When it comes to correctional insulin administration in the hospital, the old adage, “Better late than never,” is simply unsafe. Timeliness is important concerning the critical interval between checking a patient's blood glucose (BG) reading and administering correctional insulin for the result. As the population living with diabetes in America grows, so does the number of people admitted to the hospital with diabetes as a comorbidity. Current best-practice recommendations guide hospitals to use insulin for diabetes management while patients are in the hospital.
Insulin is a high-risk medication, with potential for serious harm or even death when errors occur. A 2010 study revealed that the most-common medical errors in critical care patients were insulin administration errors.1 A survey conducted by the Institute for Safe Medication Practices (ISMP) in 2014 surveyed pharmacists and nurses, and showed that subcutaneous insulin ranked ninth among almost 40 drugs and drug classes identified as high-alert medications that concerned practitioners.2 Yet, of all the high-alert medications, subcutaneous insulin came in last place when pharmacists and nurses were asked to rank how confident they were regarding the effectiveness of hospital-wide precautions to prevent serious errors.2 The survey findings suggest a consensus among pharmacists and nurses that hospitalized patients are vulnerable to errors with subcutaneous insulin, and that more must be done to prevent patient harm with this high-alert medication.
Many insulin errors result in serious hypoglycemia, especially when point-of-care BG monitoring is not coordinated well with meals and insulin therapy. Coordinating insulin with meals and glucose monitoring in inpatient settings is a nationwide challenge.3 Studies suggest that the timing of glucose monitoring and insulin administration occur within an acceptable range less than half of the time in hospitalized patients prescribed insulin.4,5 Studies suggest that less than half of patients met the goal of receiving a rapid-acting insulin within 10 to 15 minutes of a meal, and only 35% received glucose monitoring within 1 hour prior to insulin administration.4,5 Timing for meals, BG testing, and rapid-acting insulin administration varied significantly and was not well synchronized among the various facilities. Coordinating BG monitoring with correctional insulin administration is a significant challenge in the inpatient hospital setting.
BG checks were performed at inconsistent times on the authors' unit; insulin was not being administered in coordination with BG checks, and the staff was unaware of the ISMP recommendations. This article outlines an intervention aimed at reducing the interval between BG checks and correctional insulin administration, with a target of 30 minutes or less of the BG check as per ISMP best-practice recommendations.
Baseline chart audits on a busy 19-bed cardiac intermediate care unit at a medium-sized nonprofit suburban hospital in the Southwestern US revealed that the timing of correctional insulin administration was suboptimal. When the authors surveyed nurses on the unit about the recommended time from checking BG level to administering correctional insulin, the answers were shocking: “1 hour...Before eating...As soon as I can...Whenever I get to it...I don't know.” The authors conducted chart audits on 477 correctional insulin administrations and found that the average time from BG check to correctional insulin administration was 52 minutes. According to the ISMP, correctional insulin should be given within 30 minutes of BG check.6 The root cause analysis revealed that not only were nurses unaware of the 30-minute recommendation, but also that patient-care technicians (PCTs) did not understand the significance of checking a BG too early. When BGs are checked too early, and a previous insulin dose is still in effect, the BG level may still be falling. This phenomenon, known as insulin stacking, is a significant contributor to hypoglycemia in the inpatient setting.7
For diabetes management at home, patients on insulin are taught to measure their BG, immediately take insulin, and then eat all in one coordinated process. Hospital practices create a vast discrepancy between what patients have become accustomed to with diabetes self-management at home, which can cause distrust and dissatisfaction in hospitalized patients with diabetes.
The primary goal of the authors' intervention was to reduce the time between BG checks and correctional insulin administration, with a target of 30 minutes or less. In the patient's electronic health record (EHR), the rapid-acting insulin given before meals is split into two separate administrations on the medication administration record (prandial and correctional). Both orders have separate administration instructions. Correctional insulin is ordered for hyperglycemia, and prandial insulin is ordered to cover carbohydrates consumed. For this project, the authors focused solely on correctional orders. The authors sought to improve documentation, teamwork, and interdisciplinary collaboration. To address this problem, they established a project lead and created a team including three day-shift nurses, three night-shift nurses, one day-shift PCT, and one night-shift PCT. Clinical nurse champions were guided by the hospital inpatient diabetes care and education specialist.
Based on feedback from nursing staff and their observations, the nurse champions identified several factors contributing to the gap in care. The first one was the most easily rectified: that nurses were unaware of the 30-minute standard of care. Another significant contributor was workflow and process issues related to the timing of meal trays and the timing of BG checks. An additional issue was the division of labor between PCTs and nurses. The authors identified that unit PCTs tended to lump BG checks with vital signs checks often resulting in BG checks being performed prematurely, sometimes an hour or more before unit meal trays were delivered. Best-practice recommendations suggest checking BGs as close to a patient's meal as possible.
The project team leader approached the problem by first educating staff on evidence and recommended practice guidelines. Education took place in informal short sessions taught by the project lead and champions. Education was required and recorded on a sign-in sheet. The team also created flyers to remind staff of the new process. Flyers were posted on the unit along with email communications. Staff had the opportunity to ask questions and share concerns.
When educated, nursing staff were highly motivated to find ways to rectify the problem. Using this energy, nurse champions solicited ideas to create a consensus-based and nurse-friendly process. First, the team collected data via a time study on when unit meal trays were delivered. The time study was completed over 2 weeks, and the unit secretary recorded when meal trays arrived on the unit. The authors found that meal trays delivery times were consistently reliable +/-15 minutes from 0800 for breakfast, 1200 for lunch, and 1700 for dinner. Using these data and staff input, the team set goal times for BG checks that coordinated within the 30-minute goal window. The group found that assigning BG checks and subsequent meal trays to specific times helped organize workflow and reduce care deviations. The team worked to ensure the new process and timing was easy to understand with BG checks occurring at set times of 0730, 1130, 1630, 2030. Reducing the focus on individual tasks assigned to different disciplines helped redirect attention to overall care issues. Unit staff devised innovative strategies to mark time and ensure everyone's focus remained on the 30-minute deadline, even as distractions and other priorities occurred on the busy unit. Their solution was to have the PCT announce via clinical communication device or in person that BG checks have been completed and the meter has been docked. This made all unit staff aware of the time countdown.
Predata were collected in November 2017 via chart audits. These data were then shared with the project team to design an intervention that occurred in December 2017. After staff education on the new process was completed in January 2018, the intervention went live in February 2018. The authors continued monitoring data each month in March and April 2018. Using these new processes, nurses were expected to assess BG results and administer insulin accordingly within 30 minutes of the results, as recommended by the ISMP. Another process change was that the second nurse who double-checks insulin administration also verifies that BG results are not older than 30 minutes. Results exceeding the 30-minute window require a recheck before insulin is administered. (See New standardized BG-check process.)
Staff received consistent, frequent feedback through baseline and ongoing chart audits, both individually and as a group. With the support of the unit manager, the team performed random audits to assess progress with this process. The team shared results of monthly audits and average time results with the nursing unit. Individualized feedback was shared with staff using the insulin administration feedback form. Nurses and PCTs with more than five out-of-range results in a month were addressed individually via email. Nurses and PCTs with fewer than five incidences were included in a drawing for a reward every month. Nurses were spurred to action primarily by their desire to improve outcomes. They appreciated having their efforts rewarded and were further prompted to action through peer competition. (See Insulin Administration Feedback Form.)
Results and outcomes
The bundle of nurse-led interventions improved all outcomes within a very short period. Using chi-square analysis and t-tests, the authors compared a 1-month baseline to a 2-month postintervention period for these three outcomes:
- BG check done at assigned time (Yes/No)
- Insulin administered within 30 minutes of BG check (Yes/No)
- Time (in minutes) from BG check to insulin.
With a sample size of 1,085 insulin administrations, statistically significant improvements with added clinical importance were identified. Preintervention data were gathered in November 2017, and postintervention data were collected from February to April 2018. There were 477 insulin administrations in the baseline period and 608 insulin administrations post intervention.
The EBP project improved the proportion of BG checks done on time from 45% to 92%. There was a statistically significant association between the intervention and whether BG checks were on time. Χ2 (1) = 282.573, P < .0001, OR 13.48, 95% CI [9.58-18.96]. Based on the odds ratio, the odds of BG checks being done on time was about 13 times higher after the intervention.
After implementing the intervention, the proportion of times insulin was given in the appropriate 30-minute window improved from 33.1% to 82%. This was a statistically significant improvement, Χ2 (1) = 267.907, P < .0001, OR 9.19, 95% CI [6.94- 12.17]. The odds ratio indicates that the odds of insulin being given within 30 minutes was about nine times higher after the intervention. (See Results.)
There were several key ingredients to this successful translation of research evidence into practice. First, the project capitalized on nurses' inherent quest for excellence. Nursing staff were surprised and dismayed to learn about existing standards and the gap in their practice. However, education alone was insufficient to change practice. A second ingredient was frequent updates. Consistent frequent feedback to all nursing staff was a powerful motivator for change. Frontline nurses became fully engaged in the project and were energized by daily or weekly successes. Another essential component was a strategy and process that involved nurses in chart audits. Peer involvement in the education and feedback transformed practice on a busy intermediate care unit. The final critical building block to success was passionate project leaders who were determined to overcome barriers. These frontline nurse leaders marshalled expertise, organizational skills, and enthusiasm to drive project activities. The combination of nurses' education, empowerment, and clinical leadership plus consistent performance updates propelled this project to success.
Challenges and limitations
Despite the success of this project, there were some limitations and challenges. Through the course of the project a few key barriers emerged. The BG meters used at this facility require docking to transmit the information to the EHR. As with all technology, especially when interfacing between multiple software programs, there can be delays in the transfer of information beyond our control. Some capabilities, such as meters with wireless uploading, could help in overcoming this identified challenge.
Also, because the unit is not a closed unit, float staff are often present to help meet staffing challenges. Because these staff are not based on the project unit, they were not included in the education related to this process change. In the course of this project, the authors identified a need to consistently educate float staff on expectations related to the unit process. Although project champions would try to talk to float staff about the process, this was inconsistent across shifts, especially when they were not working. Not following the new standardized process for both BG checks being performed on time and insulin given within 30 minutes often arose when float staff were on the unit and were unaware of the established new process. Other gaps occurred as the result of the busy nature of the unit with nurses having competing priorities for patient care. This unit also uses tray-delivery meal service. If this floor transitions to room service where patients can order food on demand, this would create a need to reevaluate the established process.
Additionally, the audits to provide data and consistent feedback to staff were time-consuming. An opportunity to explore further would be the creation of reports within the EHR that could pull this information. This study was limited to patients with provider orders for BG checks before meals and at bedtime who were eating by mouth. Patients with provider orders for BG checks every 4 or 6 hours or who were on tube feedings or parenteral nutrition were not included. Patients were not followed if transferred to another unit within the hospital.
The authors did not audit hyper- and hypoglycemic events during this study. Hypoglycemia data were not assessed during this study, as the authors did not have a reliable way to capture these data for the specific population on correctional insulin. Much of this work involved intense chart reviews to obtain accurate data. Although giving correctional insulin based on a BG that was greater than 30 minutes old could result in hypoglycemia, there are many other potential contributing factors to hypoglycemia in the hospital setting. Resources to intensively review each hypoglycemia event were not available as part of this project.
The scope of this project also did not include prandial or basal insulin administrations. Further research or evidence-based practice projects should include these aspects of insulin administration in the inpatient setting. This may prove challenging since the authors found that there is inconsistent documentation of carbohydrates consumed. However, in facilities that use carbohydrate-based prandial insulin orders, administration of insulin in relation to meals documentation will likely be more accurate.
Insulin is a high-risk medication, with potential for serious harm or even death when errors occur. Correctional insulin should be given within 30 minutes of BG check per the ISMP. Coordinating correctional insulin administration with BG monitoring in a busy cardiac intermediate care unit setting is a significant challenge for the care team. In order to decrease the time interval between BG check and insulin administration, it is important to establish clear expectations for staff. Staff education and unit champions are critical to achieve established goals in practice. The strategy design in order to achieve a desired evidence-based practice change must include input for direct care staff to overcome identified barriers and integrate the change into their daily practice. Monitoring successes and missed opportunities through chart audits and providing individual feedback are key strategies to close the gap between BG checks and insulin administration. Take advantage of identified gaps to drive practice improvement efforts on units. Partnerships between subject-matter experts (diabetes care and education specialists) and clinical nurses drove evidence-based practice improvement efforts.
1. Garrouste-Orgeas M, Timsit JF, Vesin A, et al. Selected medical errors in the intensive care unit: results of the IATROREF study: parts I and II. Am J Respir Crit Care Med
2. ISMP. Survey suggests possible downward trend in identifying key drugs/drug classes as high-alert medications. ISMP Medication Safety Alert!
3. Ryan DB, Swift CS. The mealtime challenge: nutrition and glycemic control in the hospital. Diabetes Spectr
4. Freeland B, Penprase BB, Anthony M. Nursing practice patterns: timing of insulin administration and glucose monitoring in the hospital. Diabetes Educ
5. Lampe J, Penoyer DA, Hadesty S, Bean A, Chamberlain L. Timing is everything: results to an observational study of mealtime insulin practices. Clin Nurse Spec
7. ISMP. ISMP Acute Care Guidelines for Timely Administration of Scheduled Medications. Institute for Safe Medication Practices