Abstract
While several studies have shown that insulin pens are more convenient and accurate than conventional administration with syringes and vials (syringes/vials), there is a frequent need for low-dose insulin injections administered by nurses using syringes/vials in hospital settings, particularly for critically ill patients. However, there is a lack of research investigating factors related to the accuracy of low-dose insulin administration using syringes/vials, particularly in hospital settings. We therefore performed a cross-sectional study to assess the accuracy of low-dose insulin administration by registered nurses using syringes/vials and to determine whether time of day, years of experience and adherence to proper injection procedures (vertical insertion/drawing and air bubble checking) affected the accuracy. The participants were 33 registered nurses working in the diabetes ward, and a total of 198 trials were analyzed. Using syringes/vials, the median errors converted to insulin units were found to be 0.6 units for the 2- and 6-unit target doses, and 0.7 units for the 10-unit target dose. In cases with the largest error, errors for the 2-, 6-, and 10-unit target doses were observed to be 2.3, 4.0, and 3.3 units, respectively. In our study, time of day, years of experience and vertical insertion/drawing did not correlate with errors, but errors were significant in the participants who did not check for air bubbles. Nurses can make non-negligible dosage errors when administering low-dose insulin using syringes/vials, and this is particularly likely when air bubble checks are missed.
Supplementary Information
The online version contains supplementary material available at 10.1007/s13340-024-00726-5.
Keywords: Insulin, Syringe, Vial, Inpatient, Accuracy
Introduction
Since the introduction of insulin pen in 1985 [1], several studies have shown that insulin pens are easier and more convenient for patients to use than conventional administration with syringes and vials (syringes/vials). [2–4]. Moreover, several reports have examined the accuracy of insulin pens versus syringes/vials, and the results have shown that insulin pens are more accurate when used by patients [5] or healthcare professionals [6]. Thus, insulin pens are commonly prescribed for outpatients who require long-term insulin use.
On the other hand, there is a frequent need for low-dose insulin injections administered by nurses using syringes/vials in hospital settings, particularly for critically ill patients. For example, nurses may need to administer low-dose insulin to patients during treatment of infectious disease or in the perioperative period, when the patient's appetite is diminished or fasting is required.
Since insulin solution is typically measured as 0.01 mL = 1 unit, and even minor errors can lead to significant variations in insulin dosage, resulting in hypoglycemia or hyperglycemia, the Institute for Safe Medication Practices considers insulin a high-alert drug [7, 8]. However, studies investigating factors associated with the accuracy of low-dose insulin administration using syringes/vials are lacking, especially in hospital settings.
We therefore performed a cross-sectional study to assess the accuracy of low-dose insulin administration by registered nurses using syringes/vials and to determine whether time of day, years of experience and adherence to proper injection procedures (vertical insertion/drawing and air bubble checking) affected the accuracy.
Methods
Participants
This cross-sectional study was performed at Yamaguchi Red Cross Hospital, a 333-bed regional hospital in Yamaguchi City, Japan. The study was approved by the Institutional Review Board of Yamaguchi Red Cross Hospital (approval number Y202103). The recruitment of participants was conducted over the 4-weeks period between April and May 2021. A total of 33 registered nurses working in the diabetes ward participated in the study. The ward primarily cares for patients with diabetes requiring treatment for hyperglycemia, infectious diseases, renal failure, and other conditions. The nurses use insulin pens as well as syringes/vials for insulin delivery. The participants had a wide range of experience with insulin injection, ranging from one to 34 years (interquartile range (IQR) = 3, 18). They were divided into three groups according to the experience quartile range of the participants.: least- (1–3 years, n = 9), intermediate- (4–18 years, n = 16), and most experienced (> 19 years, n = 8). All participants provided written informed consent to participate in this study.
Materials
The 0.5 mL insulin syringes for U-100 insulin (Myjector, TERUMO, Tokyo, Japan), and U-100 concentration vials of regular insulin (Novolin R, Novo Nordisk A/S, Bagsvaerd, Denmark) were used. The syringes were marked in 1-unit increments. A Novolin R FlexPen (Novo Nordisk A/S) was used with a 34-gauge 4-mm needle (Nanopass, TERUMO). Insulin was brought to room temperature before use.
Assessment of dosing accuracy
The participants performed a simulated administration in the area where they always prepare insulin injections, as described below. All simulated administrations were conducted during the participants' shifts to approximate the conditions of actual operation. Using the syringes/vials, the participants drew up doses of 2, 6, and 10 units of insulin twice each: before lunch and dinner, when insulin injections for patients were usually prepared (66 cases per unit). In a control experiment using FlexPen, which was considered to ensure accuracy, doses of 2, 6, and 10 units were each dispensed twice by four randomly selected participants (eight cases per unit). Quantification of insulin was performed in accordance with previous reports [9]. Briefly, a beaker with a liquid paraffin layer was placed on an electronic precision balance (LIBROR EB-330H, SHIMADZU, Kyoto, Japan) and set to zero before each insulin dose was dispensed and weighed. Weighed insulin was converted to insulin units (0.01 g = 1 insulin unit) as in previous report [6, 9]. Dosing accuracy was defined as follows:
absolute percentage error =| (target dose—actual dose dispensed) ÷ target dose |× 100.
All cases (66 syringes and eight pen cases per unit) were included in the comparison of syringes/vials and pen dosing accuracy.
Assessment of injection procedures
For one participant classified as intermediate-experienced, the injection procedure could not be assessed due to missing data. To check the participants' injection procedures, three nurses who were not involved in the study observed the following two factors: inserting the needle vertically into the vial and keeping the syringe and vial vertically while drawing up insulin, and checking for air bubbles in the syringe. The judges did not interrupt participants during the process regardless of their level of adherence to the proper injection procedures. A majority decision was taken in the cases of disagreement between the three judges.
Statistical analysis
Group comparisons were performed using the Mann–Whitney U test and Kruskal–Wallis test, as appropriate. The Bonferroni method was used to correct p-values for multiple comparisons. In the box-and-whisker plot, the box is drawn from the lower to the upper quartile, with a horizontal line drawn in the middle to represent the median. Data were analyzed using EZR version 1.61 (Saitama Medical Centre, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria) [10]. Results were considered statistically significant at p < 0.05.
Results
Figure 1 shows the absolute percentage error with syringes/vials and pen insulin. Over the dose range studied, pen insulin showed a lower absolute percentage error than syringes/vials (Fig. 1a-c, all p < 0.001). With syringes/vials, the absolute percentage error decreased with increasing insulin dose, i.e. the lower the target dose, the greater the absolute percentage error (p < 0.001 for 2 vs. 6 units and 2 vs. 10 units, p = 0.0012 for 6 vs. 10 units) (Fig. 1d). Comparing syringes/vials and pen insulin, the median absolute percentage errors were 30.0% (IQR = 15.0, 40.0) vs. 6.7% (IQR = 2.9, 8.8), 10.0% (IQR = 6.7, 12.1) vs. 2.7% (IQR = 0.9, 4.0) and 7.0% (IQR = 5.0, 9.0) vs. 2.1% (IQR = 1.5, 3.1) for the 2-, 6- and 10-unit target dose, respectively. In syringes/vials, the median error converted to insulin unit were 0.6 units for 2- and 6-unit target dose, and 0.7 units for 10-unit target dose. Focusing on the cases with the largest errors, the errors were 2.3, 4.0 and 3.3 units for the 2-, 6-, and 10-unit target doses, respectively.
Fig. 1.
Accuracy (%error) with syringe/vial and pen insulin. The data for 2-unit. a 6-unit. b and 10-unit. c target dose are shown (n = 8 trials in pen, n = 66 trials in syringe/vial each). Over the dose range studied, pen insulin had lower %errors than syringe/vial (all p < 0.001). d With syringe/vial, the %error diminished with increasing doses of insulin: 2 > 6 > 10 units of insulin (p < 0.001 for 2 vs. 6 units and 2 vs. 10 units, p = 0.0012 for 6 vs 10 units). (a-c) ** p < 0.001 by the Mann–Whitney U-test. d * p < 0.01, ** p < 0.001 by the Kruskal–Wallis test with Bonferroni correction
Figure 2 shows a comparison of dosing accuracy for syringe/vial administration by time of day. There was no difference in dosing accuracy by time of day. Supplementary Fig. S1 shows the results of each participant's before-lunch and dinner trials.
Fig. 2.

Comparison of dosing accuracy for syringe/vial administration by time of day. There was no difference in dosing accuracy by time of day (n = 99 trials both before lunch and dinner)
Figure 3 shows analyses of association between dosing accuracy for syringe/vial administration and three factors: years of experience as a registered nurse, vertical insertion/drawing and air bubble checking, based on all data. In terms of adherence to proper procedures, in 14% of all 192 trials, vertical insertion/drawing was not adhered to and in 42% of all 192 trials, the syringe was not checked for air bubbles. Years of experience and vertical insertion/drawing did not affect dosing accuracy for syringe/vial administration (Fig. 3a-b). On the other hand, the median percentage errors between actual and target doses were significantly greater for participants who did not check for air bubbles (p < 0.001) (Fig. 3c). The results of the analysis of the association between dosing accuracy and each factor by target dose are shown in Supplementary Fig. S2-4.
Fig. 3.
Analyses of association between dosing accuracy for syringe/vial administration and three factors: years of experience as a registered nurse, vertical insertion/drawing and air bubble checking. The association between dosing accuracy for syringe/vial administration and three factors: years of experience as a registered nurse (a, n = 198 trials), vertical insertion/drawing (b, n = 192 trials) and air bubble checking (c, n = 192 trials) are shown. There was no difference in dosing accuracy by years of experience as a registered nurse and vertical insertion/drawing. On the other hand, the median percentage errors between actual and target doses were greater in participants who did not check for air bubbles (p < 0.001). (c) ** p < 0.001 by the Mann–Whitney U-test
Discussion
In our study, we made two important clinical observations.
First, in all experimental conditions using the syringes/vials, although the median error between the target and actual insulin dose was less than one unit, the error was more than two units in some participants.
Second, checking for air bubbles has a significant impact on the dosing accuracy of insulin administration using a syringe/vial.
In our study, the insulin dosage was accurate, even with the syringes/vials, however, there were sometimes large errors. In terms of the median absolute percentage error, our data showed 30% for 2 units, 10% for 6 units, and 7% for 10 units. The errors between the target and actual insulin doses were slightly larger than those in previous reports of nurses. Gnanalingham et al. [9] evaluated the accuracy of syringes/vials in 1, 2, 5, and 10 units with five pediatric nurses; the mean errors were 23%, 12%, 3.6%, and 0.7%, respectively [9]. Keith et al. [11] examined the accuracy of 2 and 5 units of insulin administration by two investigators: one to draw up the doses and another to weigh them. The mean error was 12–15% for two units and 1.2–3.6% for five units [11]. Asakura et al. studied the delivery of 10 units of insulin by 30 insulin therapy-experienced healthcare professionals and 20 insulin therapy-naive healthcare professionals; the mean error was 1.8% for the former and 2.6% for the latter [6]. Differences between the previous studies and ours may be attributed to a unique feature of our study, namely that we experimented with conditions as close as possible to those of daily clinical practice: in the middle of the participants' shift, by the same procedures and at the same time as routine insulin injections. Moreover, the larger error in our study may be related to the fact that we surveyed a larger number of nurses to assess accuracy, including human error, while previous reports had fewer investigators to focus on the differences in accuracy between syringes/vials and pens.
Large dosage errors were not necessarily made by participants who were judged not to have checked for air bubbles. Even participants who were judged to have checked for air bubbles sometimes made large dosage errors. These results can be explained by the fact that they checked for air bubbles but did not remove enough of them, or that they did not check for air bubbles but there were few air bubbles initially. Alternatively, participants might misunderstand the target dose in trials where large errors occurred. Based on the results of each participant's two trials (Supplementary Figure S1), we do not assume that any large errors were made solely due to low technical proficiency of the participants, since no participant made a large error in both trials. In fact, the adherence rates for air bubbles checking and vertical insertion/drawing were similar or superior to the previous report from Greece [12], which investigated compliance with international guidelines. Unfortunately, our study did not elucidate why some participants failed to adhere to the proper procedure, but there appears to be room for improvement in dosing accuracy by reminding the participants of the proper injection procedure.
Asakura et al. showed that the dosing accuracy of insulin therapy-experienced healthcare professionals was superior to that of insulin therapy-naive healthcare professionals [6]. The nurses in our study had a wide range of experience, ranging from one to 34 years. We initially assumed that there would be differences in dosing accuracy between the least experienced and the other two groups, however, we found no differences. This discrepancy may be due to the fact that participants in the least experienced group were not entirely inexperienced with insulin injections.
While previous studies have reported that medication errors are more likely to occur at specific times of the day [13, 14], we found no differences in accuracy between the time periods in which the trials were conducted. These differences may be more pronounced in the actual setting, where insulin injections must be prepared in the middle of busy medication rounds, rather than in an experimental setting. Morning insulin injections require special attention because they are prepared and administered during the night shift, when fewer nurses are available, and errors are more likely to occur [15].
Our study and several reports suggest that pen use is preferable in terms of accuracy, convenience, and time efficiency [16–19], but multiple factors are involved, making it difficult to determine the superiority of pens over syringes/vials during hospitalization. We also have to consider the cost. In Japan, as of April 2024, the prescription cost is 2,640 yen for a U-100 concentration vial (containing 1000 units of Novolin R) and 1296 yen for a pen (containing 300 units of Novolin R). If pens are used for patients who do not require a lot of insulin, the amount of wasted insulin will be substantial. Thus, from a drug cost perspective, the cost difference between the vials and pens is likely to be even greater when vials are used for multiple patients. On the other hand, the study of Shogbon-Nwaesei suggests that the use of pens saves nursing time, resulting in reduced labor costs, and that a complete economic cost analysis needs to take into account other direct and indirect costs of each delivery method [16]. There is also the safety issue of infection from pens being inadvertently shared among patients, which is why one hospital decided to switch from pens to vials [20].
Our study had several limitations. The small sample size and the fact that all participants belonged to only one ward in one hospital may have increased the variability of individual data. Therefore, our results need to be clarified in future studies, either by comparison with other wards or by external, larger-scale validation. Another limitation is that only one type of syringe was used, while other syringes may yield different results. In addition, we need to consider the Hawthorne effect: a phenomenon where participants improve their performance in response to the awareness of being observed when they take part in research [21]. Therefore, although our study was performed with registered nurses in conditions as close to clinical practice as possible, dosage errors may be more significant in real-world situations in which nurses must interact with other staff, patients and medication storage. Finally, further studies are needed to clarify whether emphasizing the importance of air bubble checking to nurses can reduce dosage errors.
Conclusions
Nurses can make non-negligible dosage errors when administering low-dose insulin using syringes/vials. Dosage errors can occur at any time, regardless of experience with insulin injection, and are particularly likely to occur when air bubble checks are missed. When administering low-dose insulin using syringes/vials, in addition to checking the indicated insulin dose, an emphasis on checking for air bubbles can help to prevent dosage errors.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to thank Editage (www.editage.com) for English language editing.
Author contributions
All authors contributed to the study conception, design and analysis. Material preparation and data collection were performed by Rika Saji, Yuka Yamada, Mari Hayashi, and Tomoko Shirai. The first draft of the manuscript was written by Manabu Kondo and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Data availability
Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.
Declarations
Conflict of interest
None of the authors has any potential conflict of interest associated with this research.
Ethical approval
The study was approved by the Institutional Review Board of Yamaguchi Red Cross Hospital (approval number Y202103, March 26, 2021). Also, the study was in accordance with the Helsinki Declaration of 1964 and its later versions.
Informed consent
The recruitment of participants was conducted over the 4-weeks period between April and May 2021. All participants provided written informed consent to participate in this study.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Hyllested-Winge J, Jensen KH, Rex J. A review of 25 years’ experience with the NovoPen family of insulin pens in the management of diabetes mellitus. Clin Drug Investig. 2010;30:643–74. 10.2165/11584360-000000000-00000 [DOI] [PubMed] [Google Scholar]
- 2.Korytkowski M, Bell D, Jacobsen C, Suwannasari R. A multicenter, randomized, open-label, comparative, two-period crossover trial of preference, efficacy, and safety profiles of a prefilled, disposable pen and conventional vial/syringe for insulin injection in patients with type 1 or 2 diabetes mellitus. Clin Ther. 2003;25:2836–48. 10.1016/S0149-2918(03)80337-5 [DOI] [PubMed] [Google Scholar]
- 3.Coscelli C, Lostia S, Lunetta M, Nosari I, Coronel GA. Safety, efficacy, acceptability of a pre-filled insulin pen in diabetic patients over 60 years old. Diabetes Res Clin Pract. 1995;28:173–7. 10.1016/0168-8227(95)01092-R [DOI] [PubMed] [Google Scholar]
- 4.Summers KH, Szeinbach SH, Lenox SM. Preference for insulin delivery systems among current insulin users and nonusers. Clin Ther. 2004;26:1498–505. 10.1016/j.clinthera.2004.09.009 [DOI] [PubMed] [Google Scholar]
- 5.Lteif AN, Schwenk WF. Accuracy of pen injectors versus insulin syringes in children with type 1 diabetes. Diabetes Care. 1999;22:137–40. 10.2337/diacare.22.1.137 [DOI] [PubMed] [Google Scholar]
- 6.Asakura T, Seino H, Nakano R, Muto T, Toraishi K, Sako Y, et al. A comparison of the handling and accuracy of syringe and vial versus prefilled insulin pen (FlexPen). Diabetes Technol Ther. 2009;11:657–61. 10.1089/dia.2009.0006 [DOI] [PubMed] [Google Scholar]
- 7.Institute for Safe Medication Practices. ISMP guidelines for optimizing safe subcutaneous insulin use in adults. 2017. https://www.ismp.org/guidelines/subcutaneous-insulin. Accessed 2 November 2023.
- 8.Cousins D, Rosario C, Scarpello J. Insulin, hospitals and harm: a review of patient safety incidents reported to the National Patient Safety Agency. Clin Med (Lond). 2011;11:28–30. 10.7861/clinmedicine.11-1-28 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gnanalingham MG, Newland P, Smith CP. Accuracy and reproducibility of low dose insulin administration using pen-injectors and syringes. Arch Dis Child. 1998;79:59–62. 10.1136/adc.79.1.59 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kanda Y. Investigation of the freely available easy-to-use software “EZR” for medical statistics. Bone Marrow Transplant. 2013;48:452–8. 10.1038/bmt.2012.244 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Keith K, Nicholson D, Rogers D. Accuracy and precision of low-dose insulin administration using syringes, pen injectors, and a pump. Clin Pediatr (Phila). 2004;43:69–74. 10.1177/000992280404300109 [DOI] [PubMed] [Google Scholar]
- 12.Theofanidis D. In-hospital administration of insulin by nurses in Northern Greece: an observational study. Diabetes Spectr. 2017;30:175–81. 10.2337/ds16-0001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Isaacs AN, Ch’ng K, Delhiwale N, Taylor K, Kent B, Raymond A. Hospital medication errors: a cross-sectional study. Int J Qual Health Care. 2021. 10.1093/intqhc/mzaa136. 10.1093/intqhc/mzaa136 [DOI] [PubMed] [Google Scholar]
- 14.Manias E, Street M, Lowe G, Low JK, Gray K, Botti M. Associations of person-related, environment-related and communication-related factors on medication errors in public and private hospitals: a retrospective clinical audit. BMC Health Serv Res. 2021;21:1025. 10.1186/s12913-021-07033-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tehewy MA, Fahim H, Gad NI, Gafary ME, Rahman SA. Medication administration errors in a university hospital. J Patient Saf. 2016;12:34–9. 10.1097/PTS.0000000000000196 [DOI] [PubMed] [Google Scholar]
- 16.Shogbon-Nwaesei AO, Jacob BC, Kimble LP, Peasah SK, Burtner J, Moody L, et al. Time-motion analysis of nurses using insulin pen devices compared to vial and syringe in an inpatient setting. J Pharm Pract. 2019. 10.1177/0897190017743132. 10.1177/0897190017743132 [DOI] [PubMed] [Google Scholar]
- 17.Costa SD, Brackenridge B, Hicks D. A comparison of insulin pen use in the United States and the United Kingdom. Diabetes Educ. 2002;28(52–6):59–60. [DOI] [PubMed] [Google Scholar]
- 18.Davis EM, Bebee A, Crawford L, Destache C. Nurse satisfaction using insulin pens in hospitalized patients. Diabetes Educ. 2009;35:799–809. 10.1177/0145721709340057 [DOI] [PubMed] [Google Scholar]
- 19.Sheldon KA, Seoane-Vazquez E, Szeinbach SL, Tubbs C. Exploring risk and ease of use for insulin delivery by nurses. J Eval Clin Pract. 2010;16:199–201. 10.1111/j.1365-2753.2009.01131.x [DOI] [PubMed] [Google Scholar]
- 20.Trimble AN, Bishop B, Rampe N. Medication errors associated with transition from insulin pens to insulin vials. Am J Health Syst Pharm. 2017;74:70–5. 10.2146/ajhp150726 [DOI] [PubMed] [Google Scholar]
- 21.Purssell E, Drey N, Chudleigh J, Creedon S, Gould DJ. The Hawthorne effect on adherence to hand hygiene in patient care. J Hosp Infect. 2020;106:311–7. 10.1016/j.jhin.2020.07.028 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.


