Background
Reason's (1990) Human Error theory offers a generic explanation of error. Error happens when a planned activity fails in outcome and when the failure is not due to chance. Reason argues that all errors can be categorised as slips, lapses, violations or mistakes. Barber (2002) suggests the theory may help to understand non-compliance in medicine-taking. In this study we tested Barber's application in undergraduates taking Vitamin C tablets.

Method
We used a prospective cohort design, with data collected from participants before taking any tablets (T1) and 10 days later (T2) at the end of the "course". We recruited 100 undergraduate students using opportunity sampling (posters and snowballing). At T1 participants completed three measures that might predict non-compliance: the Cognitive Failures Questionnaire (CFQ) to assess the frequency of everyday cognitive failures; the Social Motivation Questionnaire (SMQ) to assess mild social deviance; and a bespoke Theory of Planned Behaviour questionnaire (TPB) to assess intentions to take the tablets. Participants were given a bottle of 20 vitamin C tablets and instructed to take 2 each day, morning and night. At T2 compliance was assessed by participant recall, pill count and Horne's MARS-5 scale. Any participant who had missed a tablet or taken it at a different time to instructed, was helped to classify it, using Reason's categorisation of error. To aid recall, we did this for the first and last instances of non-compliance only.

Results
T2 data were available for 84 (84%) participants. The 3 measures of non-compliance varied and showed that 14% did not miss any doses (by self-recall), 19% (by MARS-5), 11% (by pill count). Of the first missed doses, 70% were classified as unintentional (34% slips, 36% lapses) and 30% were intentional (24% mistakes, 6% violations). Of the last missed doses, 54% were unintentional (34% slips, 20% lapses) and 46% were intentional (22% mistakes, 24% violations). None of the 3 T1 measures (CFQ, SMQ, TPB) was a strong predictor of non-compliance. Participants who used a strategy to help them remember to take the tablets were more likely to do so (p<.05).

Conclusions
The three measures of compliance gave different ratings. The measures we chose were not good predictors of compliance. Non-compliance is complex, as evidenced by the change in rates of intentional and unintentional errors between the first and last error doses. Reason's theory offers a useful means of categorising errors in medicine-taking and the possibility of increasing our understanding of it. Moreover, targeted strategies based on knowledge of the specific underlying problem, i.e. intentional or unintentional compliance, are likely to lead to greater efficacy in intervention. It would be useful to compare the findings from studies of people taking a prescribed medicine.

References
Barber, N. (2002) Should we consider non-compliance a medical error? Quality and Safety in Health Care, 11, 81-84.
Reason, J.T. (1990) Human Error. Cambridge University Press: Cambridge