On October 26, 2026, Florence Healthcare announced new tools intended to automate work connecting clinical trial sites and sponsors, including feasibility responses, document quality checks, startup activities, and monitoring workflows. As a vendor announcement, it should not be treated as independent evidence that automation has solved site burden. It is still a useful indicator of where clinical trial technology is heading: away from systems that simply store records and toward systems that initiate, route, review, and reconcile trial work.
That direction is reasonable. Manual movement of information between the EHR, eSource, EDC, eISF, sponsor portals, safety systems, laboratory platforms, and email consumes substantial coordinator time. A 2026 multi-study assessment published in the AMIA Joint Summits proceedings also found fewer transcription errors in the evaluated direct EHR-to-EDC workflow than in manual abstraction. The authors were equally clear, however, that mappings require rigorous validation and that automation remains a governance challenge, not just a technical one.
For sites, the central question is therefore not, “Can this task be automated?” It is, “After automation, can we still identify the authoritative record, reconstruct what happened, and assign responsibility when something goes wrong?”
The system-of-record decision must come before automation
An automated workflow may move a document from a site-controlled eISF into a sponsor-controlled eTMF, extract data from an EHR into EDC, or populate a feasibility questionnaire from previously stored answers. Each function can remove repetitive work. It can also create uncertainty about which version is authoritative.
Consider a delegation log updated in the site’s system and transmitted to a sponsor platform. If a coordinator later corrects a role or effective date, does the correction overwrite the sponsor copy, create a new version, or wait for approval? Which timestamp governs? Can the site see whether transmission succeeded? Does the sponsor’s copy preserve the site system’s metadata? These questions are operational controls, not minor IT details.
Before implementation, the sponsor, CRO, site, and vendor should map record ownership at the level of the individual object. “The eISF is the system of record” is often too broad. Training evidence, safety correspondence, consent documentation, source data, monitoring follow-up, and financial records may each have different authoritative locations and retention requirements.
- Identify where the original record is created.
- Identify which system holds the authoritative version.
- Define whether downstream copies are reference copies or certified copies.
- Specify which metadata and audit-trail elements move with the record.
- Document who may correct, approve, reject, or supersede the record.
Automate routine work, but design the exception path first
Most demonstrations show the expected path: a correct document is uploaded, classified, checked, transferred, and filed. Site operations are dominated by exceptions. A license expires during startup. A principal investigator signs with an incorrect date. Two staff members upload the same certificate. An amendment changes required training while documents are in transit. A participant’s name differs between connected systems. A user loses access during staff turnover.
The workflow needs to distinguish between an administrative discrepancy and an issue that could affect participant protection, eligibility, treatment, safety reporting, or data reliability. It should then route the exception to someone with the authority and context to resolve it.
The phrase “human in the loop” is not enough. The operating plan should name the reviewer by role, define the queue being reviewed, establish response expectations, and describe escalation. If everyone assumes someone else is monitoring the exception dashboard, the automation has not removed work. It has made the work less visible.
- Test rejected, corrected, and superseded records, not only successful transfers.
- Define what happens during vendor or internet downtime.
- Confirm how urgent safety and eligibility issues bypass routine queues.
- Create alerts for unresolved exceptions that are approaching protocol deadlines.
- Maintain a documented manual fallback that staff can actually execute.
Auditability has to survive movement between systems
FDA’s guidance on electronic systems emphasizes trustworthy and reliable records, preserved metadata, controlled access, training, and audit trails capable of showing who changed a record, when it changed, and what changed. ICH E6(R3) similarly emphasizes proportionate, risk-based controls focused on factors critical to trial quality.
An integration should not flatten a traceable electronic record into an unexplained PDF or move a value without enough context to establish its origin. The receiving system must preserve or link to the information needed to understand the record. That may include the originator, creation time, modification history, reason for change, units, time zone, and transformation rules.
This is especially important when automated tools classify documents or recommend answers. The site should be able to tell the difference between information copied directly from an approved source, information transformed by a rule, and content proposed by an algorithm and accepted by a staff member. Those actions do not carry the same operational or compliance meaning.
- Retain enough metadata to reconstruct the record’s history.
- Make automated and human actions distinguishable in the audit trail.
- Restrict administrative privileges and periodically review access.
- Confirm that records and audit trails can be exported in a usable format.
- Define record custody and access after study closeout or vendor termination.
Feasibility should measure workflow fit, not just platform availability
Technology feasibility questions frequently stop at whether the site has an EHR, eISF, CTMS, or compatible browser. That does not establish whether an integration is practical. Two sites using the same named platform may have different versions, configurations, security policies, staffing models, and contracting constraints.
A useful assessment examines the full operating environment. Who can approve an interface? How long does local IT review take? Can the site provide a testing environment? Are laboratory units standardized? Does the EHR consistently capture protocol-required data as structured fields? Can participant identifiers be mapped without increasing privacy risk? Who performs reconciliation after go-live?
Community sites may have faster local decisions but fewer dedicated technology staff. Health systems may have deeper IT support but longer security and contracting reviews. Sites in Miami and Puerto Rico may also be managing bilingual documents, multiple referral sources, and systems that were not designed around research workflows. A realistic feasibility process recognizes these differences rather than labeling one model as universally more technology-ready.
- Request a workflow diagram, not a yes-or-no technology answer.
- Separate technical compatibility from organizational approval.
- Estimate site hours for mapping, testing, training, and reconciliation.
- Assess whether the integration covers high-volume data or only a small subset.
- Confirm that the site can continue operating if the connection is unavailable.
Automation has a site budget and staffing impact
Automation can reduce repetitive work after implementation, but implementation itself requires labor. Site personnel may need to review field mappings, test sample records, reconcile discrepancies, complete system training, maintain user access, investigate transmission failures, and support validation evidence.
Those activities should be visible in the study budget and responsibility matrix. Otherwise, the sponsor may record a technology efficiency while the site absorbs the configuration and quality-control work. This is particularly risky for smaller sites, where the same coordinator may be responsible for participant visits, data entry, regulatory maintenance, and technology troubleshooting.
Sponsors and CROs should provide one accountable support route rather than requiring sites to determine whether a problem belongs to the EDC vendor, integration vendor, CRO help desk, sponsor data manager, or local IT team. Issue ownership and escalation times should be established before first participant activity.
- Budget startup testing and configuration separately from routine data entry.
- Include ongoing reconciliation and access administration.
- Define payment for substantial retraining after system changes.
- Provide support coverage that matches visit and safety-reporting schedules.
- Include closeout exports and archival support in the implementation plan.
A practical readiness exercise for sites and study teams
Before activating an automated connection, select five representative records or data elements and follow each from creation through review, transmission, correction, reconciliation, and retention. Include at least one failed transfer and one post-transmission correction.
Ask a staff member who did not configure the workflow to reconstruct what happened using the available records and audit trails. If that person cannot identify the source, current version, responsible reviewer, and final disposition, the workflow is not ready, regardless of how smoothly the demonstration ran.
The most useful automation will not simply move work faster. It will reduce duplicate effort while making ownership, status, and exceptions easier to see. That is the standard sponsors, CROs, sites, and technology vendors should apply as connected and agent-assisted workflows become more common.
- Can the team identify the authoritative record within minutes?
- Can it see whether transmission succeeded or failed?
- Can it explain every automated transformation?
- Can it locate unresolved exceptions and their owners?
- Can it reconstruct the process after staff turnover or system closure?
Sources and further reading
- Florence Automating the Connection Between Study Sites and Sponsors, with AI Tools Already Live
- Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Questions and Answers
- E6(R3) Good Clinical Practice Guidance for Industry
- Advancing Clinical Trial Efficiency and Data Accuracy through Direct EHR-to-EDC Integration
- Clinical Trial Site Technology Challenges and Solutions Guidance