More information about text formats
We thank Dr Jones and Professor Franklin’s insightful and constructive response to our systematic review of the incidence and prevalence of intravenous medication errors in the UK. We appreciate and are grateful for their consideration and the opportunity to respond to their observation.
They are absolutely right to suggest that this is an example of both the limitations of our systematic review methodology, and the importance of grey literature accessing wider datasets as part of these reviews. Our protocol allowed us to contact authors of papers for more detailed data, however did not provide for occasions where authors were not contactable, and did not include grey literature. Two independent data extractors flagged that the data in the original publication  was ambiguous. When it was clear that further data was not accessible through direct contact with the author a consensus decision was taken to present only the data that we could reliably associate with IV medication errors from the paper and acknowledge this limitation.
As the correspondents rightly suggest, by supplanting the thesis data into the analysis, we identify 1773 intravenous doses, and 789 errors, resulting in a weighted prevalence estimate of 451/1000 administrations (95% CI 420–482), however the limitations related to the definition and operationalisation of errors, and how they affect estimates still hold, particularly around the impact of including “wrong time” errors into the...
As the correspondents rightly suggest, by supplanting the thesis data into the analysis, we identify 1773 intravenous doses, and 789 errors, resulting in a weighted prevalence estimate of 451/1000 administrations (95% CI 420–482), however the limitations related to the definition and operationalisation of errors, and how they affect estimates still hold, particularly around the impact of including “wrong time” errors into these syntheses.[2,3]
This experience has guided us well in subsequent reviews. In a systematic review on the prevalence and nature of drug-related problems in hospitalised children and young people in England  the protocol explicitly permitted the use of grey literature and thesis data in line with Cochrane recommendations. Thus we identified the summarised nature of Ghaleb’s data in the published article and extracted granular data direct from the thesis.
We join Jones and Franklin in reminding future reviewers to be mindful of the data that exists in grey literature such as theses and government reports, and to ensure that strategies are incorporated into protocols and search strategies to accommodate these important data sources.
1 Ghaleb MA, Barber N, Franklin BD, et al. The incidence and nature of prescribing and medication administration errors in paediatric inpatients. Arch Dis Child 2010;95:113–8. doi:10.1136/adc.2009.158485
2 McLeod MC, Barber N, Franklin BD. Methodological variations and their effects on reported medication administration error rates. BMJ Qual Saf 2013;22:278–89. doi:10.1136/bmjqs-2012-001330
3 Keers RN, Williams SD, Cooke J, et al. Causes of Medication Administration Errors in Hospitals: a Systematic Review of Quantitative and Qualitative Evidence. Drug Saf 2013;36:1045–67. doi:10.1007/s40264-013-0090-2
4 Sutherland A, Phipps DL, Tomlin S, et al. Mapping the prevalence and nature of drug related problems among hospitalised children in the United Kingdom: a systematic review. BMC Pediatr 2019;19:486. doi:10.1186/s12887-019-1875-y
5 Lefebvre C, Glanville J, Briscoe S, et al. Chapter 4: Searching for and selecting studies. In: Higgins J, Thomas J, Chandler J, et al., eds. Cochrane Handbook for Systematic Reviews of Interventions. Cochrane 2021. http://www.training.cochrane.org/handbook
We read this article with great interest. Given differences among countries in preparation and administration practices for intravenous medicines, it is an important contribution to the literature.
Sutherland et al. state that their calculation of the incidence of intravenous medication errors may be an underestimate (1), as they were not able to clearly differentiate intravenous from non-intravenous administrations in the study by Ghaleb et al. (2) We are writing to highlight refined data related to the Ghaleb et al. study, which will be useful to readers interested in interpreting the review’s findings.
Specifically, Table 2 of the review reports that in the Ghaleb et al. study, 85 infusions (5.5%) of a total of 1,554 contained at least one error. However, the Ghaleb et al. study reports data relating to all routes of administration, and not just the intravenous route (2). The total of 1,554 observed doses therefore includes both intravenous and other routes of administration, with the number of intravenous doses not reported in the published paper. Consequently, the incidence of intravenous medication errors reported Table 2 for the Ghaleb et al. study is artificially low. This is likely to considerably influence the systematic review’s pooled estimate of the incidence of intravenous medication errors, (1) as Ghaleb et al. (2) contributes 60% of the observations included in this calculation.
The PhD thesis on which the paper by Ghaleb et al. is bas...
The PhD thesis on which the paper by Ghaleb et al. is based contains more information. (3) It states that 751 intravenous doses were observed, which therefore represents a more appropriate denominator than 1,554 to use in calculating the incidence of intravenous medication errors.
As noted in the footnote to Table 2 of the systematic review (1), it was only possible to extract information on ‘intravenous administration rate’ errors from the Ghaleb et al. paper, (2) giving a total of 85 such errors. However, more detail is available from the thesis, which states that 190 intravenous errors were observed in total. (3) This number might therefore be a more appropriate numerator for the incidence of intravenous medication errors in the Ghaleb et al. study, although the thesis does not state if the 190 intravenous errors occurred in 190 intravenous doses, or if more than one error was observed during some administrations.
These refined data result in a maximum intravenous error incidence of 25% for the Ghaleb et al. study (190 errors in 751 intravenous doses). Use of these refined data is therefore likely to result in a considerably higher pooled estimate of the incidence of intravenous medication errors than that originally calculated in the systematic review. (1)
This case serves to highlight some of the potential limitations of systematic review methodology and the importance of drawing on supplementary data and grey literature in selected situations when specific details are not available in the peer-reviewed article.
1. Sutherland A, Canobbio M, Clarke J, et al. Incidence and prevalence of intravenous medication errors in the UK: a systematic review. Eur J Hosp Pharm 2018;27:3-8. http://dx.doi.org/10.1136/ejhpharm-2018-001624
2. Ghaleb MA, Barber N, Franklin BD, et al. The incidence and nature of prescribing and medication administration errors in paediatric inpatients. Arch Dis Child 2010;95(2):113-8. http://dx.doi.org/10.1136/adc.2009.158485
3. Ghaleb MAA. The incidence and nature of prescribing and administration errors in paediatric inpatients [PhD]. University of London, 2006.