Intended for healthcare professionals

Education And Debate

Changing clinical behaviour by making guidelines specific

BMJ 2004; 328 doi: (Published 05 February 2004) Cite this as: BMJ 2004;328:343
  1. Susan Michie, reader in clinical health psychology (s.michie{at},
  2. Marie Johnston, professor of psychology2
  1. 1Centre for Outcomes Research and Effectiveness, Department of Psychology, University College London, London WC1E 7HB
  2. 2School of Psychology, King's College, University of Aberdeen, Aberdeen, AB24 2UB
  1. Correspondence to: S Michie
  • Accepted 8 December 2003

The problems of getting people to act on evidence based guidelines are widely recognised.1 An overview of 41 systematic reviews found that the most promising approach was to use a variety of interventions including audit and feedback, reminders, and educational outreach.1 The effective interventions often involved complicated procedures and were always an addition to the provision of guidelines. None of the studies used the simplest intervention—that is, changing the wording of the guidelines. We examine the importance of precise behavioural recommendations and suggest how some current guidelines could be improved.

Importance of wording

The wording of a behavioural instruction affects the likelihood that it will be followed, by influencing comprehension, recall, planning, and behaviour. For example, work by Philip Ley shows that using specific concrete statements increases the extent to which information is both understood and remembered.2 3 Individuals who intend to change behaviour are more successful if they have a specific behavioural plan.4 A recognised technique of behaviour modification is to define the target behaviour in specific and concrete terms.5

In a study of attributes of 10 national clinical guidelines, general practitioners followed the guidelines on 67% of occasions if they were concrete and precise but on only 36% of occasions when they were vague and non-specific.6 Implementation of guidelines might therefore be improved by defining the crucial elements precisely.6 7 Clear specification also makes it possible to assess whether the guideline has been followed.1

Specifying behaviour

Recommendations in guidelines are rarely specified in precise behavioural terms such as what, who, when, where, and how. An example is the first clinical guideline to be developed and published by the National Institute for Clinical Excellence, which has been widely circulated throughout the NHS.8 Despite the high quality review of evidence underpinning the guideline, the recommendations are not behaviourally specific. The guideline is long, with the recommendations in the short form exceeding 20 pages. This may result in recommendations getting lost.9 Furthermore, the style of presentation of the guidelines does not optimise implementation.

1 gives the main (grade A, based on randomised controlled trials) recommendations with suggestions on how they could be behaviourally specified. We have limited the suggestions to what and who, but where, when, and how could be similarly added. We have used active verbs rather than general exhortations (should) or descriptions (may). Translating the recommendations in this way revealed two areas of uncertainty. In the first example, it is unclear whether day hospital treatment should be encouraged or merely offered; in the fifth, it is unclear whether depot preparations should be prescribed for everyone with schizophrenia.

BSuggested amendments to published guidelines on schizophrenia to improve specification of behaviour

View this table:

Embedded Image

Constructing the who component of the recommendation showed that the specification of the health professional is slightly different in each of the five recommendations. In practice, the same person may carry out each of the defined functions, but it could be two or more people. The person specification clarifies the initial steps of implementing the guideline for trust boards.

This analysis suggests that these guidelines may have been developed to offer general guidance rather than prescriptive action. However, similar patterns are present in the subsequent three evidence based guidelines published by the National Institute for Clinical Excellence. 2 gives one grade A recommendation from each guideline and illustrative questions that remain unanswered by the guidelines' wording.

Table 2

Illustrative grade A recommendations from three National Institute for Clinical Excellence guidelines and questions about behavioural specifications

View this table:

Analysing and changing behaviour

Specifying behaviour precisely serves two functions in implementing guidelines. Firstly, as described above, it makes implementation more likely; there is greater clarity about what is required and greater certainty about whether it has been accomplished. Secondly, it allows a systematic investigation of the way that the behaviour is linked to what occurs before (antecedents) and what occurs after (consequences). Antecedents and consequences can operate as either facilitators of, or barriers to, behaviour. Changing these is a powerful way of changing behaviour. The scientific analysis of the relations between antecedents, behaviour, and consequences (ABC) is known as behavioural analysis or functional analysis13 and was developed within clinical psychology.14

Examples of manipulating antecedents and consequences and observing the effect on a defined behaviour come from two experimental studies of referral rates of general practitioners. The first introduced an antecedent cue to increase referral rates for mammography.15 The cue was a sticker in a schematic breast shape with a bright orange dot placed on the charts of women aged over 50 years. Referral rates were higher in this condition than in the control condition.

The second study evaluated an intervention to reduce the frequency that general practitioners ordered radiography of the lumbar spine and knee.16 Each time the general practitioner ordered an x ray examination, the results were accompanied by an educational sticker that reiterated the guidelines. This reduced the frequency of ordering by 20%, in line with the recommendations of the Royal College of Radiologists.


Ample evidence exists to support the argument that the simplest, most cost effective intervention to increase the implementation of guidelines is rewriting guidelines in behaviourally specific terms. Future research could compare the effectiveness of rewritten guidelines with that of the original before assessing more complex interventions. Interventions should be developed on the basis of an analysis of the antecedents and consequences controlling implementation behaviours. This analysis should be informed by relevant psychological theory.

Summary points

Psychological research shows that the more precisely behaviours are specified, the more they are likely to be carried out

Rewriting guidelines to increase behavioural specificity may be the simplest, most effective method of increasing implementation

Specifying what, who, when, where, and how will assist implementation

Behavioural analysis of the controlling antecedents and consequences of implementation may help develop effective interventions


  • Contributors SM and MJ are chartered health and clinical psychologists. Their research has covered many areas of health, illness, and the healthcare system, including the application of psychological theory to the implementation of evidence based practice. This article was informed by discussions in a seminar series funded by the British Psychological Society (the scientific understanding of the psychological processes involved in the implementation of evidence based practice in health services) and in an advisory group to a programme grant funded by the Medical Research Council (process modelling in implementation research: selecting a theoretical basis for interventions to change practice; MJ, Martin Eccles, Jeremy Grimshaw, Nigel Pitts, Nick Steen).

  • Competing interests None declared


View Abstract