Find thresholds instantly, with sources.

What Clinical Thresholds Does

Clinical Thresholds is a feature within the CliniSearch platform that automatically identifies, extracts, and presents critical numerical values from UK national clinical guidelines. It provides clinicians with immediate access to validated thresholds—such as diagnostic criteria, treatment targets, and referral triggers—alongside their original source citations and decision context. This tool is designed for all healthcare professionals, including GPs, hospital doctors, nurses, and allied health professionals, who need to make rapid, evidence-based decisions involving numerical criteria during patient care.

The feature is particularly valuable in high-pressure environments like primary care clinics, A&E departments, and inpatient settings, where quick access to precise, authoritative numbers can directly impact patient safety and pathway adherence. By centralising thresholds from multiple, often-updated sources, it eliminates the need to consult disparate, potentially outdated documents, thereby standardising practice and reducing cognitive load for busy clinicians.

How It Works

The system operates through a structured process to ensure accuracy and reliability. It begins by ingesting clinical guidelines from authoritative UK sources like NICE, SIGN, and Royal Colleges, as well as incorporating validated local trust or CCG policies where applicable. This content is then processed using natural language processing and clinical coding to index every numerical threshold, categorising them by clinical concept (e.g., HbA1c, blood pressure) and the action they signify (e.g., 'refer', 'treat', 'monitor'). Each threshold is version-controlled, meaning any update to the source guideline is automatically detected, and the corresponding threshold in the system is flagged as changed.

The output for the user is a clear, searchable interface. When a clinician queries a term, the system returns relevant thresholds with the exact numerical value, the source guideline (including section and publication date), and a tag indicating the clinical decision (e.g., "Consider specialist referral"). For enterprise users, the system can also generate audit artefacts, showing which thresholds were accessed and when, supporting clinical governance and training needs.

Underlying this process is a robust technical architecture. The NLP engine is specifically trained on medical text to accurately distinguish between different types of numerical statements (e.g., a target value versus an exclusion criterion). The indexing system uses a standardised clinical ontology to ensure consistency, and the versioning system maintains a complete history of each threshold, allowing users to see not just the current value but also when and why it was updated.

Example Workflow

The following steps illustrate a typical user journey for checking a clinical threshold, contrasting the traditional method with the CliniSearch process.

Before CliniSearch (Traditional Workflow)

1. Clinician recognises a need to check a threshold (e.g., INR target for a patient with atrial fibrillation).
2. Searches through local trust intranet or paper-based protocols, which may be outdated.
3. If local policy is unclear, searches NICE website directly, navigating multiple pages.
4. Manually locates the relevant numerical value and its context within a lengthy PDF.
5. Process can take 5-10 minutes, with a risk of using an incorrect or obsolete value.

After CliniSearch (Streamlined Workflow)

1. Clinician enters a search term like "INR target AF" into the CliniSearch platform.
2. System instantly returns the current threshold (e.g., "2.5") with the source (NICE CG196) and decision tag ("Treatment target").
3. User can see if this value has recently changed and view the rationale from the guideline excerpt.
4. Total time taken is under 15 seconds, with confidence in the provenance of the information.

This efficiency gain is compounded over multiple patient encounters, saving significant time per clinician per day and reducing the likelihood of errors stemming from inaccessible or outdated reference materials. The system's ability to surface the precise source and context also aids in documentation and justification of clinical decisions.

Governance & Safety

Clinical safety and information governance are paramount. The provenance of every threshold is meticulously recorded, linking it directly to the source guideline paragraph. A rigorous update handling process ensures that when a source guideline is revised, the change is detected, and the new threshold is highlighted within the system, with the previous version retained in an audit trail. This prevents the silent propagation of outdated information.

Wording controls are applied to ensure that the presentation of thresholds is clear and unambiguous, avoiding any phrasing that could be misinterpreted as a direct clinical instruction. The system is designed to support, not replace, clinical judgement. A comprehensive audit trail logs all user interactions with thresholds, providing transparency for clinical governance, safety audits, and compliance with standards like DCB0129 and DCB0160 where applicable.

The governance model includes regular reviews by a clinical advisory panel to validate the accuracy of extracted thresholds and the appropriateness of their presentation. All data processing adheres to the NHS Data Security and Protection Toolkit, with hosting in UK-based, secure data centres. For enterprise deployments, the system supports integration with clinical safety management systems to facilitate hazard logging and risk mitigation as part of the organisation's overall clinical risk management framework.

Frequently Asked Questions (FAQ)

Implementation

How is Clinical Thresholds implemented within an NHS trust?
Implementation typically involves integrating the CliniSearch platform with the trust's existing clinical systems, such as the EPR or intranet, via API or a single sign-on (SSO) portal. The CliniSearch team works with clinical safety and IT leads to ensure a smooth rollout with minimal disruption. A phased approach is often used, starting with a pilot department to validate usability and workflow integration before trust-wide deployment.

Can the system be customised for local formularies or pathways?
Yes. The platform can be configured to prioritise or highlight thresholds from specific local guidelines or formularies where these have been formally approved and provided to CliniSearch. This ensures that national guidance is presented in the context of local implementation policies.

Data Handling

What patient data does the system handle?
Clinical Thresholds is a reference tool and does not process or store any patient-identifiable data. Searches are based on clinical terms (e.g., "eGFR threshold CKD"), not patient-specific information. All platform data is hosted securely in the UK in compliance with NHS Data Security and Protection Toolkit requirements.

How are searches and usage data handled?
Aggregated, anonymised data on search terms and feature usage may be collected to help improve the service and understand clinician information needs. This data is never linked to individual users or patients and is used solely for analytical purposes to enhance the product.

Limitations

What are the limitations of the Clinical Thresholds feature?
The feature is designed for rapid reference and decision support. It does not calculate patient-specific scores or integrate directly with patient records to auto-populate values. It is essential that clinicians use the thresholds within the context of the full clinical picture and adhere to local policy where it exists. The system is not a medical device.

Does it cover all specialties and conditions?
Coverage is extensive and continuously expanding, focusing on areas with well-defined national guidance. However, it may not include every possible threshold for every rare condition or highly specialised area. The system clearly indicates the source and scope of the guidance presented, and clinicians should exercise judgement for cases falling outside standard guidelines.

Core capabilities