Return to PJ Online Home Page
The Pharmaceutical Journal Vol 265 No 7114 September 16, 2000
Pharmacy Practice Research
Papers presented at the British Pharmaceutical Conference, Birmingham, September 10 to 13, 2000 pR70

Check mate: an internet-based qualitative study of the processes involved in error checking

By Angela M. Alexander

Introduction A study of dispensed items in hospital pharmacy has indicated that errors are significantly reduced when items are checked by a second person.1 Courses, allowing technicians to perform final checks, have described the mechanistic activities of checking2, but the cognitive processes are less well defined. This study used qualitative methods to collect pharmacists' descriptions of the processes used in checking. The aims of the research were to build up a picture of the processes involved and to investigate the use of e-mail to obtain qualitative data for analysis.

Methods An open question was e-mailed to Private-Rx,3 the internet forum for pharmacists in the UK. The question "checking - how do you do it?" asked what senses were used and what thought processes were involved. The e-mail said that responses would be used as the basis for a study. Members were asked to e-mail their responses privately to the researcher, to avoid the bias of others' thoughts. Responses were analysed using NUD*IST4 to classify themes.

Focal points

  • Qualitative methods were used to collect pharmacists' descriptions of the cognitive processes used in checking
  • The internet provided a low-cost method of gathering primary research material
  • The data obtained support some of the theories for understanding human error and the cognitive processes involved in its avoidance
  • Checking may be seen as an automatic process, but there is evidence that pharmacists employ a variety of different processes to accomplish it
  • A self-audit of the processes employed in checking, as part of risk management in pharmacy practice, is recommended

Results Fifteen responses were received from an anticipated readership of the e-mailed question of 300 pharmacists. Several themes were identified which indicate some of the cognitive processes.
Automatic - "It's like being asked how you drive, once you've learned it's instinct." "My first thought is that I don't know how I actually do it." "Most of it is habit." "It's not something you consciously think about much." "Have never given the process much thought - you just do it."
Dual task interference - "Wait for as long as possible between dispensing and checking; this avoids you getting complacent through memory of the dispensing process." "The mechanics of actually setting aside the whole job is an aid to breaking the image." "Try to leave a gap between dispensing and checking to allow myself a fresh look."
Self monitoring - "Self checking is more difficult; it is harder to re-read objectively." "The essentials of own work checking are bound to be fraught." "It differs whether I am checking myself or others."
Anticipation error - "Pay extra attention to those that are likely to be confused." "Expect to find slips and look for the unexpected."
Confirmation bias - "Confirmation of the matching identities is a large part of it".
Focused attention - "If a different sequence of mechanisms is employed, the likelihood of error being repeated is reduced." "Try to walk away from the current prescription." "Consider a triangle with prescription, label and item at each point." "By working in reverse it seems safer." "If you run your fingertip along the label as you read it, you are more likely to actually read what is in front of you, instead of seeing what you expect to see."
Sensory input - "Read it out aloud." "Visually confirm the product, smell for that characteristic aroma." "Use colour and smell to confirm right drug." "Shake pack to see if it sounds full." "Sometimes the sixth sense pings and something else is spotted." "Use sixth sense to wonder whether there is some gross mistake."

Discussion The internet provided a low-cost method of gathering research material of sufficient quality to provide an insight into the processes used to check prescriptions. This study highlighted some of the advantages of the internet for qualitative research; responses can be obtained regardless of geographic locations, with comments recorded directly to file for analysis, eliminating transcription costs.
The data obtained support some of the theories for understanding human error5 and the cognitive processes involved in its avoidance. These are worthy of consideration in self-audit of error checking as part of risk management in pharmacy practice.

Maidenhead, Berkshire

References

1. Spencer MG, Smith AP. A multicentre study of dispensing errors in British hospitals. Int J Pharm Pract 1993;2: 142-6.
2. Culshaw MA, Swanson DG. Introduction of a training scheme for technician final dispensing checks. Hosp Pharm 1994;1:126-7.
3. Private-Rx. The information network for UK pharmacists. At: http://www.private-rx.net/
4. NUD*IST 4.0 (Non numerical unstructured data indexing searching and theory-building) QSR Ltd, Australia.
5. Reason J. Human error. Cambridge: Cambridge University Press, 1990.