NICE Updated Its Real-World Evidence Framework with Unstructured Data Guidance in 2025

The National Institute for Health and Care Excellence (NICE) published its real-world evidence framework on 23 June 2022 to improve the quality of real-world evidence (RWE) used in its guidance. NICE describes it as a living framework that will be updated periodically to reflect user feedback, lessons from implementation, exemplar case studies and developments in RWE methods. The framework is published as corporate document ECD9 and is used across NICE’s guidance programmes.

The framework identifies when RWE can reduce uncertainty in NICE decisions, for example in characterising a health condition, its care pathway and outcomes, or in estimating the effects of interventions. It gives particular attention to comparative treatment effects, which require greater scrutiny before decision-makers can rely on them. For such studies, the framework’s methods guidance follows the target trial approach, in which the observational study is designed to emulate a hypothetical randomised trial.

NICE states that the framework does not set minimum acceptable standards for the quality of evidence. Instead it describes good practice in planning, conducting and reporting studies so that committees can judge whether the evidence is trustworthy and relevant for the decision.

According to a summary published by Becaris Publishing, the most recent revision, published in September 2025 and described as the fourth update, adds a section on the validity of information extracted from unstructured data. This covers sources such as imaging, laboratory reports and clinical notes. The framework states that validation studies should be undertaken and reported for methods used to derive structured data from unstructured data, an issue of growing importance as natural language processing and other automated extraction tools are applied to health records.

The same revision adds a reporting tool for studies that use a target trial approach and refers to the TARGET Statement published in JAMA. The framework notes that the checklist is intended for reporting observational studies of interventions that emulate a parallel-group, individually randomised target trial.

Earlier revisions had added guidance on planning, conducting and reporting qualitative research, references to the PRINCIPLED framework for target trial studies, and material on external validity and transportability of evidence from one setting to another. The update numbering in secondary sources is not fully consistent, and NICE’s own update history should be consulted to confirm the date and content of each revision.

A central element of the framework is the Data Suitability Assessment Tool (DataSAT), included as an appendix with a case study. DataSAT provides a reporting template that helps evidence developers set out the key elements of data suitability so that NICE committees are adequately informed. It covers the provenance of the data, data quality and the relevance of the data to the decision problem.

Commentators have noted that DataSAT focuses on justifying the choice of dataset to the committee, and that not every element of an RWE study yet has a dedicated checklist. In practice, evidence developers are expected to combine DataSAT with established reporting guidance appropriate to the study design, now including the TARGET Statement for target trial emulations.

Companies preparing evidence for NICE should treat the framework as the reference point for any submission that relies on routinely collected data, registries or other non-randomised sources. For comparative effectiveness questions, this means specifying the target trial explicitly, describing how confounding and other biases were addressed, and reporting in line with the TARGET Statement.

Where key variables were derived from free text or other unstructured sources, submissions should include the validation evidence the 2025 revision calls for, such as the accuracy of the extraction method against a reference standard. Medical writers should ensure that study reports present DataSAT information in full, as committees and external assessment groups use it to judge whether the dataset is fit for purpose.

Because the framework is a living document, teams should check the current version on the NICE website at the start of each project rather than relying on an earlier download.

The framework also has a bearing on how NICE committees read evidence submitted from other countries. Its material on external validity and transportability asks evidence developers to consider whether results obtained in one population or health system can be applied to patients in England, and to explain any adjustments made. Companies relying on overseas registries or claims databases should therefore set out the differences between the data source and NHS practice, such as treatment pathways, patient characteristics and outcome recording, and describe how those differences were handled in the analysis.