Aim DRC software aims to reduce dosing errors,¹ however, it has been linked to ‘alert fatigue’, a phenomenon which causes prescribers to potentially override alerts despite being clinically relevant.^{2 3} The project aim is to determine the effectiveness of the DRC software implemented into the electronic prescribing software used in a paediatric tertiary hospital.

Method A retrospective clinical audit was undertaken to investigate the DRC alerts produced during July and August 2020. The DRC alert subtypes and the top 10 drugs that produced alerts were described. Alerts generated due to a ‘missing value’ (height, weight, or dose unit) were counted but not analysed further. Dose-based alerts were clinically screened to ascertain whether they were appropriately or inappropriately overridden. This involved considering whether the DRC recommendations differed from BNFc recommendations, local in-house guidance, or deviations occurred due to individual patient idiosyncrasies or specialist prescribing. Data from dose-based alerts was interrogated to determine whether increasing the acceptable dose range limit from 5% to 10% would have significantly reduced the number of alerts fired. Alerts that were inappropriately overridden and resulted in a medication error were categorised by severity using the EQUIP study scoring tool.³

Results 1778 alerts generated in July and August 2020 were analysed. 48% (n=846) of those alerts were produced due to a ‘missing value’. The DRC software did not recognise whether the alerted drug was recorded at the same dose in the patients’ ‘home medications’ (verified drug history) or whether in-house dose guidance was used. If recognised, the number of alerts would have reduced by 21.4% (n=200). Conversely, increasing the DRC acceptable dose range from 5 to 10% would have reduced the number of alerts only by 4.5% (n=42).

Overall, 741 alerts were clinically screened. 95% (n=704) of these were not actioned by prescribers. Of those alerts 5.2% (n=37) should have been acted upon and this led to medication errors. 35% (n=13) of the errors were significant and 22% (n=8) were serious according to the EQUIP study tool. 62.5% (n=5) of serious medication errors involved a ‘Narrow Therapeutic Index’ drug, such as gentamicin and liposomal amphotericin. 38% (n=5) of significant errors related to no prescribed maximum frequency for intravenous ondansetron.

Conclusion DRC systems are effective tools for preventing prescribing errors,¹ but it is concerning to see that several significant and serious prescribing errors occurred despite an alert generation. This potentially suggests that alert fatigue may counteract the error preventing effects of DRC alerts. Therefore, further refinement of DRC systems is required to reduce alert fatigue. Unfortunately, increasing the acceptable dose range limits from 5 to 10% does not appear to be a simple way of sorting this problem. Removing ‘missing value’ alerts would significantly reduce the number of alerts generated. Including in-house guidelines alongside BNFc dose recommendations into the DRC software would also reduce unnecessary alerts.

References

1. Sahota H, Hughes P, Barrass C, et al. Delivering electronic prescribing and medicines administration in challenging areas such as paediatrics and maternity at King’s College Hospital NHS Foundation Trust. Clinical Medicine Journal [Internet] 2015;15:s5. Available at: https://www.rcpjournals.org/content/clinmedicine/15/Suppl_3/s5

2. Neame M, Moss J, Saez Dominguez J, et al. The impact of paediatric dose range checking software. European Journal of Hospital Pharmacy [Internet] 2020. Available at: https://ejhp.bmj.com/content/early/2020/03/31/ejhpharm-2020-002244

3. Dornan T, Ashcroft D, Heathfield H, et al. An in-depth investigation into causes of prescribing errors by foundation trainees in relation to their medical education. EQUIP study final report. December 2009. Available at: https://www.gmc-uk.org/-/media/documents/FINAL_Report_prevalence_and_causes_of_prescribing_errors.pdf_2893 5150.pdf