catherine-henderson profile photo

Catherine Henderson

A word cloud made of words in different sizes and colours, such as care, patients, labour, blood, emergency, birth, covid, surgical, nhs, children, arterial, insulin, detection, medicine, hospital, risks, decontamination, thromboembolism, learning, secure, symptoms and disability.

Development of a systems-based classification scheme to support aggregation of safety intelligence data from healthcare safety investigations.

The Healthcare Safety Investigation Branch (HSIB) published 73 national investigations prior to December 2022, making over 200 safety recommendations to high levels within the healthcare system, such as regulators or governing bodies. These investigation reports contain a wealth of data identifying system level contributory factors to safety incidents within the NHS. Analysis of multiple investigation reports can be used to identify trends, improve learning and inform the development and implementation of recommendations.

The word cloud above was created using the titles of the 20 HSIB national investigations analysed in this project.

Project phases

The picture illustrates the two phases of the project

In phase one the proposed classification scheme was developed. The researchers identified contributory factors in 20 published HSIB national investigation reports, classifying them using existing codes from the internal HSIB SEIPS-based maternity classification scheme, or adapting and creating new codes where needed.

In phase two the proposed classification scheme was assessed for reliability using the index of concordance, then the face validity was considered by both researchers.

A diagrammatic display of the project phases. In the background a large arrow pointing to the right. On top of the arrow on the left hand side two boxes, one says '20 HSIB national investigation reports' and the second says 'HSIB maternity classification scheme' leading to a third box saying 'Proposed national classification scheme'. These three boxes are bracketed below and labeled as phase one. Next on the arrow is a box saying 'Assess reliability' then a box saying 'consider validity', again with a bracket below labeled as phase two.
A table showing the six head codes in the left hand column and then three columns of subcodes. For example: The head code 'external environment' has the subcodes policy regulation, policy recommendation, policy government, covid, geographical, evidence factor, societal, economic, manufacturing and design. The head code 'internal environment' has 7 subcodes such as physical layout factor, environmental factor and time of day. The head code 'organisation' has 19 subcodes such as teamworking factors, documentation factors, change management and procurement. The head code 'job/task' has 7 subcodes such as care planning, medication and risk assessment. The head code 'technologies and tools' has 7 subcodes including usability, accessibility and functionality. The last head code 'person' has 18 subcodes including those about patients, such as physical or psychological characteristics and those about staff, again including physical and psychological characteristics as well as distractions and interruptions or decision making.

Proposed classification scheme

The image shows the final proposed classification scheme consisting of six SEIPS-based head codes and 67 subcodes.

525 contributory factors were identified within 20 HSIB national investigation reports, with only 43 disagreements in coding between the researchers. The index of concordance was found to be acceptable at 91.8% and face validity was considered high by both researchers.

The analysis of coding demonstrated that HSIB carry out system-wide investigations identifying contributory factors at all levels. This is vital to ensure the responsibility for recommendations made from investigations is allocated to those at the appropriate level, with the power and resource to implement change.

The proposed classification scheme enables aggregation of contributory factors from multiple investigations to identify trends and patterns that can be interrogated further or used as supporting evidence for recommendations.

The scheme is currently limited in reliable application to the healthcare areas and subject matter described in the 20 investigations used. Therefore, iterative development should continue as further investigations are completed in other healthcare areas, such as primary care or mental health. Accimaps may be useful for visualisation of the interdependent factors across the healthcare system and a training program should be developed with a guidebook for novice users of the scheme at HSIB.

Other Work

An Accimap diagram showing contributory factors to an aircraft near miss across five levels: 1) Regulatory bodies and external organisations 2) Airline 3) Equipment and environment 4) Pilots 5) Outcome

An Accimap was created in the human factors and systems module to illustrate factors that contributed to an Airbus A321-231 near-miss into Khartoum in 2005. Air Accident Investigation Branch report 5/2007 was used as the source material.

A photo of an RNLI crewman wearing a headset and a yellow jacket in a lifeboat. The photo is taken from behind, looking over the right shoulder and showing the position of the computer screen and the tracker ball for operating it.

Human factors and safety developments in the Royal National Lifeboat Institution (RNLI) since 1990 were considered during the transport module. PPE, such as lifejackets and helmets, were reviewed, as well as looking at the design features of the hulls, such as self-righting capabilities, and the shock-absorbing seating with ergonomic design of equipment, as shown in the picture.

Link analysis diagram of a kitchen showing the position of the dishwasher and kitchen cupboards and drawers. Lines are drawn between these areas to indicate the movement around the kitchen that people have taken when unloading the dishwasher and putting the crockery, cutlery etc. away.

In the data collection and analysis module, link analysis diagrams were created to show the movement around a kitchen by different people (person 1 in red, person 2 in green, grey lines represent movements made by both people) when unloading a dishwasher (in position 1).


First Class Honours BSc Adult Nursing


20+ years aviation experience as a pilot, flight safety manager and crew resource management instructor. 8 years experience in the NHS as a healthcare assistant and registered nurse, currently working as an intelligence analyst for the Healthcare Safety Investigation Branch.

Visionary Thinkers

Visionary Creators

Visionary Makers