See every change, not just the headline update.

What it does and who it's for

The Guideline Updates module automatically monitors and highlights changes within clinical guidelines from sources like NICE, SIGN, and Royal Colleges. It provides granular, recommendation-level tracking to ensure clinical teams are always aware of the latest evidence-based advice. This tool is designed for clinical governance leads, patient safety teams, and departmental heads responsible for ensuring their teams practice in line with current national standards.

In practice, the module addresses the critical challenge of 'silent drift', where clinical practice gradually diverges from updated guidelines because changes go unnoticed or are not effectively communicated. By providing a systematic and auditable process for tracking and reviewing updates, it supports organisations in meeting their clinical governance obligations, particularly those outlined in frameworks like the NHS Patient Safety Strategy and the Duty of Candour. The system is equally valuable for individual clinical specialties managing complex, rapidly evolving evidence bases, such as oncology, cardiology, or infectious diseases.

How it works

The system operates through a structured process of input, processing, and output to deliver actionable intelligence on guideline changes.

Inputs

CliniSearch continuously ingests guidelines from authoritative UK sources. It also allows organisations to upload and track their own local policies and protocols alongside national guidance, ensuring a unified source of truth.

The system's ingestion process is automated and monitors publisher feeds, websites, and official repositories for new publications or revisions. Supported national sources include, but are not limited to, the National Institute for Health and Care Excellence (NICE), Scottish Intercollegiate Guidelines Network (SIGN), Royal Colleges, specialist societies, and relevant government health bodies. For local policies, administrators can upload documents in standard formats (PDF, DOCX), and the system will extract and index the textual content for version tracking.

Processing

Each guideline is indexed at the individual recommendation level. When a new version is published, the system performs a detailed comparison (diff) against the previous version. It automatically versions each recommendation and tags changes based on pre-defined criteria such as clinical impact or wording alteration.

The processing engine uses natural language processing (NLP) techniques to parse guidelines into discrete recommendations, identifying key elements like population, intervention, comparator, and outcome where possible. The diff algorithm compares text at a semantic level, not just character-by-character, to intelligently identify substantive changes. Changes are automatically categorised into types such as 'New Recommendation', 'Recommendation Removed', 'Wording Change', 'Strength of Evidence Change', or 'Dosage/Threshold Change'. This categorisation helps users quickly assess the nature and potential impact of an update.

Outputs

Users receive targeted alerts for changes relevant to their specialty or interests. The system provides clear, side-by-side comparisons of old and new recommendations and generates audit-ready artefacts that document which changes were reviewed, by whom, and when.

Alerts can be configured to deliver via email, in-platform notifications, or integrated messaging systems (e.g., Microsoft Teams). The comparison view highlights added text in green and removed text in red, providing an intuitive visual diff. For governance purposes, the system generates standardised reports that include a summary of changes detected, a list of users notified, timestamps of reviews, and any comments added by reviewers. These artefacts can be exported for inclusion in clinical governance meeting packs or for external audit purposes.

Example workflow

The following steps illustrate a typical workflow for managing a guideline update, contrasting the process before and after using CliniSearch.

Before CliniSearch

1. A new guideline version is published on a publisher's website.
2. The change may go unnoticed unless manually checked by a team member.
3. If noticed, a clinician must read the entire document to identify specific recommendation changes.
4. Changes are manually communicated via email or meeting, with no formal tracking of acknowledgement or review.
5. No definitive audit trail exists to prove the team is practicing to the latest standard.

After CliniSearch

1. The system automatically detects the new guideline version upon publication.
2. An alert is sent to relevant clinical leads, highlighting the update and its tagged severity.
3. The lead accesses the platform to view a precise diff of only the changed recommendations.
4. After review, the lead logs an acknowledgement or note directly within the system.
5. The audit trail is automatically updated, and a summary can be generated for the next governance meeting.

This streamlined workflow significantly reduces the time from guideline publication to team awareness and documented review. For instance, a critical update from NICE on sepsis management can be flagged to the emergency department lead and the patient safety team within hours of release. The lead can then quickly review the specific changes to antibiotic recommendations and fluid resuscitation protocols, add a note about disseminating the changes to the rota, and have a complete record for the Trust's Clinical Effectiveness Group. This process replaces an ad-hoc, error-prone system with a reliable, time-stamped governance loop.

Governance & safety

Robust governance features are central to the module's design, ensuring clinical safety and accountability.

Provenance

Every piece of guidance is tagged with its source and publication date, providing clear provenance. Links to the original source document are always maintained for verification.

The system records the exact URL and timestamp of source retrieval, creating an immutable record of the guideline's origin. For local policies, the uploader, upload date, and document checksum are recorded to ensure integrity. This level of provenance is essential for meeting standards like the DSPT (Data Security and Protection Toolkit), where demonstrating the use of authoritative sources is a key requirement.

Update Handling

The system differentiates between major and minor changes, allowing teams to prioritise reviews. The previous version of each recommendation remains accessible for a configurable period to support retrospective audit.

Change severity is determined by a combination of automated tagging and configurable rules set by the organisation. A major change, such as a new first-line treatment recommendation, might require acknowledgement within 7 days, while a minor typographical correction might simply be logged. Organisations can define their own review timelines and escalation paths for different change types. Archive policies ensure that previous versions are retained for a minimum period (e.g., 7 years) to support clinical negligence claims or historical audits.

Wording Controls

To prevent misinterpretation, the system highlights nuanced wording changes (e.g., "should consider" vs. "should offer") that might significantly alter clinical meaning.

The NLP engine is trained to recognise the significance of modal verbs and qualifying phrases common in clinical guidance. A change from "may be considered" to "is recommended" represents a substantial shift in the strength of a recommendation, and the system will flag this as a high-priority wording change. This helps prevent clinical teams from overlooking subtle but critical updates that could affect patient management decisions and consent processes.

Audit Trail

A comprehensive, timestamped log records every interaction with an update, including when it was detected, who was alerted, who reviewed it, and any notes added. This creates a defensible record for clinical governance and safety committees.

The audit trail is non-editable and forms a complete chain of custody for each guideline change. It can be filtered by date, user, guideline, or change type, making it easy to generate reports for specific incidents, time periods, or audits. For example, if a question arises about when a new anticoagulation protocol was implemented, the audit trail can provide unequivocal evidence of the review date and the responsible clinician, supporting accountability and demonstrating proactive clinical governance.

Frequently Asked Questions

Implementation

How quickly can we implement this module? Implementation can typically be completed within a few weeks, depending on the integration requirements with existing clinical systems. Our team provides full support throughout the process.

Does it require training for staff? The interface is designed for intuitive use, but we recommend a short, focused training session for clinical governance leads to ensure they maximise the benefit of all features.

Can we pilot the module in one department first? Yes, a phased rollout is common. We can enable the Guideline Updates module for a single directorate or specialty (e.g., Cardiology) as a pilot before rolling it out across the entire organisation. This allows for workflow refinement and demonstrates value to stakeholders.

Data Handling

Where is our data stored? All data is stored securely within UK-based data centres that comply with NHS Digital's Data Security and Protection Toolkit (DSPT).

How do you handle local policies? Locally uploaded policies are treated with the same level of security and version control as national guidelines. Access is restricted to authorised users within your organisation.

Is user activity data anonymised? User activity related to guideline reviews is logged with user identifiers to maintain the audit trail for governance purposes. This data is processed under the legal basis of 'public task' for NHS organisations and is covered by our Data Processing Agreement. All access is controlled by role-based permissions.

Limitations

Does it cover all clinical specialties? The system covers all major national guideline publishers. Coverage of highly specialised or very localised guidelines may vary and can be supplemented with manual uploads.

Can it integrate with our Clinical Decision Support (CDS) system? Yes, we offer API-based integrations to feed update alerts and new recommendations directly into other clinical systems, subject to technical compatibility.

How does it handle rapid or emergency guideline updates? The system is designed to process updates as soon as they are published. For time-critical updates, such as those from UKHSA during a public health emergency, the alerting system can be configured to send immediate, high-priority notifications to specific user groups, bypassing standard review queues to ensure rapid dissemination.

What about guidelines that are not in a standard digital format? While the system works best with digitally native documents, our team can assist with processing scanned PDFs or other non-standard formats. However, the accuracy of recommendation extraction and diffing may be reduced for non-digital sources.

Core capabilities