startup is the one taking responsibility for initiating the investigation and funding it. The hospital and the investigators run the study operationally, but the sponsor role sits with the manufacturer.

Annex XV Chapter I §2.7 and §3 — the clinical investigation plan and the investigator's brochure must be prepared under the sponsor's responsibility. The sponsor is accountable for the scientific quality and ethical conduct of the investigation.

EN ISO 14155:2020+A11:2024 is the harmonised standard for good clinical practice in clinical investigations of medical devices for human subjects. Using this standard provides a presumption of conformity with the relevant parts of MDR Articles 62 to 82 and Annex XV. Your hospital partners will typically already know this standard. Many university research infrastructure teams are built around it.

Two provisions are often misread. First, MDR does not prohibit investigator-initiated studies. But if you want to use the data from an investigator-initiated study in your clinical evaluation under Article 61, the study must still meet the methodological and ethical standards that make the data admissible. Second, the fact that a study is conducted at a university hospital does not automatically make it compliant with Article 62. Compliance depends on how the study is designed, approved, consented, and documented.

A worked example

A startup builds an AI-based triage tool for emergency radiology. A senior radiologist at a teaching hospital wants to run a prospective study on 500 patients comparing the tool against standard reading. She drafts a study protocol, submits to her local ethics committee, and asks the startup for a letter of support.

Path A — handshake collaboration. The startup provides the software and a letter of support. The hospital runs the study. Twelve months later the results are published in a good journal. The startup asks to include the data in its CER. The Notified Body asks: who was the sponsor? Who was responsible for the investigation plan? How was informed consent worded? Was this a clinical investigation under Article 62 or a retrospective data analysis? The consent form only authorised academic use. The protocol was not written to Annex XV Chapter II structure. The Notified Body accepts the publication as supporting literature but not as primary clinical evidence. The startup has spent a year and has to run a second study.

Path B — structured collaboration. Before the study starts, the startup and the hospital sign a clinical investigation agreement. The startup is the sponsor under Article 62(2). The hospital is the investigation site. The principal investigator is the radiologist. The clinical investigation plan is written jointly and follows Annex XV Chapter II and EN ISO 14155:2020+A11:2024. Informed consent explicitly covers use of the data for CE-mark purposes. The startup holds the data, grants the hospital a publication right after a pre-agreed embargo, and retains all IP on device improvements. Twelve months later the same 500 patients generate a dataset the Notified Body accepts as primary clinical evidence.

Same patients. Same science. Different paperwork. Different outcome.

The Subtract to Ship playbook

Step 1 — decide sponsor role first. Before anything else, decide whether the startup is the sponsor under Article 62(2) or whether this is genuinely an independent academic study you will cite as third-party literature. Both are legitimate. Mixing them up is not.

Step 2 — use EN ISO 14155:2020+A11:2024 as the common language. Hospital research offices know this standard. Building the investigation plan around it avoids reinventing processes and gives you a presumption of conformity route.

Step 3 — write the collaboration agreement before the protocol. The collaboration agreement covers sponsor role, data ownership, IP allocation, publication rights, budget flow, insurance, liability, and authorship. The protocol comes after. Do not let enthusiasm flip this order.

Step 4 — put data ownership in writing. The startup should own the clinical dataset it paid for. The hospital typically retains rights to publish and to use the data for non-commercial academic purposes. Specific wording matters: "sponsor owns all data generated under the investigation" is a clean starting point.

Step 5 — handle IP at three layers. Background IP brought by each party stays with that party. Foreground IP created during the study on the device, the algorithm, or the use case typically belongs to the startup because it is tied to the device. Independent academic contributions, such as new clinical insights, belong to the investigator. Write these three layers explicitly into the agreement.

Step 6 — align informed consent with intended use. The consent form must allow the patient data to be used for the purpose the startup actually needs, which includes regulatory submission and post-market surveillance. A consent form written only for academic research is a time bomb.

Step 7 — keep Article 10 visible on your side. The hospital runs the study, but you remain the manufacturer. PMS planning under Article 83, risk management under EN ISO 14971:2019+A11:2021, and technical documentation under Annex II all stay with you. The collaboration agreement should reinforce this, not blur it.

Reality Check

  • Have you decided in writing whether you are the sponsor under Article 62(2) for each ongoing hospital collaboration?
  • Does your collaboration agreement specify data ownership, IP allocation, and publication rights, or are these covered only verbally?
  • Does the informed consent form authorise use of the data for CE-mark purposes and post-market surveillance, not just academic research?
  • Is the clinical investigation plan structured to Annex XV Chapter II and EN ISO 14155:2020+A11:2024?
  • If your hospital partner published tomorrow, could the data still be used in your CER under Article 61?
  • Have you documented how Article 10 manufacturer obligations remain with you despite operational execution by the hospital?
  • Do you know which of the three IP layers (background, foreground, academic) each improvement made during the study falls into?
  • If the collaboration ended tomorrow, would you walk away with a usable dataset and a clean IP position?

Frequently Asked Questions

Can a hospital be the sponsor of a clinical investigation intended for our CE mark? Technically yes, but it is unusual and rarely advisable. If the hospital is the sponsor under Article 62(2), the hospital takes responsibility for the investigation plan, sponsor obligations under Articles 71 and 72, and data ownership. Startups lose control over the evidence they need for their own conformity assessment.

Can we use data from an investigator-initiated study in our CER? Yes, if the data meets the methodological quality required by Article 61 and Annex XIV, the consent form authorises the intended use, and you have a legal basis to use the data. Notified Bodies assess investigator-initiated data as one data source among several.

Who owns the algorithm improvements made during a hospital study? This depends entirely on the collaboration agreement. Without an agreement, default IP law in the relevant jurisdiction applies, which is rarely in the startup's favour. A clear foreground IP clause assigning device improvements to the startup is standard in structured collaborations.

Do we need an ethics committee approval even for a small study? Any clinical investigation under Article 62 requires ethics committee review under Article 62(4). National rules add additional requirements. Even small feasibility studies should go through ethics review if they involve patients and generate data intended to support CE marking.

Is a memorandum of understanding enough? No. A memorandum of understanding signals intent but does not allocate rights. For any collaboration that will generate CE-mark evidence, you need a full clinical investigation agreement with schedules covering protocol, data, IP, insurance, and publication.

What if the hospital wants to own the data? Then you have a choice. Either the hospital is the sponsor and you cite the data as third-party literature, or you walk away and find a site that agrees to sponsor-owned data. For evidence you will rely on in your CER, sponsor-owned data is the safer path.

Sources

  1. Regulation (EU) 2017/745 on medical devices, consolidated text. Articles 10, 61, 62, 63, 71, 72, 80, 83. Annexes II, XIV, XV.
  2. EN ISO 14155:2020+A11:2024 — Clinical investigation of medical devices for human subjects — Good clinical practice.
  3. MDCG 2023-7 (December 2023) — Clinical investigation exemptions under Article 61(4)-(6).