Impact of a focussed teaching programme on practical prescribing skills among final year medical students

Abstract

AIM

To assess the impact of prescribing teaching on final year medical students.

METHODS

Students randomly allocated to two hospitals completed a prescribing assessment. Prescribing teaching was delivered to the intervention group while no additional teaching was provided for the control group. All students then completed a second prescribing assessment.

RESULTS

Teaching improved the assessment score: mean assessment 2 vs. 1, 70% vs. 62%, P = 0.007; allergy documentation: 98% vs. 74%, P = 0.0001; and confidence. However, 30% of prescriptions continued to include prescribing errors.

CONCLUSION

Medical students make significant errors in prescribing. Teaching improves ability and confidence but is insufficient alone in eradicating errors.

WHAT THIS STUDY ADDS

• Focussed prescribing teaching can lead to an improvement in prescribing ability.

• Prescribing confidence can be significantly improved through education.

• Education is insufficient alone in eradicating prescribing errors.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Medication errors, and particularly prescribing errors, are common in UK hospitals.

• Junior doctors make the majority of prescribing errors.

• Deficiencies in prescribing education and training have been closely linked to the high frequency of medication errors.

Introduction

The Audit Commission's report ‘A Spoonful of Sugar’ highlighted the frequency of medication errors in UK hospitals [1]. With over 80 000 medication incidents reported annually to the UK National Patient Safety Agency (UK NPSA), they are the third most common cause of patient safety incidents [2]. Individual and system failures have been identified but poor prescribing is likely the most common cause [3].

Junior doctors make the majority of prescribing errors. They also make most prescribing decisions despite a lack of experience [4, 5]. Establishing the rate of medication errors is difficult, not least due to differences in error definition [6]. A recent systematic review reported an error rate by junior doctors of 2–514 per 1000 prescription items and 4.2–82% of patient charts reviewed [7].

Clinical pharmacologists have often led the debate on the importance of prescribing education [8], but medical students and newly qualified doctors have themselves highlighted a lack of prescribing education [9, 10] and prescribing confidence [11]. A study commissioned by the General Medical Council (GMC) examining the preparedness to practice of newly qualified doctors graduating from three UK medical schools identified prescribing as a significant weakness [12]. Although a number of recommendations for undergraduate prescribing education have been published [13–15], there remains little guidance as to how these competencies should be achieved.

At the University of Edinburgh prescribing education is delivered under the remit of Clinical Pharmacology and Therapeutics. We assessed current prescribing standards among final year students, as well as the impact of a focused doctor- and pharmacist-led prescribing teaching programme.

Methods

The study group comprised final year medical students at the University of Edinburgh (July 2008–July 2009). University administration staff randomly allocated students to two teaching hospitals to undertake an 8 week module in general medicine. All students completed a baseline prescribing assessment in week 1 consisting of six scenario-based questions. Topics were identified as typical scenarios frequently encountered by Foundation Year 1 doctors. A stringent marking scheme was agreed by consensus between medical and pharmacy staff (see Supplementary Methods in Supporting information).

During the 8 week module, students in the intervention group received focused doctor- and pharmacist-led practical prescribing teaching, in addition to the standard teaching programme already provided. This consisted of four 1 h tutorials with specific emphasis on practical prescribing. Included was a detailed discussion of the assessment 1 scenarios including example kardexes. A further 1 h per week of prescribing teaching was delivered at the bedside as part of the student's clinical teaching, providing a total of 12 additional hours of prescribing teaching. Students in the control group received no additional prescribing teaching over that already provided.

In week 8, a second prescribing assessment including six different questions was completed by both groups (see Supplementary Methods in Supporting information). Five students from each group were excluded having not completed both assessments leaving an intervention group of 50 students and a control group of 28 students.

Students completed a questionnaire at the start and end of the module to assess prescribing confidence on a scale of 1 (no confidence) to 5 (very confident). The study protocol was approved by the local research ethics committee. Statistical analyses were undertaken using Student's t-test and two-tailed Fisher's exact test where appropriate.

Results

The scale of the problem

Assessment 1 examined students at baseline and the results are shown in Table 1. Both groups completed more than 90% of all available drug kardexes. However, between 35% and 44% contained prescribing errors, with dose errors being most common (Table 1).

Table 1.

Results of the assessments for the teaching and control groups

		Teaching group (n = 50)			Control group (n = 28)
		Assessment 1	Assessment 2	P value	Assessment 1	Assessment 2	P value
Prescriptions completed
Total available drug kardexes	n	300	300		168	168
Drug kardexes completed	n (%)	284 (95)	300 (100)		154 (92)	168 (100)
Prescriptions completed	n	691	1099		407	553
Prescriptions per student	Mean (95% CI)	14 (13, 14)	22 (21, 23)	<0.0001	15 (13, 16)†	20 (18, 21)‡	<0.0001
Allergy documentation	Mean % (95% CI)	40 (27, 52)	98 (96, 100)	<0.0001	62 (44, 79)§	74 (59, 90)¶	NS
Prescribing confidence	Mean (95% CI)	1.5 (1.3, 1.6)	3.8 (3.7, 3.9)	<0.0001	1.5 (1.2, 1.8)	3.2 (2.9, 3.5)**	<0.0001
Prescribing errors
Total prescribing errors	n	707	608		302	302
Dose errors	n (%)	346 (49)	334 (55)	0.07	214 (71)	196 (65)	NS
Route errors	n (%)	91 (13)	91 (15)	NS	30 (10)	45 (15)	NS
Typographical errors*	n (%)	269 (38)	182 (30)	0.009	57 (19)	60 (20)	NS
Prescriptions including errors	n (%)	305 (44)	349 (32)	<0.0001	141 (35)††	168 (30)‡‡	NS
Errors per prescription	Mean (95% CI)	1.0 (0.8, 1.3)	0.6 (0.5, 0.7)	0.0003	0.8 (0.6, 0.9)§§	0.6 (0.4, 0.7)	NS

The number of drug kardexes completed by each group is given as a proportion of total available kardexes. The remaining data demonstrate how each groutp performed in the two assessments.

^*Typographical errors refer to errors involving the signature, date, time, or inappropriate abbreviations.

^†NS (not significant) for the mean number of prescriptions per student teaching vs. control group assessment 1.

^‡P = 0.009 for the mean number of prescriptions per student teaching vs. control group assessment 2.

^§P = 0.003 for allergy documentation teaching vs. control group assessment 1.

^¶P = 0.0001 for allergy documentation teaching group vs. control group assessment 2.

^**P = 0.0002 for prescribing confidence teaching vs. control group assessment 2.

^††P = 0.002 for the proportion of prescriptions containing errors in the teaching vs. control group assessment 1.

^‡‡NS for the proportion of prescriptions containing errors in the teaching vs. control group assessment 2.

^§§NS for the mean number of errors per prescription for the teaching vs. control group assessment 1.

Improvement in prescribing

Following clinical exposure a significant improvement in assessment score was observed in all students (Figure 1). The degree of change was more prominent in the teaching group who, despite achieving a lower mean score in assessment 1 (mean (95% CI) assessment 1 score teaching group vs. control group: 49% (46, 52) vs. 56% (52, 60), P = 0.01, Figure 1), demonstrated a significantly higher mean score in assessment 2 (70% (67, 73) vs. 62% (56, 68), P = 0.007, Figure 1). The improvement in the teaching group's score was associated with a significant reduction in the mean number of errors per prescription (Table 1). These improvements, however, only brought the teaching group in line with results already demonstrated by the control group.

Figure 1.

Bar chart demonstrating overall assessment scores for the teaching and control groups in assessment 1 and assessment 2. Data shown are mean score (%) with error bars representing 95% confidence intervals (95% CI). Teaching Group (); Control Group ()

Allergy documentation improved in all students, but to a greater degree in the teaching group [mean (95% CI) assessment 2 teaching group vs. control group: 98% (96, 100) vs. 74% (59, 90); P = 0.0001; Table 1]. No difference in prescribing confidence was observed at baseline (Table 1), but significantly higher post-teaching confidence levels were reported by the teaching group (mean (95% CI) teaching group vs. control group: 3.8 (3.7, 3.9) vs. 3.2 (2.9, 3.5), P = 0.0002, Table 1).

Prescribing errors persist

Despite improved assessment scores, allergy documentation and confidence, prescribing errors continued to be present in 30% of prescriptions with dose errors continuing to make up the largest proportion (Table 1).

Discussion

Principal findings

Consistent with previous data [5, 7], up to 44% of all prescriptions at baseline included errors. Significant improvements in prescribing performance were observed in all students following clinical exposure, with additional improvements noted in those who received prescribing teaching. Despite this, 30% of prescriptions in assessment 2 continued to include prescribing errors. Improved confidence in the presence of continued prescribing errors may suggest an inappropriate level of confidence among students with significant implications for patient safety.

Strengths and limitations of the study

The major strength of this study was the inclusion of a control group to account for any temporal improvements in students' ability as they progressed through training, allowing the true effect of the teaching programme to be assessed. Comparable with other studies in this area [16–18], almost 50% of all medical students in the final year group were recruited with a minimal dropout rate.

Our study has some limitations. Comparison between groups was complicated by the control group achieving better baseline scores, even after random allocation of students who were not stratified according to academic performance. This is a problem with studies of this type where formal randomization or controlling for students' academic ability is not possible. The assessments were also conducted in a controlled environment, without the external distractions present on a busy hospital ward, a recognized problem in these studies [19].

Meaning of the study

Prescribing is a complex task that requires theoretical knowledge combined with practical skill. Despite this it is often considered a skill that students will develop themselves through exposure to the clinical environment. Failure of undergraduate prescribing education is closely linked to the frequency of medication errors made by junior doctors [3, 19, 20]. However, there remains a lack of consensus on the skills required for safe and rational prescribing, and a valid and reliable mechanism for assessment [19].

We, like others [18, 21] have shown that specific prescribing teaching can lead to improvements in prescribing ability and confidence. Even with additional teaching, however, a significant proportion of prescriptions continue to include errors. While undergraduate education is important, changes in training are likely only to be part of the solution. Systems-orientated interventions like electronic prescribing may increase prescribing quality through reducing prescribing errors and pharmacists' interventions [22], but their use is not without risk [23].

Eradication of all errors in any system is impossible. It is important, however, to identify causes and minimize risk. Education has an important role but there may be a ceiling to the benefit it can achieve. There is a need for a robust multi-faceted system-based approach to minimize the risk associated with prescribing errors.

Competing Interests

There are no competing interests to declare. No funding was required for this study.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Appendix S1 Supplementary Methods.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.