Actionable knowledge

What is actionable knowledge?

Actionable knowledge involves processing and packaging knowledge into decision-ready formats that can be acted upon to solve problems.  In the health and social care context this often means decision-ready tools that support decisions about patient care and improving services.

Examples might include mobile apps, clinical guidelines presented as visual pathways, shared decision aids, prompts and reminders in electronic care record systems, risk scoring tools and calculators, evidence summaries with action-focused recommendations, infographics.

‘Actionable knowledge is not only relevant to the world of practice, it is the knowledge that people use to create that world.' (Argyris, 1993). 

ARGYRIS, C., 1993, Knowledge for Action: A Guide to overcoming barriers to organizational change. San Francisco, Calif.: Jossey-Bass Inc.

How do I provide evidence of competency in this area? 

Can you...

  • explain what is meant by actionable knowledge?
  • describe examples of actionable knowledge?
  • understand why actionable knowledge is important to decision support and can be regarded as a type of decision support?

 

Blooms level 2:   Understand

DDAT Framework roles: Relevant to Data analyst, Data scientist, Data engineer, Business analyst, Product manager.

Digital knowledge resources

What are digital knowledge resources?

Digital knowledge resources includes all sources where the information is available in digital formats and accessible with a help of computers. Examples include library resources published in digital format, such as e-journals,  e-books, and PDF reports as well as webs pages, blogs, digital video and audio recordings.

Importantly, digital knowledge resources are human-readable - i.e. they can be easily read and understood by humans. This is in contrast to machine-readable knowledge, which is structured so that computers can process it. Computable knowledge, described in the competency below, is based on machine-readable knowledge. 

Decision support tools exist in both formats.  Digital knowledge resources such as web pages providing guidelines, pathways and formularies are decision support tools.  Machine-readable knowledge such as calculators and scoring tools are also forms of decision support.

How do I provide evidence of competency in this area? 

Can you...

  • Understand and describe the difference between human-readable digital knowledge resources and machine-readable knowledge?
  • Describe examples of digital knowledge resources?
  • Understand the difference between unstructured (narrative), semi-structured and structured forms of human-readable digital knowledge, and illustrate these with examples?
  • Describe how different types of digital knowledge resources can act as decision support tools?

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Computer-executable knowledge

What is computer-executable knowledge?

Computer-executable  knowledge is structured in a format that can be reasoned with or processed by a computer to carry out a task. It is sometimes called computable or computational knowledge.

‘Computer-executable knowledge’ is held in a way that can be used by a computer when it is working on a task. (Wyatt and Scott, 2020)

WYATT, J. and SCOTT, P., 2020. Computable knowledge is the enemy of disease, BMJ Health Care Inform, 27(2)

Translating narrative clinical knowledge such as guidelines, recommendations and evidence summaries into computable knowledge such as executable code or models is a key goal of decision support. Computer-executable knowledge requires:

1) Knowledge to be written in a format that can be read by computers - usually as computer code or a model.

2) A software application to read and process the knowledge and deliver outputs. For example, this could be a drug dosage calculator, a risk scoring tool, or a system that delivers prompts and alerts in electronic care record systems. 

How do I provide evidence of competency in this area? 

Can you...

  • Explain what computer-executable  knowledge is, and why it is important to decision support?
  • Explain how it differs from human-readable digital knowledge?
  • Provide examples of a range of types of computer-executable knowledge and human-readable computerised knowledge?

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Decision support and artificial intelligence (AI)

What is the relationship between decision support and artificial intelligence (AI)?

Decision support uses two main types of artificial intelligence:

1) Knowledge –based expert systems 

These are a long-established form of AI, dating back to the 1950s. They are still widely used in healthcare, in calculators, scoring tools and prompts and reminders and electronic care record systems. Knowledge-based expert systems are:

  • Created by humans based on well-defined knowledge structures.
  • Often based on algorithms derived from validated and transparent evidence. 

2) Machine learning

  • Machine learning is less structured and controlled than knowledge-based AI. In simple terms, machine learning involves feeding data into a computer and seeing what insights the computer provides. Machine learning can provide insights beyond human analysis of data, however it can be less transparent and harder to validate.

The Right Decision Service currently uses knowledge-based expert systems - for example, for calculators, scoring tools and prompts within electronic care record systems.  It is exploring use of machine learning and in time it is likely to make more use of machine learning as well. 

How do I provide evidence that I understand the relationship between decision support and artificial intelligence?

Can you...

  • Describe the different types of artificial intelligence that underpin decision support systems, and outline the pros and cons of each approach?
  • Explain which type of artificial intelligence is used by a particular decision support system - e.g. a Right Decision Service calculator or high risk prescribing decision support embedded in electronic care record systems? 

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Algorithms and decision trees - knowledge-based and computer-generated

What are algorithms and decision trees – knowledge-based and computer-generated?

An algorithm is a process or set of rules to be followed in decision-making, problem-solving or calculations.
A decision tree is a way of modelling the decision-making process. It maps out different courses of action, as well as their potential outcomes.
Algorithms and decision trees are both sometimes visualised as flowcharts.
Algorithms and decision trees can be created by people, based on human knowledge. Using machine leaning, they can also be generated by computers.

How do I provide evidence of competency in this area? 

Can you...

  • Define "algorithm" and "decision tree"?
  • Describe  examples of human-readable algorithms and decision trees, and computer-generated and decision trees?

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Knowledge-based expert systems

What are knowledge-based expert systems?

An expert knowledge-based system is a computer system which emulates the decision-making ability of a human expert. These systems:

  • Use a form of artificial intelligence
  • Are based on structured knowledge created by humans, often  in the form of algorithms and decision trees.
  • Are derived from a body of explicit, validated knowledge.

In healthcare, knowledge-based expert systems may include calculators, risk scoring tools and screening tools, interactive question and answer systems, and patient-specific prompts in electronic health record systems.  The Right Decision Service provides many 

How do I provide evidence of competency in this area? 

Can you...

  • Describe what is meant by a " knowledge-based expert system," and the key components of an expert system?
  • Explain how these systems depend on a validated knowledge base?
  • Provide examples of a range of knowledge-based expert systems in operation in health and care?

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Neural networks and machine learning

What are neural networks and machine learning?

Machine learning is one type of artificial intelligence (AI). It creates computer systems that are able to learn and adapt without following explicit instructions. It achieves this by using computer-generated algorithms and statistical models to analyse and draw inferences from patterns in data. 

A neural network uses a particular type of machine learning process to process data in a way that is inspired by the connections between neurons in the human brain. The term "deep learning" refers to machine learning processes that use multiple layers within a neural network. 

IBM, 2023a-last update, IBM. Available: https://www.ibm.com/topics/machine-learning#:~:text=Machine%20learning%20is%20a%20branch,rich%20history%20with%20machine%20learning. [Dec 11, 2023]

IBM, 2023b-last update, IBM. Available: https://www.ibm.com/topics/neural-networks#:~:text=Neural%20networks%2C%20also%20known%20as,neurons%20signal%20to%20one%20another [Oct 31, 2023]

How do I provide evidence of competency in this area? 

Can you...

  • Describe what is meant by machine learning, neural networks and deep learning?
  • Explain in simple terms to others how machine learning, neural networks and deep learning operate, and the relationships between these concepts?
  • Provide examples of these approaches in operation, in healthcare and beyond?

Blooms level 2:  Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager

Natural language processing

What is natural language processing?

Natural language processing (NLP) is a machine learning technology that gives computers the ability to interpret, manipulate, and comprehend human language. Common examples include:

  • Email filters. 
  • Smart assistants.
  • Search engines such as Google
  • Predictive text.
  • Automated language translation tools

IBM, 2023-last update, IBM. Available: https://www.ibm.com/topics/natural-language-processing#:~:text=Natural%20language%20processing%20(NLP)%20refers,same%20way%20human%20beings%20can [Oct 31, 2023]

How do I provide evidence of competency in this area? 

Can you...

  • Define what is meant by natural language processing?
  • Explain in simple terms to others how natural language processing works?
  • Provide examples of this approach in operation, in healthcare or beyond?

Blooms level 2: Understand

DDAT Framework roles: Data analyst, Data scientist, Data engineer, Business analyst, Product manager