Tools and resources
Terms used in the ESF
Terms used in the ESF
This section defines terms that have been used in a particular way for the ESF and the user guide.
Artificial intelligence
Artificial intelligence (AI) covers a range of computational methods for performing tasks that would ordinarily need human-level intelligence.
In healthcare, AI can be used to analyse large amounts of data to find patterns that are linked to an outcome, such as analysing data from MRI scans to find patterns that are linked to the presence of a tumour. AI can also be used to extract value from text information, such as patient notes, to identify patterns associated with health outcomes like risk of disease progression. Similarly, it can be used to analyse data on healthcare service use, to help to understand how to most efficiently deploy healthcare staff.
The precise definition and scope of the term 'AI' can vary between different contexts, and the level to which different digital health technologies (DHTs) use or rely on AI can vary greatly. In light of the variation in the way that the term 'AI' is used in healthcare, the NICE ESF instead refers to 'data-driven' DHTs, because this is a term that is easier to define in clear terms.
Company
Any commercial entity that is selling or planning to sell a DHT to a healthcare provider. The company may be the same as the developer who created the DHT, or it may be another organisation who is trying to promote the use of the technology in the health and care system.
Data driven
A data-driven DHT is a DHT that meets any of the following descriptions:
-
It contains algorithms that were trained using patient data or datasets. These algorithms could be adaptive, meaning they change over time, or are fixed.
-
It uses decision thresholds or cut-off values (such as for diagnosing a condition or triaging patients for different treatments) that were created using patient data or datasets.
Digital health technology
DHTs include standalone software and apps that are used to improve health outcomes or to improve how the health and care system runs. These can include:
-
regulated medical devices classed as software as a medical device (SaMD) or AI as a medical device (AIaMD)
-
software and apps designed to help people to manage their own health and wellbeing
-
software that is designed to help the health and care system to run more efficiently or to help staff manage their time, staffing or resources
-
apps or software designed to work alongside a medical device.
Software that is embedded in a physical medical device is excluded from this definition.
End user
Any person who is operating the DHT. For software as a medical device or imaging software, this is likely to be the healthcare professional. For health and wellbeing apps, this is likely to be the service user.
Evaluator
Any person or group of people who judges the quality or value of a DHT based on information and evidence provided. These could include NHS commissioners, buyers of DHTs and local evaluators.
Intended purpose
The intended purpose is the objective intent of the manufacturer regarding the use of a DHT. It should state the indication and target population, including when, how and by whom the DHT should be used. The intended purpose of the DHT should be reflected in the information provided by the manufacturer but also needs to take into account how the technology is likely to be used generally. Use outside of an intended purpose may impact the performance and safety of the device. For technologies which fall under the medical device regulations, the intended purpose should allow consistent determination of the regulatory medical device classification and facilitate the development of an adequate risk management, clinical evaluation, quality management and post-market surveillance system.
This page was last updated: