Background
There is growing awareness of the risks associated with health care provision within the NHS and the need for systematic risk management. (1) Many risks and adverse events are associated with medicine use. Until recently, the emphasis on risk management has focussed on secondary care; much less attention has been paid to primary care, although the scale of medicine use is much greater. Greater still is the use of NPMs by the public in community settings. Community pharmacists and their support staff are concerned about the safe supply of NPMs, and the avoidance of harm to their customers. (2) Because of this emphasis on safety, it has been suggested that the failure to achieve the EB supply of NPMs should be considered as an "error" and that HET may provide an appropriate theoretical framework for the development of strategies to change this behaviour. HET (3) has been used by organisations, such as the oil and gas industry, in the process of risk management. This theory categorises errors as slips, lapses, mistakes and violations. Mistakes can be further defined as skill-, rule-, or knowledge-based. If a systems-based approach is used, the system which generated the error is examined and then barriers, defences and safeguards are identified and incorporated to reduce the occurrence of similar errors. This paper applies this model to the supply of NPMs.

Method
All pharmacies in Grampian, Scotland were stratified according to their type and geographical location. One pharmacy was randomly selected from each of the nine strata and invited to participate in a participant observation study, the purpose of which was to determine how evidence is used to inform the supply of NPMs. The 364 consultations recorded during this study were categorised in two ways: 1) by the observer using themes that emerged from earlier interviews. A second researcher is confirming category assignment and disagreements are resolved by discussion. 2) On completion of the categorisation process, an expert group of health and industrial psychologists will be convened. A sample of consultations from each category will be presented to the group members who will be asked to classify the different categories according to HET (i.e. slip, lapse, mistake, violation or no error made). They will then indicate their strength of confidence with their selection of error type and indicate whether it is a skill-, rule- or knowledge-based error. The group will also consider whether the "error" has occurred at an individual or system-level. The process will be repeated until all types of consultation have been classified.

Results
The themes that were used to categorise the consultations were: evidence; communication; safety/risk/ harm; critical incident and, lack of confidence. This work is ongoing and the results of the HET analyses will be ready for presentation by March 2004.

Discussion
The results will form the basis for the development of strategies to address the different "types" of error, which are associated with the supply of NPMs. It is anticipated that different strategies may be appropriate for different types of error and that the type of consultation will inform the content of any intervention strategy.

1. Conference Proceedings: Improving Patient Safety in Scotland. Edinburgh, 2003.
2. Watson MC, Bond CM. The evidence based supply of non-prescription medicines: barriers and beliefs. IJPP (accepted for publication)
3. Reason J. Human Error. Cambridge University Press, 1998.