OBJECTIVES

Insulin is a high-risk medication, and its dosing depends on the individualized clinical and nutritional needs of each patient. Our hospital implemented an insulin dose calculator (IDC) imbedded in the electronic medical record with the goal of decreasing average wait times in inpatient insulin ordering and administration. In this study, we evaluated whether implementation of an IDC decreased the average wait time for insulin administration for hospitalized pediatric patients.

METHODS

This pre- and postintervention cohort study measured wait times between point-of-care glucose testing and insulin administration. Patients admitted to the inpatient pediatric services who were treated with subcutaneous insulin during the study period were included. Additionally, nurses completed satisfaction surveys on the insulin administration process at our hospital pre- and post-IDC implementation. Descriptive statistics, χ2, Fisher’s exact test, and Student t tests were used to compare groups. Statistical process control charts were used to analyze data trends.

RESULTS

The preintervention cohort included 79 insulin doses for admitted pediatric patients. The postimplementation cohort included 128 insulin doses ordered via the IDC. Post-IDC implementation, the average wait time between point-of-care glucose testing and insulin administration decreased from 37 to 25 minutes (P < .05). The statistical process control chart revealed a 5-month run below the established mean after implementation of the IDC. Before IDC implementation, 15.6% of nurses expressed satisfaction in the insulin-dosing process compared with 69.2% postimplementation (P < .05).

CONCLUSIONS

Implementation of an IDC reduced the average wait time in ordering and administration of rapid-acting insulin and improved nursing satisfaction with the process.

In the United States, 2.4 per 1000 people aged <20 years are estimated to have insulin-dependent diabetes mellitus, and this number continues to increase.1  The rates of hospitalization among children and young adults with diabetes have also increased by 38% between 1993 and 2004.2  Insulin, discovered in 1921, remains the standard of care for glycemic control across hospital settings.1,2  The Institute for Safe Medication Practices classifies all forms of insulin as high-risk medications and recommends instituting safeguards to reduce the risk of errors and minimize harm with these medications through standardization of ordering, storage, preparation, and administration of these medications.3 

Rapid-acting insulin dosing requires complex calculations based on blood sugar corrections and carbohydrate intake.4  In the hospital setting, rapid-acting insulin administration is challenging because of variable nutritional intake, changing clinical status, and frequent blood glucose monitoring. The American Diabetes Association recommends dosing rapid-acting insulin 15 to 30 minutes before eating.5  In a previous study, Slattery et al6  reported that the optimal timing for subcutaneous insulin administration is 15 to 20 minutes before meals, with later administration of insulin increasing the risk of postprandial hypoglycemia.

Sliding scale insulin has been employed to avoid wait times in insulin administration; however, sliding scales require consistent carbohydrate intake per meal.5  Improved communication between patients, bedside nurses, and providers are cost-effective interventions to decrease insulin-dosing wait times.7  However, such interventions are difficult to standardize. Previous research revealed that electronic medical record (EMR)–imbedded insulin dose calculators (IDCs) do not cause an increased risk of hypoglycemia.8  This is the first study in which the efficiency of an EMR-imbedded IDC has been evaluated. The authors hypothesized that the IDC would decrease the average wait time for insulin administration for hospitalized pediatric patients. The IDC provides an innovative approach to the complex and time-consuming process of dosing, ordering, and administering rapid-acting insulin in the hospital setting.

A multidisciplinary team of pediatric resident physicians, hospitalists, pharmacists, endocrinologists, and nurses was assembled to design the IDC tool and implement its use. After the tool was developed, the multidisciplinary team met to identify potential safety concerns, develop a hospital workflow for implementation of the new tool, and design a staff education plan. Information technologists, pharmacists, physicians, and end users performed multiple rounds of testing of the new IDC to evaluate the accuracy of calculations, ease of use, clarity of technology, and new workflow. This team provided live education sessions to all pediatric and psychiatric providers, pediatric pharmacists, and nurses on included units. Education sessions were mandatory for all staff and were offered over a 2-month period before and during implementation. The hospital safety event reporting system was used to review safety concerns.

Before implementation of the IDC, a resident physician ordered each rapid-acting insulin dose after receiving notification from the patient’s bedside nurse of blood glucose measurement and estimated carbohydrate intake. On systems analysis, the process of dosing, ordering, and administering rapid-acting insulin was 7 steps with 6 areas of potential delay (Fig 1).

Implementation of the IDC within the EMR shifted many of the steps for dosing of rapid-acting insulin to the time of admission (Fig 1). In the new workflow, pediatric endocrinology is consulted on admission for all patients with insulin-dependent diabetes mellitus. The standardized consultant note outlines the following parameters: time-specific blood glucose targets, hyperglycemia correction factors, and meal-specific carbohydrate coverage (Supplemental Fig 4). These parameters match the fields within the electronic prescription of the IDC (Supplemental Fig 5).

After verification of the IDC order by pharmacy, a patient-specific insulin vial is dispensed. Once the medication is received and undergoes bar code medication scanning, the bedside nurse enters the patient’s current blood glucose measurement, selects the appropriate corresponding mealtime (breakfast, lunch, dinner, or snack), and enters the number of carbohydrates the patient plans to consume (Supplemental Fig 6). The IDC tabulates the appropriate dose on the basis of the parameters outlined by the physician’s order. As a safety measure, the dose is capped at 0.4 U/kg per dose. In addition, a second nurse checks the input information and calculated dose before administration. In both the pre- and postintervention groups, blood glucose was measured by using point-of-care testing that integrates electronically with the EMR. Similarly, time of insulin administration is recorded on barcode scanning in both groups.

The cohorts were derived from data obtained via EMR reports of rapid-acting insulin administration for patients aged ≤18 years at a 110-bed tertiary care, urban, imbedded children’s hospital. These data included patients admitted to the general pediatric floor, pediatric psychiatric unit, and PICU. The preintervention cohort represented rapid-acting insulin doses administered between September 2017 and March 2018. The postintervention cohort included rapid-acting insulin doses ordered via the IDC in the 6 months postintervention (October 2018 to April 2019). In the postintervention cohort, the data were reviewed retrospectively on a monthly basis. The institutional review board approved this study.

This work is a retrospective, observational, pre- and postintervention cohort study. The times between most recent blood glucose measurement, insulin order placement, and insulin administration were extracted from the EMR and medication administration record. Subcutaneous insulin was primarily used in the general pediatric and psychiatric units. Although subcutaneous insulin is administered in the PICU, this usually precedes transfer to a lower level of care. Subcutaneous insulin ordering and administration procedures did not differ between hospital units at our institution. Statistical process control charts were plotted for the pre- and postintervention to assess the impact of the IDC tool on process efficiency. An X-bar chart, used for continuous data on average wait times, was analyzed for consecutive points above or below the established mean to identify meaningful change attributable to the implementation of the IDC. P-charts, used for proportional data on doses administered in allotted time frames (<30 minutes, 30 to 60 minutes, and >60 minutes), were analyzed for successive points outside of the previously established control limits to suggest change beyond special cause variation.

To evaluate nursing satisfaction with insulin use at our hospital, pediatric nurses completed a modified 5-point Likert scale survey pre- and post-IDC implementation. An anonymous 6-question survey was developed by using Google Forms online software (Alphabet, Inc, Mountain View, CA). Nurses were asked to estimate wait time between meal delivery and insulin order availability, gauge patients’ and/or families’ satisfaction with the wait time associated with insulin processes, estimate how often an insulin dose required multiple phone calls to providers, and estimate the frequency of discrepancy between the nurse-calculated and provider-calculated insulin dose. Finally, nurses were asked to rank their satisfaction with the current process for insulin ordering and administration. Surveys were distributed via e-mail to the nursing pool on all 3 hospital units. The preimplementation survey was open for 4 weeks in April of 2018, and the postimplementation survey was open for 4 weeks in April of 2019. Surveys were anonymous, and consent was not obtained. One initial e-mail introducing the survey and 3 reminder e-mails were sent to the nursing pool. Descriptive statistics, χ2 tests, Fisher’s exact tests, and Student t tests were used to compare groups.

The preintervention group included 79 individual insulin doses from 14 patients. The postintervention cohort included 128 individual insulin doses from 16 patients (Table 1). Distribution of gender was similar between groups; however, a difference was seen between the racial and ethnic breakdown of the pre- and postintervention cohorts.

Before implementation of the IDC, the average wait time between blood glucose testing and insulin ordering was 21 (SD: 21) minutes. The average wait time between blood glucose measurement and insulin administration was 37 (SD: 29) minutes (Fig 2). In the postintervention period, IDC use eliminated the wait time between blood glucose measurement and insulin ordering. The average wait time between blood glucose measurement and insulin administration was 25 (SD: 26) minutes, representing a 12-minute reduction (P < .05). Wait times plotted on an X-bar chart reveal a sustained run below the mean for the 5-month period postimplementation (Fig 2). Additionally, the percentage of insulin doses given in <30 minutes increased from 51% preintervention to 68.8% postintervention (P < .05). Figure 3 highlights the increase in mean postintervention for insulin dose administered in <30 minutes and subsequent decreased mean in doses administered in >60 minutes.

Nursing surveys were distributed to 119 members of the pediatric nursing staff. Thirty-two (27%) nurses completed the preintervention survey. Thirty (25%) nurses completed the postintervention survey. Preintervention survey results revealed that 73.5% of nurses were “very unsatisfied” with the process before implementation of the IDC. After implementation, no nurses answered “very unsatisfied.” The most common response postintervention was “very satisfied” (45%), compared with no nurses responding “very satisfied” preintervention (P < .05) (Table 2).

Through a multidisciplinary approach, our children’s hospital sustained a significant decrease in the wait time between blood glucose measurement and insulin administration for pediatric patients with insulin-dependent diabetes mellitus. Implementation of the IDC streamlined the arduous process for inpatient insulin dosing, ordering, and administration. Review of the institutional safety event reporting system revealed no significant safety events associated with use of the IDC. Overall, this project revealed improvement in insulin-dosing efficiency with minimal safety concerns.

In 2019, the Centers for Medicare and Medicaid Services mandated that all eligible hospitals adopt electronic health record technology.9  Given the near-universal use of EMRs in the United States, this IDC technology provides an important resource for interventions to improve the quality of inpatient diabetes care.10  Our study reveals significant improvement in inpatient insulin management using EMR technology.

Although a significant change in the average wait times between blood glucose measurement and insulin administration was noted in this study, there is room for improvement. From our data analysis, the percentage of doses administered after >30 minutes was 32.3%. We hypothesize that many factors may contribute to persistent wait times. One, a cultural shift among nursing and housestaff on these units may be required for continued change. Because significant wait times have been tolerated for so long, more education may be needed for nursing and pediatric resident physicians to reinforce the importance of timely insulin dosing. Two, many pediatric patients do not eat on a regular meal schedule. However, nursing staff require blood glucose measurement at specific breakfast, lunch, and dinner sweep times. Therefore, a breakfast blood glucose measurement may occur long before the patient desires to eat breakfast. Three, patients admitted with a new diagnosis of diabetes often require teaching on insulin administration that increases the time between blood glucose monitoring and insulin administration. Four, nursing availability for dosing review and documentation may impact ongoing wait times. Since the completion of this study, a reminder has been added to the ordering instructions to advise nursing staff that the maximum recommended time between the most recent blood glucose measurement and IDC administration is 30 minutes.

Limitations to this study include the absence of glucose monitoring to determine the IDC’s impact on glycemic control. Specific research should focus on episodes of hypo- and hyperglycemia associated with the IDC. Each dose of insulin was not an independent factor but may represent multiple doses to the same patient or from the same physician. Nursing perception of patient satisfaction was used as a surrogate for patient and family satisfaction. Future studies should be used to directly assess patients’ and families’ perceptions. Although the IDC revealed improved efficiency in insulin administration, further study on the overall impact on inpatient length of stay is an important area for future investigation. To the best of our knowledge, no concurrent interventions were undertaken at the time of this study that may confound the data. Although patients were selected from 3 different clinical units, the workflow for subcutaneous insulin administration is the same across all areas.

The implementation of an IDC within the EMR significantly decreased patient wait times for insulin administration. This intervention also increased both nursing satisfaction with the insulin use process and nursing perceived patient satisfaction with the insulin ordering process. The input of a multidisciplinary team was integral to the successful design and implementation of the IDC.

Dr Yanell built and tested the EPIC tool. We also acknowledge Michael Yanell, PharmD and the institutional review board of Rush University Medical Center, Rush Children’s Hospital, Chicago.

FUNDING: No external funding.

Dr Ruddock-Walker conducted all postintervention data analysis, drafted the initial manuscript, and approved the final manuscript; Dr Shaaban conceptualized this study and contributed to, edited, and approved the final manuscript; Drs Ruddock-Walker and Shaaban conducted all pre- and postintervention chart review and preintervention data analysis; Drs Meltzer, Minutti, and Jacobson contributed to, edited, and approved the final manuscript; Dr Hovey conceptualized the electronic medical record tool and led its development and implementation, developed and administered the nursing satisfaction survey, and edited and approved the final manuscript; and all authors approved the final manuscript as submitted.

Deidentified individual participant data will not be made available.

1.
Menke
A
,
Orchard
TJ
,
Imperatore
G
,
Bullard
KM
,
Mayer-Davis
E
,
Cowie
CC
.
The prevalence of type 1 diabetes in the United States
.
Epidemiology
.
2013
;
24
(
5
):
773
774
2.
Lee
JM
,
Okumura
MJ
,
Freed
GL
,
Menon
RK
,
Davis
MM
.
Trends in hospitalizations for diabetes among children and young adults: United States, 1993 2004
.
Diabetes Care
.
2007
;
30
(
12
):
3035
3039
3.
Institute for Safe Medication Practices
.
ISMP list of high-alert medications in acute care settings
.
2018
.
4.
American Diabetes Association
.
Standards of medical care in diabetes--2014
.
Diabetes Care
.
2014
;
37
(
Suppl 1
):
S14
S80
5.
Moghissi
ES
,
Korytkowski
MT
,
DiNardo
M
, et al
.
American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control
.
Endocr Pract
.
2009
;
15
(
4
):
353
369
.
6.
Slattery
D
,
Amiel
SA
,
Choudhary
P
.
Optimal prandial timing of bolus insulin in diabetes management: a review
.
Diabet Med
.
2018
;
35
(
3
):
306
316
7.
Wilde
MI
,
McTavish
D
.
Insulin lispro: a review of its pharmacological properties and therapeutic use in the management of diabetes mellitus
.
Drugs
.
1997
;
54
(
4
):
597
614
.
10.2165/00003495-19975 4040-00006
8.
Ateya
MB
,
Aiyagari
R
,
Moran
C
,
Singer
K
.
Insulin bolus calculator in a pediatric hospital. Safety and user perceptions
.
Appl Clin Inform
.
2017
;
8
(
2
):
529
540
9.
Centers for Medicare and Medicaid Services
.
Promoting interoperability programs
.
10.
Henry
J
,
Pylypchuk
Y
,
Searcy
T
,
Patel
V
.
Adoption of electronic health record systems among U.S. non-federal acute care hospitals ONC data brief 35
.
2016
.

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.