Inpatient Insulin Pen Implementation, Waste, and Potential Cost Savings: A Community Hospital Experience

Abstract

Background:

Insulin pen injectors (“pens”) are intended to facilitate a patient’s self-administration of insulin and can be used in hospitalized patients as a learning opportunity. Unnecessary or duplicate dispensation of insulin pens is associated with increased healthcare costs.

Methods:

Inpatient dispensation of insulin pens in a 240-bed community hospital between July 2018 and July 2019 was analyzed. We calculated the percentage of insulin pens unnecessarily dispensed for patients who had the same type of insulin pen assigned. The estimated cost of insulin pen waste was calculated. A pharmacist-led task force group implemented hospital-wide awareness and collaborated with hospital leadership to define goals and interventions.

Results:

9516 insulin pens were dispensed to 3121 patients. Of the pens dispensed, 6451 (68%) were insulin aspart and 3065 (32%) were glargine. Among patients on insulin aspart, an average of 2.2 aspart pens was dispensed per patient, but only an estimated 1.2 pens/patient were deemed necessary. Similarly, for inpatients prescribed glargine, an average of 2.1 pens/patient was dispensed, but only 1.3 pens/patient were necessary. A number of gaps were identified and interventions were undertaken to reduce insulin pen waste, which resulted in a significant decrease in both aspart (p = 0.0002) and glargine (p = 0.0005) pens/patient over time. Reductions in pen waste resulted in an estimated cost savings of $66 261 per year.

Conclusions:

In a community hospital setting, identification of causes leading to unnecessary insulin dispensation and implementation of hospital-wide staff education led to change in insulin pen dispensation practice. These changes translated into considerable cost savings and facilitated diabetes self-management education.

Introduction

Diabetes is a prevalent metabolic disorder in the United States, affecting over 34.2 million Americans as of 2018.¹ At this rate, it is predicted that by 2050 every third American will be diagnosed with diabetes.² The annual cost of lost productivity and wages stemming from diabetes is $327 billion, with a major driver ($97 billion, or 23% of the cost) in the form of inpatient care.² Insulin therapy is the preferred method of glycemic management in hospitalized patients with diabetes, and most require multiple daily insulin injections.^3,4

Traditionally, insulin was administered using a vial and syringe in the inpatient and outpatient settings.⁵ In 1985, insulin pen injectors (“pens”) were introduced to the market.⁶ The insulin pen is available as both reusable and disposable pens, but the more commonly used is the disposable type that is prefilled and used for multiple doses. These pens can be discarded when either empty or expired according to the manufacturer (eg, NovoLog FlexPen expires 28 days after first use). A growing number of insulin pens are currently available in the United States, and most of them are intended to facilitate safe and accurate patient self-administration of insulin.⁷ Insulin pens are used for many reasons, particularly in the inpatient setting, including 1) reducing the risk of erroneous dose (inaccurate dose); 2) ready for immediate administration; 3) decreasing nursing time to prepare and administer insulin; 4) decreasing needlestick injuries; and 5) preventing inadvertent use for more than one patient and/or cross-contamination.⁸

On the other hand, the use of insulin pens has several disadvantages, including the inability to mix compatible insulin types, which may increase the number of injections required, and inconvenience and pain for the patient.⁹ Also, unlike vial and syringes, it is necessary to prime each new insulin pen needle, also called an “air shot,” which represents a waste of two units of insulin prior to each injection.¹⁰ Moreover, pen needles are typically more expensive than insulin syringes, although typically only by a marginal amount.¹¹ Furthermore, the ease of staff in sharing insulin pens may inadvertently lead to pen-sharing among multiple patients, which may place individuals at risk for transmission of blood-borne pathogens.¹² Lastly, while insulin pens have been associated with lower overall healthcare utilization rates and costs in comparison to insulin vials, some hospitals have discontinued inpatient insulin pen use due to large upfront costs.¹³

There are several barriers to effectively and safely dispensing insulin within the hospital: narrow therapeutic index, risk of hypoglycemia, inaccurate medication reconciliation, multiple daily injections (with potentially differing doses at each time point), multiple types of insulin (eg, rapid-, short-, intermediate-, and long-acting), and prompt and potentially frequent dose adjustments during the hospitalization.¹⁴ The Institute for Safe Medication Practices considers insulin a high-alert medication as its use can lead to patient harm when used in error.¹⁰ Additionally, unnecessary or duplicate dispensation of medication can occur due to misplacement, miscommunication, or systematic error, resulting in increased cost, waste, and dosing delays.

It has been estimated that hyperglycemia can affect up to 38% of hospitalized patients, while 12% of this population has no known history of diabetes. Most of these patients will require at least one type of insulin.^15-18 However, to our knowledge no study to date has examined the cost or frequency of unnecessary or duplicate inpatient insulin dispensation (“waste”) by hospital pharmacies. This manuscript describes the implementation of insulin pens at the community hospital, examines the etiologies of insulin pen waste, and evaluates the effects of an intervention specifically targeted to reduce this waste.

Methods

Setting

This quality improvement project was carried out at Suburban Hospital, a 240-bed community hospital in the Johns Hopkins Health System located in suburban Maryland, between July 2018 and July 2019, to assess inpatient insulin pen dispensation.

Until 2018, Suburban Hospital only used insulin vials; the insulin pen program was implemented to align with practices at other health institutions within the health system. Our institution used two types of insulin pens: insulin aspart (NovoLog FlexPen) for bolus insulin and insulin glargine (Lantus SoloSTAR) for basal insulin requirements. The cost of each pen was $16.19 and $25.08, respectively, and each pen could hold 300 units of insulin.

Of note, in 2015, Suburban Hospital implemented an inpatient diabetes management service led by an Endocrine Hospitalist team, a multidisciplinary coordinated effort to manage glycemia of patients with diabetes, as previously described.^19-24

Implementation

A pharmacist-led insulin pen task force consisting of nurses, Epic Willow-certified systems analyst, certified diabetes care and educator specialist (formerly certified diabetes educator), hospitalist and endocrinologist implemented a hospital-wide awareness campaign and collaborated with the hospital leadership to define goals and interventions. Formal institutional review board approval was not required, as this was a quality improvement project not involving patient data. Additional information about the pre-, and post-launch of the insulin pens is presented in Table 1.

Table 1.

Timeline of Insulin Pen Implementation.

Pre-launch	• Multidisciplinary committee was prepared a year before the launch of insulin pens.
	• Acknowledged higher costs associated with insulin pens along with insulin pen waste.
	• Recognized that insulin vial and syringe method led to risk for needlestick injuries for nursing staff, despite the use of safety needles (unpublished data).
	• Multiple delays encountered throughout the process before the implementation.
During	• Met weekly and then monthly to review data and evaluate root causes.
	• Pharmacy began tracking daily patient insulin pen reports.
	• Epic Willow disabled defaulted duplicate label printing function and issued Best Practice Alerts with “Do Not Dispense.”
	• Pharmacy technicians monitored duplicate orders of insulin pens.
	• Mandated that all multi-dose medications such as insulin pens be kept in patient bins located inside BD Pyxis MedStation ES.
	• Nursing staff instructed to look for insulin pens upon patient transfer to another unit before asking pharmacy for a replacement pen.
Post-launch	• Nursing education made mandatory including teaching proper use of insulin pens.
	• Competency sheet for nursing staff.
	• Monthly insulin pen education incorporated into nursing orientation.
	• Diabetes educator performed impromptu checks to ensure proper use.
	• Diabetes nurse champions serve as resources for nursing staff.
	• Director of Pharmacy presented the project to executive team members, the Pharmaceutical and Therapeutic and Suburban Hospital Quality Committees 6 and 13 months post-implementation.

Insulin Pen Order Set

The insulin pen order set was applied and harmonized with two other Johns Hopkins Medicine institutions. The insulin pens were a part of the order set that was implemented in the clinical decision support tool within the electronic health record (EHR; Epic, Verona, WI) used to assist clinicians in the management of insulin, as previously described.¹⁴

Study Design

The number of aspart and glargine insulin pens dispensed were quantified via Epic medication record master file, ERX, which is a subset of the Chronicles database. A pen must be dispensed by the pharmacy and scanned at the bedside by a nurse to submit a charge status. Pens erroneously dispensed by the pharmacy (but returned by the nurse unused) would not be charged and therefore were not counted in the analysis.

For the purposes of the analysis, a pen was considered unnecessarily dispensed if the same type of pen was already dispensed within the preceding 48 hours. This cut-off was chosen based on the highest user amount after factoring in the availability of pharmacy resources and staffing. Manual review of individual patient records for necessary and unnecessary insulin dispensation was performed for the final three months in the dataset. The calculated three-month average of the “necessary” aspart and glargine pens/patient was extrapolated as an estimated value for the entire dataset.

To help understand the causes of insulin pen waste, informal interviews were conducted with Suburban Hospital staff.

Results

Insulin Pen Dispensation

Over the 12-month study period, 3121 patients received insulin, of which 2977 (95%) were ordered aspart and 1461 (47%) were ordered glargine. A total of 9516 insulin pens was dispensed, of which 6451 (68%) were aspart and 3065 (32%) were glargine. Among patients on aspart, an average of 2.2 aspart pens was dispensed per patient (monthly average range: 1.4–3.4 pens; Figure 1A). Similarly, for inpatients prescribed glargine, an average of 2.1 pens was dispensed by the pharmacy (monthly average range: 1.5–2.9 pens; Figure 2A). Linear regression analysis demonstrated a significant decrease in both aspart (p = 0.0002) (Figure 1B) and glargine (p = 0.0005) pens/patient over time (Figure 2B).

Figure 1.

Monthly inpatient dispensation of aspart pens. (A) Total numbers of aspart pens dispensed and inpatients prescribed aspart. (B) Number of aspart pens/patient. Linear regression analysis demonstrates that aspart pens dispensed/patient significantly decreased over time (p = 0.0002). Horizontal dotted line represents the estimated number of necessary aspart pens dispensed/patient (1.2 pens).

Figure 2.

Monthly inpatient dispensation of glargine pens. (A) Total numbers of glargine pens dispensed and inpatients prescribed glargine. (B) Number of glargine pens/patient. Linear regression analysis demonstrates that glargine pens dispensed/patient significantly decreased over time (p = 0.0005). Horizontal dotted line represents the estimated number of necessary glargine pens dispensed/patient (1.3 pens).

Over the final three months of the study period, an average of 1.2 aspart and 1.3 glargine pens/patient were deemed appropriately dispensed. Conversely, 19% of aspart pens (0.3 aspart pens/patient) and 22% of glargine pens (0.4 glargine pens/patient) were deemed unnecessarily dispensed.

Etiologies of Insulin Pen Waste and Challenges

The causes of insulin pen waste were identified via structured meetings and discussions with a multidisciplinary team, in addition to nursing and pharmacy staff being interviewed. Identified etiologies of insulin pen waste are discussed in Table 2. During a patient transfer from one unit to another, insulin pens were misplaced and required frequent dispensation of a new insulin pen. Within a patient’s assigned medication bin, medications at times were not being stored and returned to the BD Pyxis™ MedStation™ ES in the patient’s bin. Medications were occasionally stored in the patient room or elsewhere such as a nurse’s pocket for “convenience.” Insulin pens would be lost in patients’ rooms that were cleaned due to any items remaining being discarded. Although a tamper-evident seal is affixed to the insulin pen at the time of dispensing to indicate whether the pen had been used to help avoid waste of unused pens, the insulin pens were still misplaced, perhaps due to their convenient size and shape.

Table 2.

Etiologies and Challenges Associated with Insulin Pen Implementation.

Patient transfer	• During the transfer of patients from one unit to another insulin pens were misplaced and required frequent dispensation of a new insulin pen.
Patient’s assigned bin	• Medications were not being stored and returned to the BD Pyxis™ MedStation™ ES in the patient’s bin.
	• Medications were occasionally stored in the patient room or elsewhere for convenience.
	• Insulin pens could be lost in rooms that were cleaned and any remaining items were discarded.
	• Insulin pens were easily misplaced due to their convenient size and shape.
Delay in administering patient insulin	• Time was spent searching for the insulin pen (if not found in the patient’s medication bin).
	• Time was spent by the pharmacy for replacing the insulin pen as the dispensing process requires placing electronic “sticky notes” in the medication administration record to indicate that the medication was sent.
EHR system error	• A best practice alert had been set to not print a medication label for either new or modified aspart orders if the patient received a dose within the previous 72 hours.
	• This was not initially the case for glargine, as a label would print for each new order.
Patient-specific barcode	• The pharmacy team’s procedure was to cover the manufacturer’s barcode of each pen and to add a patient-specific barcode to prevent administering insulin to the wrong patient.

Financial Cost

For the first four months of post-insulin pen implementation, an estimated 58% of all insulin pens (aspart: 62%, glargine: 51%) were unnecessarily dispensed. The average monthly cost of this insulin pen waste was $11 820.68, which projects to a yearly cost of total waste of $141 848.16 (Figure 3).

Figure 3.

Estimated monthly cost of waste of unnecessary insulin pens.

After the main causes of insulin pen waste were identified, hospital-wide education was initiated, and the default insulin label printing function was disabled in Epic Willow, the monthly waste dropped to 48% in month 5 and 42% in month 6. The average monthly waste plateaued from months 7 to 12 at 21% (aspart: 18%, glargine: 25%). This represented an average monthly cost of waste of $2359. Extrapolated over 12 months, the cost of waste would have been $28 307. Overall, the cost of unnecessary insulin pen dispensation over the 12-month study period was estimated to be $75 588. Therefore, due to the insulin pen waste reduction strategies and education campaign, the hospital had an estimated potential cost savings of $66 261.

Discussion

In this study, we investigated the associated waste and cost of transitioning from insulin vials to pens in an inpatient setting, as well as the effects of a quality improvement effort to mitigate this waste. A methodological plan involving nursing education, task force communication, pharmacy leadership, and proper handling of insulin pens was key in this cost-saving effort.

Inpatient diabetes management teams may not be available at other institutions; caution should be taken in extrapolating these results.²⁵ In a large academic medical center, a quality improvement project concluded that with proper medication safety features in place there are major benefits of using insulin pens in the inpatient setting.²⁶ In a pilot study conducted at the University of California, San Francisco Medical Center, it was determined that using the EHR to directly integrate patient safety improvements regarding insulin pens offered an opportunity to enhance patient safety.²⁷ Furthermore, in a prospective, randomized study comparing insulin pens and the vial and syringe method of insulin administration in the hospital setting, was associated with decreased frequency of hyper- and hypo-glycemic events among the insulin pen group.¹¹

Insulin vials are associated with a greater risk of measuring inaccurate dosing units, as insulin syringes utilize small lines to mark the volume of insulin being drawn up. A nurse may mistake the quantity being drawn up, or miscalculate the conversion in the use of concentrated insulin (eg, U-500 vial), leading to potentially catastrophic consequences.^28,29 Insulin pens simplify and improve dosing accuracy, as the number of units is plainly displayed on the pen. Overall, the use of insulin pens has resulted in greater satisfaction among nurses, physicians, and patients.⁵ In another study, the handling and dosing accuracy of insulin pens versus vials and syringes were evaluated and it was discovered that insulin pens were easier to administer for insulin-naïve healthcare providers.³⁰ The insulin pen devices were also associated with delivering more accurate doses and fewer variations than vials and syringes.³⁰ Additionally, the use of insulin pens in the hospital allows patients to have multiple opportunities to practice insulin self-administration with the type of delivery device they may go home on. This inpatient education may better prepare patients to follow the post-discharge insulin regimen and thereby lower the risk of emergency department visits and 30-day readmission rates.^31,32 Our research may be of interest to other hospital administrators and clinicians interested in introducing insulin pens to their respective clinical settings.

There are numerous strategies to reduce the waste and costs linked to insulin administration. These include: 1) improving communication between providers, nurses, and pharmacists; 2) implementing a medication-specific financial and utilization report; 3) nursing education and quality improvement initiatives (such as keeping insulin pens in designated medication bins, and not in nurse pockets or on a counter in the patient room); 4) establishing a mechanism to securely transfer insulin pens from the emergency department to the general medical/surgical floor; and 5) hospitals having an outpatient dispensing license, which would allow insulin pens used during inpatient stay to be dispensed upon discharge. Lastly, it is essential to review the insulin usage/waste data on a recurrent basis, ideally in a structured multidisciplinary team meeting (eg, Glucose Steering Committee and/or Pharmacy & Therapeutics Committee), to assess the current and historical levels of insulin waste and enact action items as needed.

Challenges

While many hospital pharmacies may re-label insulin pens before discharge for patients to take home, the rules surrounding this practice vary from state to state, and our institution was not licensed to dispense any medications for the outpatient setting. Consequently, all duplicate/unnecessarily dispensed pens, even if unused, had to be destroyed upon patient discharge. Patients were unable to take any inpatient medications to their homes. See Supplemental Material S1 for a video interview with a nurse discussing the challenges.

Limitations

Our study has several limitations. First, patients with very high insulin doses requiring frequent insulin pen dispensations may have inadvertently been flagged as duplicate dispensation. Second, we did not evaluate the frequency of insulin vial waste (pre-pen implementation) to see if the amount or costs were similar to insulin pen waste. Third, other variables impacting the amount of insulin wastage may exist and could have been unaccounted for. Fourth, the included data were analyzed retrospectively. Finally, it was challenging to calculate the full cost of insulin pen waste, as the cost of nursing and pharmacy time and effort is difficult to quantify. Additionally, we were unable to factor in the cost of medical waste cartage, as the information available was separated by hazardous and non-hazardous material but was not medication-specific.

Conclusions

In a community hospital setting, identification of causes leading to unnecessary insulin dispensation and implementation of a multi-pronged quality improvement effort led to a change in insulin pen dispensation practice and reduction in waste. These changes may be translated into considerable cost savings.