Frequency and nature of prescribing errors made by final year medical students

Andrew Hitchings, Dagan Lonsdale, Emma Baker. St George's, University of London, London, UK

Introduction: Prescribing errors are found in 8–10% of prescriptions written by UK foundation doctors [1]. Few studies have sought to characterise the frequency and nature of prescribing errors made by senior medical students; none have done so using methodology comparable with that used in recent studies in doctors [1,2]. Our institution has recently implemented a prescribing skills test, which all final year students must pass in order to graduate. A formative version of this, comprising 10 free-text prescribing questions, is run early in the final year to facilitate students’ preparation.

Aim: In this study, the first part of the PRescribing Errors PrequAlification to diRect Education (PREPARE) study, we sought to characterise the frequency and nature of prescribing errors made by final year medical students in the context of a formative prescribing assessment.

Methods: All 2012/13 final year students were required to attempt the assessment under examination conditions. The scripts were marked against a predefined mark scheme and compared with an Angoff-derived notional pass mark, set at graduation level. Written informed consent was sought from students to subject their answers to detailed prescribing errors analysis, which was performed on a randomly-selected sample. To describe the error rate, two denominators were prospectively defined: ‘prescribing acts’ (comprising all prescriptions written, reviewed, or incorrectly omitted), and ‘medication orders’ (all prescriptions written or reviewed). Errors were classified by severity and type as described elsewhere [1].

Results: 277 students (95% of the cohort) gave written informed consent to participate in the prescribing errors study. 120 exam scripts were selected at random for prescribing errors analysis. 6185 prescribing acts were reviewed, 1958 (32%) of which contained at least one error (38 errors per 100 prescribing acts). There were 5743 medication orders, of which 1517 (26%) contained at least one error (33 errors per 100 orders). Using prescribing acts as the denominator, the severity of errors was: minor 971 (41% of errors), significant (40%), serious 365 (16%), life-threatening 82 (3%). The most common errors were omission of a required medication (19%), failure to discontinue a drug where required (12%), failure to specify duration of therapy where required (7%), and use of a non-approved drug name (6%) or prescribing abbreviation (6%). The number of errors per 100 prescribing acts correlated with examination mark (R=–0.760, P<0.001) and was significantly lower among students who achieved a ‘passing’ mark (35 per 100 acts) than for those who ‘failed’ (44 per 100 acts; P<0.001).

Conclusion: We evaluated prescribing errors made by final year medical students, using methodology comparable with that used in recent studies in doctors. Errors were observed in 32% of prescribing acts and 26% of medication orders, and this correlated with examination performance. The error rate was higher than that observed in a study on foundation year doctors [1]. More research is required to determine how students’ and doctors’ prescribing competence evolves, and how this can be facilitated.

[1] Dornan T et al. GMC 2009. http://www.gmc-uk.org/about/research/research_commissioned_4.asp

[2] Avery T et al. GMC 2012. http://www.gmc-uk.org/about/research/12996.asp