Linking Patient Data Without Sharing Identities with HMAC, KMAC and Tokenisation. Keyed hashing lets a GP system, an acute trust and a community provider join their records on the same patient without any of them, or the analysts downstream, ever seeing each other's NHS numbers. That is the quiet workhorse behind most pseudonymised linkage in… Continue reading Return of the HMAC
My map of all England secure data environments, nhs regions, trusts, icbs & councils is now live: https://charlesgibbons.github.io/sde-map/. Now includes new GSST SDE running on Snowflake, data clean rooms will simplify secure shares, HDRS take note! Next step is a European map of all environments that support data sharing and graphs
When I started building knowledge graph systems for healthcare and enterprise data, I quickly realized there's no single reference showing which tools actually work together. You have document extraction libraries, graph databases, visualization tools, and LLM integrations—but which ones belong in your stack? That's why I've created a comprehensive guide to 100+ graph libraries, organized… Continue reading Useful Graph Libraries Guide
Turning unstructured data into a graph database is a deceptively complex problem. You need to extract entities, understand relationships, define schemas, and maintain data quality whilst managing costs and supporting new data sources. The challenge isn't choosing whether to use a graph database; it's choosing how to populate it. This guide walks through seven different… Continue reading A Complete Guide to Data Ingestion Approaches from Documents to a Graph Database
Relational databases handle type hierarchies badly. A common workaround is a "type" column in a table, which flattens all subtypes into a single wide row and forces application code to pick which columns apply. Object-oriented code has classes, inheritance but Java struggles with multiple parents. Graph databases have something more profound. The relationship itself carries… Continue reading Inheritance in Graph Databases: How to Model a Type and Its Subtypes
There is no a single standard airlock. Sensitive data, whether patient records or financial information, requires control governing the sharing of that data. These controls must conform with the common common law concept of confidentiality, data protection laws (primarily UK GDPR and the Data Protection Act 2018) and variance across organisational governance. Specifically the variant… Continue reading In Space No Two Airlocks Are The Same
The DSPT is an online self-assessment tool that allows organisations to measure their performance against the National Data Guardian’s 10 data security standards and must be reported yearly if interacting with patient data. The DTAC is an assessment framework for care commissioners and providers to use when assuring digital health technology (DHT) product. It is… Continue reading NCSC Cyber Assessment Framewok, NHS Data Security and Protection Toolkit & NHS Digital Technology Assessment Criteria
Every major economy now has an AI safety body. The international network that came out of the Seoul Summit in 2024 has grown to include the UK, the US, the EU, France, Germany, Japan, South Korea, Canada, Singapore, India and Australia. On paper it looks like coordinated global governance. In practice, almost none of these… Continue reading International AI Bodies and their Powers
These three concepts get conflated constantly in healthcare informatics conversations. All deal with clinical codes and they all do fundamentally different jobs. Understanding what each does is critical if you're building anything that touches coded clinical data. Category What it does Major examples Healthcare context Terminology Clinical code systems Defines the codes, concepts, descriptions and… Continue reading Understanding AI Classifiers, Terminologies, terminology Engines
Because SNOMED and ICD Codes must be treated as categories for gradient boosting The 8 main blood types (A+, A-, B- rare, B+, O+, O- universal, AB+, AB-) are categories. If I used label encoding then each category becomes integer (e.g A+ =0, B-=1, AB+=2). This is compact but introduces a false ordering and the… Continue reading Using CatBoost.ai in Healthcare