The research operating system: why Europe's hospitals need one, and why now

Two operations, one operating system

Every university hospital, research institute, and specialist centre runs two operations. The first is care. The second runs alongside it: clinical trials, investigator-initiated studies, registries, biobanks, imaging archives, and the research data they produce.

Care is gaining its operating systems. Hospitals are adopting platforms that connect their clinical systems and carry AI services, and European capital now treats that layer as strategic infrastructure.

Research has no operating system of its own. It runs on whatever each sponsor sends, and Europe is paying for that.

Europe is losing commercial trials

Between 2013 and 2023, global commercial trial starts grew by 38 per cent. Over the same decade, the European Economic Area's share fell from 22 per cent to 12 per cent. Commercial trials with at least one EEA site included 286,159 patients in 2018 and 226,155 in 2023, a fall of more than a fifth. Trial starts fell in almost every EEA country (EEA, 2024)Source, which points to a structural problem rather than a local one.

The industry's own assessment locates much of that structure at the hospital. The IQVIA analysis for EFPIA and Vaccines Europe found site start-up and recruitment to be the longest steps in the trials it studied, and called for better site readiness, simpler contracting, and ready-to-go networks.

The burden sits with the site

European research staff describe a workload built around data and documents. In a national survey of Italian clinical research coordinators and data managers, 78 per cent considered their workload excessive. About four in five did data entry in most of the studies they managed, and a similar share kept the investigator site file. (Italy, 2025)Source Swedish clinical research nurses rated entering and checking case report form data as their most frequent data management task. (Sweden, 2022)Source

Administration costs time as well. In one Italian public hospital, studies that needed a contract averaged 100 days to activate, against 46 days for studies that did not. (Italy, 2024)Source In a UK study of a low-risk staff survey across 20 NHS trusts, it took a median of 147.5 days from first local contact to first participant, but only 6.5 days once the trust had agreed. (United Kingdom, 2024)Source The delay sat in the organisation.

Industry surveys, mostly North American, show how technology adds to this. In one 2023 survey of more than 500 research professionals, nearly 70 per cent reported six or more logins per study, and 55 per cent rated set-up and training on sponsor technology extremely or very burdensome. (North America, 2023)Source Europe has not measured its own version of these figures, and it should.

Why the tools have not fixed it

Most clinical trial software is designed around the sponsor. It collects data for one study, owned by one company, and it leaves when the study closes. Even site tools offered free of charge tend to connect sites to a sponsor's ecosystem rather than give the hospital infrastructure of its own.

Hospitals that want control face an awkward choice. They can accept a vendor-hosted service with limited say over where and how it runs, or host open tools themselves and take on validation, security, and support.

The result is a research operation with many tools and no operating layer. A sponsor's system runs its own trial. The hospital has to balance staff, rooms, participants, and compliance across every study at once.

Europe now measures execution

European policy has moved from harmonising rules to measuring results. ACT EU aims for two thirds of trials to recruit within 200 days of application by 2030; in the first quarter of 2026, 40.5 per cent did.Source Authorisation is shortening at the same time: a proposed 75 days for multinational trials, 26 days for national assessment in Germany, 14 in France's fast track.Source

As authorisation gets faster, the limiting step moves to the hospital. We set out the evidence in The clock has moved to the hospital.

The data duty arrives on a fixed date

The European Health Data Space makes hospitals health data holders. From March 2029, holders must describe their datasets for a public catalogue and supply data covered by a permit within three months. Approved users work inside secure processing environments rather than receiving patient-level files. From March 2031, clinical trial data, genomic data, and research cohorts enter the same regime.Source

A trial management tool cannot meet those duties on its own, and neither can a data warehouse that knows nothing about the trials that produced its data. From March 2031, the trial record and the research data belong on one governed core. Our readiness guide for health data holders sets out the duties in detail.

Execution after the signal

Care-side platforms are starting to generate trial candidates. Electronic record systems now offer feasibility and recruitment modules, ambient AI tools surface trial opportunities during the consultation, and national services search millions of records for eligible participants.

Finding candidates will get easier and cheaper. What happens after the signal will not: protocol intake, feasibility governance, ethics-approved outreach, coordinator queues, consent, visit scheduling, source capture, transfer to the sponsor's data system, monitoring access, and an audit trail an inspector will accept. That is where trials succeed or stall, and it is the layer no one owns.

The thesis

Hospitals should own the operating layer of their research, as they are coming to own the operating layer of their care.

A research operating system is the hospital's own platform for all of its research. It is used across every sponsor and every investigator-initiated study, holds the hospital's research data on the hospital's terms, and carries the AI services research needs. We define it by six requirements.

1. Owned by the hospital. Bought and governed by the institution, used across all sponsors, and still there when each study closes.
2. One place to work. Start-up, investigator site file, data capture, visits, participant communication, and monitoring in one system, with hospital sign-on in place of a new login per study.
3. Inspection-oriented by design. Audit trails and electronic signatures built to ICH GCP, EMA Annex 11, and 21 CFR Part 11 expectations, so every layer above can be trusted.
4. Data under hospital control. Deployable on infrastructure the hospital controls, with the research data governed by the institution and EHDS-aligned: described, traceable, and ready for permitted secondary use.
5. Open to AI and to other signals. AI services work on the hospital's own records. Pre-screening surfaces potential matches for investigator review, on the hospital's own recordsIn pilot, and candidate signals from other systems arrive in the same queue, with an investigator making every decision.
6. Measured. Every step from first contact to close-out is timestamped, so a hospital can show sponsors, funders, and networks its own pace.

What changes

For hospitals, research becomes an institutional capability instead of a cost run through other organisations' systems, with a research data asset the hospital controls. For sponsors and research organisations, it means sites that can show how fast they activate and recruit. For patients, it means more trials available close to home. For Europe, it means research infrastructure that competes on speed and quality while keeping sensitive data under European governance.

Why we are building ZINOS

ZINOS is building the research operating system for hospitals: a Clinical Site Operating System that runs every study in the hospital, a Scientific and Research Data Hub that holds the hospital's research data on infrastructure it controls, and AI services on top as each is validated.

We are not starting from a blank page. The platform underneath ZINOS was developed from 2012 inside clinical hospitals, with the investigators and nurses who run the trials, and rebuilt in 2026 as one core for the operating system and the data hub. It was designed from the start to run inside the hospital and to serve the hospital rather than a single sponsor. It is developed by MIA Solutions, the engineering partner behind ZINOS.

The care half of the hospital has found its operating system. The research half is next.

Sources

1. IQVIA for EFPIA and Vaccines Europe, Assessing the clinical trial ecosystem in Europe, October 2024.
2. Cagnazzo C. et al., The role and challenges of clinical research coordinators: insights from a national survey, BMC Medical Research Methodology 25:238, 23 October 2025.
3. Backman Lönn B. et al., Clarifying the role of clinical research nurses working in Sweden, Nursing Open 9(5), 2 June 2022.
4. Pelazza C. et al., The clinical trial activation process: a case study of an Italian public hospital, Trials, 2024.
5. Dunleavy et al., Why are organisational approvals needed for low-risk staff studies in the UK?, BMC Health Services Research 24, 1408, 15 November 2024.
6. European Medicines Agency, EU tracks progress towards 2030 clinical trial targets, 20 May 2026.
7. European Commission, proposal for a European Biotech Act, December 2025.
8. Paul-Ehrlich-Institut on the Medical Research Act; ANSM fast track, 20 November 2025.
9. Regulation (EU) 2025/327 on the European Health Data Space.