Methods for investigating critical incidents in healthcare
The aim of the project is two fold:-
(i) to develop a formal, comprehensive research based method of analysing adverse events and
to produce strategies to prevent avoidable errors in the future.
(ii) to develop a practical technique for use by risk managers in the UK to investigate
adverse events in both acute and mental health settings in association with ALARM, (the
Association of Litigation and Risk Management), with the long term aim of establishing a
clinical adverse event database to facilitate trend analysis and identify patterns of
underlying causes of adverse events.
The assessment of adverse medical outcomes has generally focused on identifying relevant
patient and treatment characteristics. Most studies have taken the form of large scale
audits of serious incidents. These questionnaire-based surveys have provided valuable
data on the nature and frequency of adverse outcomes but limited information on their causes.
Serious incidents are routinely discussed in mortality and morbidity meetings. However,
discussion is usually limited to patient and treatment characteristics, and the identification
of any clear cut departures from expected standards. There is seldom any attempt to identify
the chain of events leading to an adverse outcome and the wide range of factors that influence
clinical performance. In contrast, sophisticated methods of accident analysis have been
developed in aviation, nuclear and chemical industries based on human factors approaches
to safety. Human factors is a hybrid discipline which focuses on the human component within
complex organisational systems. The assessment of what can go wrong with large scale systems
such as Nuclear Power Plants and Transport systems has received considerable interest given
accidents such as the Kings Cross Underground Fire, Three Mile Island and Chernobyl, where
cause has been directly attributed to "human error". As a consequence of such disasters, these
industries have developed a variety of methodologies to identify human errors within a system,
to quantify the impact of human error on the system and finally to provide measures for error
reduction. We have based our work on Reason's Organisational Accident Causation Model [1,2,3]. This method essentially examines the chain of events that lead to an accident or adverse
outcome, considers the actions of those involved and then crucially looks into the conditions
in which staff were working and the organisational context in which the incident occurred.
The CRU's research has focused on developing systematic methods of investigating outcomes based
on interviews with staff, completion of a human factors questionnaire to assess the non
clinical factors, e.g. supervision and communication issues which affected the staff performance
and, finally, a review of the notes [2,4,5,6]. These methods help to build a detailed
sequence of events during the incident, incorporating clinical and non-clinical factors
operating within the event. After an adverse incident has occurred the staff involved in the
incident are interviewed, using a set of pre-defined questions (this takes approximately 20-30
minutes). They are also asked to complete the questionnaire which enables staff to consider the
organisational and environmental issues which may have affected their performance (e.g. poor
communication, supervision problems, team functioning problems, etc.). The information
generated from staff is then integrated to form a description of the event, which is
supplemented with a review of the medical notes. For further information relating to this
investigation process, see references [2,4,7].
Analysis of an adverse incident often reveals a series of errors combined with a set of
unusual circumstances which together lead to a (potentially) catastrophic outcome. This kind
of approach can reveal deep rooted unsafe features of organisations that are both inefficient
and potentially dangerous. Investigation and analysis using this process allows clinicians to
identify with more certainty why errors occur, which in turn enables the production of error
reduction strategies to prevent future problems of a similar nature.
Back to top.
This is available in PDF format:
We value comments about your experiences of using the Investigation Protocol.
We are planning to include more case examples in different healthcare settings.
Please contact us with your comments.
* NB. Some have difficulty downloading the main protocol document, where only the first page is displayed and the remaining pages are blank. Please contact Maria for an attached PDF file if you experience this problem.
Back to top.
- Vincent CA. (ed) Clinical Risk Management. BMJ Publications. 1995
- Vincent CA, Taylor-Adams S, Stanhope N. A framework for the analysis of risk and safety in medicine. British Medical Journal 1998; 316:1154-7
- Vincent CA, Reason JT. Human factors approaches in medicine. In Rosenthal M, Mulcahy L, (eds.) Medical mishaps: pieces of the puzzle. Open University (in Press).
- Stanhope N, Vincent CA, Taylor-Adams S, O'Connor A, Beard RW. Applying human factors methods to clinical risk management in obstetrics. British Journal of
Obstetrics and Gynaecology 1997; 104:1225-1232.
- Taylor-Adams S, Vincent CA and Stanhope N. A method for analysing adverse events in medicine. Safety Science 1999; 31:143-159.
- Taylor-Adams SE, Vincent C. Clinical accident analysis: understanding the
complex interactions between the task, individual, team and organisation. In Vincent C, De Mol B (eds.) Safety in Medicine. Elsevier Science, 2000, 101-16.
- Vincent CA, Bark P. Accident Investigation: Discovering why thing go wrong.
In Vincent CA. (ed) Clinical Risk Management, BMJ Publications, 1995, 391-410.
Back to top.
Home
| CRU Staff
| Research
| History
| Publications
| Courses
| Useful links
| Psychology
| UCL
_____________________________________________
This page was last updated on 12 February 2002.
Comments.