Collaborative Glossary

This is an area of the Wiki for members of the Discovery Collaborative to share definitions of common terms. It is public and can be referenced from other documents and websites as necessary

Term	Definition
Cloud	The Cloud refers to IT infrastructure, platforms, and services that are hosted in a remote data centre, managed by a third party. Typically these data centres are run and managed by large companies such as Microsoft (Azure), Amazon (AWS), and Google (GWP), but there are many other providers too. Using the cloud removes the need for customers to buy and maintain their own physical IT equipment. Instead, the cloud provider typically charges customers daily rates based on storage used, computational power used, memory, number of servers etc
Cloud Native	Cloud Native refers to building software and systems in the cloud that make use of services made available by cloud providers, to run scalable, resilient applications. For example, instead of creating a SQL database on a server in the cloud, which the customer would need to monitor, maintain, patch, back-up etc, they could make use of a “SQL database as a service” from the cloud provider, who would automatically take care of all the maintenance without any further intervention from the customer
Common Data Model	This is the Data Model which has been adopted as a standard for an organisation or group. It defines how the elements of data relate to one another and enables data to be transferred between different systems which share the Common Data Model with ease.
Conceptual Data Model	This is a high level model, describing business processes and how they relate to one another. In simple terms, it is a map of concepts and rules relating to the business which can then be used to define relationships between data entities in order to describe the business concepts – this leads to the creation of a logical data model
Data Model	Data models exist at different levels of abstraction, from conceptual data models which are high-level and consider business activity and relationships, to a physical data model which represents the implementation of a database. A data model shows how elements of data relate to one another and how they are structured, how they relate to business needs, and how they can be manipulated.
Data Platform	A Data Platform is any digital solution which enables an organisation to store, transform, aggregate, and analyse data. It integrates many different technologies for different purposes to meet the data needs of the organisation. It includes security and access controls to ensure IG compliance
Database Index	Indexes are used to quickly sort and retrieve data in a database. They may take the form of a physical, clustered index, which is the physical ordering of the data in the table (e.g. by ID or surname), or a non-clustered index which describes the logical ordering of the data separately – like the index at the back of a book: the index key is ordered (e.g. surname) and next to it is information showing where to find the record.
Database Schema	This is the definition of the structure of a relational database and defines how data is organised, including the table names, fields (also known as columns), data types, constraints, indexes, and relationships between these entities. The schema is a blueprint of how the database is constructed.
DDS Information Model Manager	The Information Model Manager is an application within DDS which provides a view of the Common Information Model and to support authoring of the content of the model. It can be used to view the model structure and content, download artefacts (e.g. value sets), and manage the ontology. It can be used to view entities and their relationships (e.g. an A&E admission, which is an Encounter and has a subject of a Patient and has a sub-encounter of a Triage Encounter etc)
Endeavour Information Model
FHIR	FHIR stands for Fast Healthcare Interoperability Resources. It is an international standard for health care data exchange and is published by HL7. It is designed to enable the exchange of healthcare related information, including clinical, administrative, publich-health, and research data. DDS stores data in the core in FHIR format before it is restructured and sent to subscriber databases
Graph Database	Graph Databases are used to represent data as elements and their relationships (known as nodes and edges). They are particularly beneficial when representing large datasets with complex and numerous relationships, such as social media contacts and relations. Unlike a relational database, where to return data it is often necessary to join many tables, in a graph database this isn’t necessary. Relationships are stored natively alongside the data elements enabling much faster querying
Logical Data Model	The logical data model describes how data objects relate to each other, but is independent of the technology or database management system used. Logical data models typically show entities (e.g. a patient, an appointment, a result), relationships (e.g. an appointment is linked to a patient via a unique identifier), and attributes (information that is useful to further describe the entities)
NHS Terminology Server	The NHS Terminology Server “is a FHIR compliant solution that holds and disseminates assured international terminologies and classifications (such as SNOMED-CT and ICD-10) and national terminologies (such as NHS Data Model and Dictionary codes).” See https://digital.nhs.uk/services/terminology-servers. Content is delivered in machine readable format and can be accessed as a real-time resource through APIs to support other applications.
Normalisation (health data)	In healthcare, this is the process of taking data from different sources, in different formats and with different code sets, and converting the data into a singular, unified clinical language or terminology – for example mapping proprietary codes from EPR systems to SNOMED codes
Normalisation (database)
NoSQL Database	NoSQL stands for Not Only SQL, and is a generic name given to databases which store data in a non-tabular format. They are not relational databases – they may use storage techniques such as key-value pairs, document stores, graph databases. In some cases, SQL can still be used to query them
On-Premise (On-Prem)	IT infrastructure build and maintained in-house – for example, database servers sitting in the basement of a hospital building which the IT department maintain and upgrade themselves
Ontology	An Ontology is a set of concepts and categories showing the relationships and properties between them in a particular domain. In healthcare, and ontology is used for modelling the semantics of medical concepts and to enable the exchange of medical data between systems. NEED SOME SORT OF REAL-LIFE EXAMPLE
Physical Data Model	The physical data model describes how a database should be structured and is a representation of table structures, columns, column names, column constraints, primary keys, foreign keys, and any other physical features of the database. A database is an implementation of a physical data model.
Relational Database	A Relational database is one where the data is stored in tables, featuring rows and columns and has predefined relationships between the data items. Typically, a table will have a primary key, which is a unique identifier for items in that table. Other tables will reference that primary key for items that are related to it by means of a foreign key. For example, in a table of patients, the primary key might be the NHS number
SQL	A Relational database is one where the data is stored in tables, featuring rows and columns and has predefined relationships between the data items. Typically, a table will have a primary key, which is a unique identifier for items in that table. Other tables will reference that primary key for items that are related to it by means of a foreign key. For example, in a table of patients, the primary key might be the NHS number
SQL Database	A SQL Database is a relational database, which is a collection of tables storing a specific set of structured data, with a fixed schema, which can be queried using SQL