Determining the Need for Standardized Infusions and Concentrations at Hospitals within North Carolina

Abstract

Background

In 2013, the American Society of Health-System Pharmacists (ASHP) endorsed a policy recommending the development of nationally standardized drug concentrations and dosing units. Although many hospitals have started standardizing their intravenous (IV) solutions, few have taken the practice beyond their institutions or local geographical areas.

Objective

This project evaluates the current IV standardization practices for adult patients across hospitals in North Carolina and compares them with each other. In addition, this project proposed standards and evaluated them for their impact on reducing observed variability.

Methods

In the fall of 2013, an e-mail request was sent to select hospital pharmacy leaders in North Carolina asking them to voluntarily submit a copy of their adult IV standard concentrations and dosing guidelines. From these lists, the data were summarized and compiled to evaluate trends and compare the various policies.

Results

A total of 18 different hospitals and health systems responded. Survey results showed more than 255 concentrations for 84 unique drugs. Of these, 37 were high-risk medications, with 135 unique drug concentrations. From this list, a single proposed concentration was developed for each medication. If utilized, this standardization would result in a greater than 65% reduction in potential drug concentrations in use. A single dosing unit was also proposed for all medications reviewed, resulting in a greater than 21% reduction in variation.

Conclusions

Standardization of IV drug concentrations and dosing units statewide could reduce the variability in IV therapies and promote safer and more consistent practices in medication administration.

The medication administration process is a complex system that requires several coordinated steps to be performed for appropriate medication administration. Because of the complexity, errors can occur at any step in this process. However, no matter how hard one strives for perfection in clinical practice, some level of variability exists. Lean methodology has been adopted in many health care organizations and teaches increasing value by improving quality and decreasing costs. One of the core principles in achieving these goals is to standardize workflow by reducing variation as a means to reduce waste. Variability can lead to errors and potentially increase health care costs. The challenge lies in being able to identify and take necessary actions in order to implement change to reduce and monitor the variability that exists within the system.

One of the foundational reports that first drew national attention to the incidence of medication errors was released by the Institute of Medicine (IOM) in 1999, titled To Err is Human: Building a Safer Health System. This report discusses how 44,000 to 98,000 Americans die each year due to medical errors; more specifically, 1 out of 854 deaths in hospitalized patients is due to a medication error. ¹ Further studies have shown how medication errors are the most common cause of adverse drug events, and 28% of medication-related adverse drug events are considered to be preventable.^1–3

Medication errors that are associated with high-risk drugs possess some of the greatest potential for causing significant harm to patients.^4,5 High-alert medications, such as those defined by the Institute for Safe Medication Practices (ISMP), ⁶ are primarily delivered through intravenous (IV) infusions (eg, heparin, insulin, propofol) within the inpatient setting and cause an increased risk of harm to patients when used in error.^4,7–9 In consequence, 61% of the most serious and life-threatening potential adverse drug events are related to IV drugs. ¹⁰ IV medications have accounted for up to 56% of medication errors and 54% of potential adverse drug events (ADEs).^11,12 Hospitalized patients may receive numerous IV medications, with the critically ill typically receiving several at a time, increasing the probability of an error.

Drug administration has been shown to be the stage of the medication use process (MUP) most vulnerable to error. ³ Compared to other stages, this step has fewer safeguards and support structures, mainly due to the fact it is performed by a single health care professional.^7,13,14 Furthermore, 38% of preventable medication errors that result in ADEs have been shown to occur during the administration stage, with only 2% being intercepted before patient interaction. ¹⁵ Studies performed within intensive care units (ICUs) observing high-risk IV medications showed that 34% to 49% of errors occurred during administration.^14,16 This level of error may be due to administration occurring at the end of the MUP, where there are no safety redundancies or further opportunities to intercept errors. Additionally, 41% of drug administration mortalities were shown to be associated with a dosing error. ¹⁷ A key issue related to administration is incorrect dosing of medications. With nursing staff turnover rates and numerous hand-offs, high variability can lead to an error. This effect can be compounded when nurses work in several settings or transfer/work within several institutions at one time.

Studies support the standardization of the medication administration process to help reduce errors. For instance, Apkon et al ¹⁸ showed how a designed safer approach to IV drug administration reduced the number of IV medication and infusion errors. Another study ¹⁹ found support in standardizing parenteral medication practices in the pediatric setting, which showed increases in operational efficiency as well as patient safety goals.

It is important to realize that most hospitals have already taken measures and, to some degree, have begun standardization efforts of IV drug libraries. A review of infusion safety software from 100 hospitals demonstrated the large variability in several factors: drug names, concentrations, dosing units, and dosing limits. ²⁰ It was noticed that drug concentrations were in fact the least variable of the infusion parameters, with dosing units demonstrating a greater variability. ¹⁸ Further, dosing units tended to differ even within a single institution depending on the specific patient care unit. Some level of standardization of IV concentrations has occurred, but more can be accomplished across institutions and nursing units.

Variability in IV medication practices drives the need for standardized practice. There are several thousand medications currently available, and this increases annually. The environment in which we deliver patient care has increased in complexity. With staff working in multiple settings, the risk for harmful medication errors is compounded.

Most large health care institutions and systems have either begun or have implemented some level of standardization of their IV libraries, but few have taken this standardization beyond their networks or geographical regions. This can create variability and errors when patients are transferred between institutions. To date, there is one published example of standardization — the metropolitan area from the San Diego Patient Safety Consortium Task Force. Utilizing a multidisciplinary task force to develop a single concentration for each medication, 17 hospitals within San Diego county demonstrated a 94% decrease in variation from the proposed standards. ¹⁰ Further, a 77% reduction in variation was observed for high-risk medications using the same criteria. A single dosage standard was also developed for all reviewed medications, resulting in a 100% reduction in variation. This group documented the implementation of these standards, demonstrating how 15 of 17 hospitals chose to adopt the standards (accepting between 72% and 100% of the standard list). This is also currently the only study of its kind that provides outcomes data demonstrating the benefits of multi-institutional standardization practices from institutions not under common ownership. The creation of a policy and procedure to standardize work processes, such as those suggested by these types of coalitions, can create a margin of safety when high-risk products as well as general-use medications are used.

The need to reduce variability and begin standardization has become a focus of several prominent organizations. IOM ¹ and ISMP ²¹ have been releasing reports since 1999 that describe the need to reduce variability. Since 2003, The Joint Commission National Patient Safety Goals program has issued statements regarding the improved safety of using medications. ²² One method is to “standardize and limit the number of drug concentrations available in the organization.” The US Pharmacopeia Safe Medication Use Expert Committee published the results of a national survey to determine trends for suggesting standardization of IV infusion concentrations. ²³ In 2013, the American Society of Health-System Pharmacists updated an endorsed policy from 2008 (Policy 1306) recommending the development of nationally standardized drug concentrations and dosing units for high-risk medication infusions. ²⁴ Each of these groups has provided information and general principles to guide development of standard concentrations for IV admixtures.

This project seeks to evaluate the current IV standardization practices for adult patients across hospitals in North Carolina and to compare them with each other. In addition, the project proposed hypothetical standards to determine how much they could impact the observed variability and to better understand the need for standardization. Adult continuous infusion concentration and dosing unit standards were the focus of this study, as they affect the largest proportion of our patient population.

Materials and Methods

In the fall of 2013, an e-mail request was sent to select hospital pharmacy leaders within 20 institutions asking them to voluntarily submit a copy of their adult IV standard concentrations and dosing unit guidelines. These leaders were selected based on their potential diverse representation, ensuring that hospitals within the same health systems that share policies were not selected. The goal was to prevent skewing of observed trends and to gain the best representation of policies used across the state. Although only a small population of the 124 licensed hospitals within the state of North Carolina were solicited (excluding 14 hospitals operated by the military), the queried institutions potentially represented 61% of institutional policies due to common ownership. Further, a broad representation of geography in the respondents was chosen, including small community hospitals, large academic medical centers, and integrated health systems.

Of the queried institutions, 18 different hospitals and health systems responded (90% response rate) with their adult IV standard concentrations and dosing recommendations. The data were summarized and compiled to evaluate trends and compare the various policies. Furthermore, medications that were deemed high risk per the ISMP classification were identified for separate analysis. ⁶

Each policy was reviewed to look for the variation in practices. In general, a need for standardization was confirmed if 4 or more institutions listed a drug in their adult IV infusion policies or if it was classified as a high-risk medication. Drugs were excluded from the final analysis if 3 or fewer institutions listed them within their adult IV infusion policies.

To determine the impact of a new set of potential standard concentrations and dosing units, we proposed a single concentration and dosing unit standard for each medication. It was determined that, with few exceptions based on clinical judgment, this would be the best method to reduce variability. The primary author utilized multiple resources to develop these proposed recommendations, including manufacturer recommendations, secondary and tertiary reference materials, package inserts, typical dosing strategies, drug packaging and IV diluent volumes, ease of mathematical calculations and dilutions, and clinical judgment.

Results

From the submitted lists, the data were summarized and compiled to evaluate trends and comparisons for adult IV infusions. Of note, 5 of the institutions reported not having standardized infusion concentrations. Furthermore, of the 13 institutions that submitted policies, only 8 (61.5%) defined dosing units for all included medications within these policies. In some cases, concentrations and dosing units in a hospital could not be determined because of a lack of formal standardization.

Based on these lists, 84 unique IV medications were identified, consisting of 255 unique drug concentrations and 102 unique dosing units ( Table 1 ). After excluding drugs that were not represented on at least 4 hospital lists, 61 common medications were placed into analysis, consisting of 225 unique concentrations and 78 unique dosing units. A full summary demonstrating the variety of standard concentrations and dosing units for analyzed medications can be seen in Table 2 (high-risk medications are designated in shaded rows). Overall, the average number of drug concentrations and dosing units per drug was 3.7 and 1.3, respectively.

Table 1.

Summary of collected adult standardized infusion policies within North Carolina health care institutions

Standardized infusion policies	n
Unique drugs identified	84
Unique drug concentrations	255 (avg 3.0 per drug)
Unique dosing units	102 (avg 1.2 per drug)
Unique common drugs analyzed	61
Unique drug concentrations	225 (avg 3.7 per drug)
Unique dosing units	78 (avg 1.3 per drug)
High-risk medications	37
Unique drug concentrations	135 (avg 3.6 per drug)
Unique dosing units	47 (avg 1.3 per drug)

Note: avg = average.

Table 2.

Summary of variety in standard concentrations and dosing units within North Carolina health care institution policies for adult intravenous infusions (n = 12)

Medication	No. of hospitals reporting	No. of unique conc.	No. of hospitals with most common conc.	Most common conc.	No. of hospitals reporting	No. of unique dosing units	No. of hospitals with most common dosing units	Most common dosing unit
Abciximab	9	2	8	29 & 36 mcg/mL	8	2	7	mcg/kg/min
Acetylcysteine	5	3	2	12 mg/mL	5	2	4	mg/kg/h
Alteplase	8	7	7	1 mg/mL	6	1	6	mg/h
Aminocaproic acid	7	5	6	20 mg/mL	6	2	5	g/h
Aminophylline	7	3	5	2 mg/mL	6	1	6	mg/kg/h
Amiodarone	11	5	9	1.8 mg/mL	7	2	4	mg/min
Argatroban	10	1	10	1 mg/mL	6	1	6	mcg/kg/min
Atracurium	4	3	2	1 mg/mL	3	1	3	mcg/kg/min
Bivalirudin	7	3	7	5 mg/mL	5	1	5	mcg/kg/h
Bumetanide	5	3	3	100 mcg/mL	5	1	5	mg/kg/h
Calcium gluconate	4	4	3	10 & 20 mg/mL	3	2	2	g/h
Cisatracurium	9	5	5	2 mg/mL	7	1	7	mcg/kg/min
Clevidipine	5	1	5	0.5 mg/mL	3	1	3	mg/h
Conivaptan	5	1	5	0.2 mg/mL	4	1	4	mg/h
Dexmedetomidine	10	1	10	4 mcg/mL	7	1	7	mcg/kg/h
Diltiazem	12	2	11	1 mg/mL	8	1	8	mg/h
Dobutamine	12	4	7	4 mg/mL	8	1	8	mcg/kg/min
Dopamine	12	6	9	3.2 mg/mL	8	1	8	mcg/kg/min
Epinephrine	12	8	7	16 mcg/mL	8	2	8	mcg/min & mcg/kg/min
Epoprostenol	2	2	2	10 mcg/mL	3	1	3	ng/kg/min
Eptifibatide	11	2	11	0.75 mg/mL	7	1	7	mcg/kg/min
Esmolol	12	3	7	10 mg/mL	8	1	8	mcg/kg/min
Fenoldopam	7	4	6	0.04 mg/mL	5	1	5	mcg/kg/min
Fentanyl	8	4	5	50 mcg/mL	6	1	6	mg/h
Furosemide	11	4	9	1 mg/mL	7	1	7	mg/h
Heparin	11	4	6	100 units/mL	7	2	6	units/h
Hydromorphone	7	5	6	0.2 & 1 mg/mL	4	1	4	mg/h
Insulin (regular)	11	3	11	1 unit/mL	7	1	7	units/h
Isoproterenol	11	5	8	4 mcg/mL	7	1	7	mcg/min
Ketamine	3	3	3	2 mg/mL	3	1	3	mg/kg/h
Labetalol	9	5	4	1 mg/mL	6	1	6	mg/min
Lepirudin	4	1	4	0.4 mcg/mL	3	2	2	mcg/kg/h
Lidocaine	12	2	8	4 mg/mL	8	1	8	mg/min
Lorazepam	9	4	5	1 mg/mL	5	1	5	mg/h
Magnesium sulfate	8	4	8	40 mg/mL	5	1	5	g/h
Methadone	1	1	1	2 mg/mL	1	1	1	mg/h
Midazolam	11	5	10	1 mg/mL	7	2	6	mg/h
Milrinone	12	3	12	0.2 mg/mL	8	1	8	mcg/kg/min
Morphine	9	4	8	1 mg/mL	5	1	5	mg/h
Naloxone	9	11	3	4 mcg/mL	7	3	5	mg/h
Nesiritide	10	1	10	6 mcg/mL	6	1	6	mcg/kg/min
Nicardipine	12	4	8	0.1 mg/mL	8	1	8	mg/h
Nitroglycerin	12	4	9	0.4 mg/mL	8	1	8	mcg/min
Nitroprusside	12	3	9	0.4 mg/mL	8	2	7	mcg/kg/min
Norepinephrine	12	6	8	16 mcg/mL	8	1	8	mcg/min
Octreotide	9	6	4	5 mcg/mL	6	1	6	mcg/h
Oxytocin	4	2	4	0.02 & 0.06 units/mL	3	2	2	units/h
Pancuronium	4	3	3	1 mg/mL	3	1	3	mcg/kg/min
Pantoprazole	8	4	5	0.32 mg/mL	6	1	6	mg/h
Pentobarbital	5	5	5	2, 5, 8, 8.33, & 50 mg/mL	5	1	5	mg/kg/h
Phenylephrine	11	10	8	80 & 160 mcg/mL	8	2	5	mcg/min
Potassium chloride	3	2	3	0.1 mEq/mL	1	1	1	mEq/mL
Procainamide	11	4	11	4 & 8 mg/mL	7	1	7	mg/min
Propofol	10	1	10	10 mg/mL	7	1	7	mcg/kg/min
Reteplase	3	3	3	0.02 & 0.05 & 1 units/mL	3	1	3	units/h
Rocuronium	5	3	3	1 mg/mL	3	1	3	mcg/kg/min
Sodium bicarbonate	4	3	2	0.13 mEq/mL	4	3	2	mEq/kg/h
Tenecteplase	1	1	1	10 mcg/mL	1	1	1	mg/h
Treprostinil	2	4	2	10 mg/mL	3	1	3	ng/kg/min
Vasopressin	12	6	5	1 unit/mL	8	2	7	units/min
Vecuronium	11	4	8	1 mg/mL	7	2	6	mcg/kg/min

Note: High-risk medications are designated in shaded rows. conc. = concentration.

No identified medications had identical standardized concentrations within participating hospitals. However, there was a majority consensus on 4 medications (argatroban, dexmedetomidine, nesiritide, and propofol) due to a single standard concentration being used by all institutions that reported this drug. Only 12 (20%) of the final analyzed medications (diltiazem, dobutamine, dopamine, epinephrine, esmolol, lidocaine, milrinone, nicardipine, nitroglycerin, nitroprusside, norepinephrine, and vasopressin) were reported to have at least one standardized concentration in every institution within North Carolina. Phenylephrine, naloxone, and epinephrine showed the most variety between hospitals, with 10, 11, and 8 separate unique standard concentrations in use, respectively. Based on ISMP classification for high-risk medications, almost 37 (60%) of the listed medications in the final analysis were deemed high risk, making up 135 (60%) of the unique drug concentrations ( Table 1 ).

Variations in dosing units (such as mg/mL or mcg/kg/min) were not as prominent between institutions, which typically resulted in a clear majority between 2 unique dosing units. However, most institutions did not report a dosing unit along with their standard concentrations ( Table 2 ). In fact, at least 4 institutions did not report a dosing unit for each medication analyzed, and more than 24 (39%) medications had 7 or more institutions fail to define a dosing unit. There was a majority consensus on 9 medications (diltiazem, dobutamine, dopamine, esmolol, lidocaine, milrinone, nicardipine, nitroglycerin, and norepinephrine) due to a single dosing unit being used by all institutions that listed these drugs ( Table 2 ). Sodium bicarbonate showed the most variety between hospitals, with 3 separate unique dosing units in use. Based on the ISMP classification for high-risk medications, 47 (60%) of unique dosing units were contributed by the 37 identified high-risk medications ( Table 1 ).

The target goal was to understand the impact of a single, standard concentration and dosing unit for each medication. If such a list existed and was adhered to, this could yield a greater than 72% reduction in variation from concentration standards and a greater than 21% reduction for dosing units. In a hypothetical final compendium of standards (eTables 1 and 2), the investigators were able to achieve a single dosing unit for each unique medication, resulting in a 21.8% reduction in variation in all medications and 21.3% in high-risk medications ( Table 3 ). This resulted in a 66.7% reduction in total variation and a 65.9% reduction in variation in high-risk medications ( Table 3 ).

Table 3.

Reduction in variation of intravenous drug concentrations with standardized concentrations and dosing units

No. of concentration variants				Percent reduction	No. of dosing unit variants			Percent reduction
	Initial	Target a	Final b		Initial	Target a	Final
High-risk medications	135	37	46	65.9%	47	37	37	21.3%
General-use medications	90	24	28	68.9%	31	24	24	22.6%
All medications	225	61	75	66.7%	78	61	61	21.8%

^aTarget goals are defined as an ideal goal of one concentration and dosing unit per unique medication.

^bSome medications were allowed 2 concentrations based on potential administration volumes.

Discussion

Medication administration, despite deliberate changes made within the MUP over the years, remains a series of steps in a complex process that is prone to errors. Medication orders themselves are a complex system, requiring the drug, frequency, dose, rate, and duration to define a patient's individual needs and reach a desirable, healthy outcome. A lack of precision in any element of the order may lead to an error; errors can also occur due to misinterpretation or lack of knowledge. This risk is compounded in an ICU, where patients are more likely to receive high-risk medication intravenously.

This study revealed the wide spectrum of practice within North Carolina in IV medication preparation and administration. An example of this (seen in Table 2 ) is epinephrine, a medication used as a continuous infusion, reserved for critically ill patients and not available in a premade formulation. All hospitals reported using a standardized concentration for this medication, but there were 8 unique concentrations utilized. The most frequent concentration was 16 mcg/mL, but the following concentrations were provided as options at different hospitals: 8, 10, 16, 32, 48, 64, 120, and 128 mcg/mL. To add further variability, 3 hospitals reported the preferred dosing unit as mcg/min, 4 preferred mcg/kg/min, and 4 did not list a preferred infusion parameter. Some medications had variations in both concentrations and dosing units. For example, furosemide had 4 concentrations and 3 dosing units in use. This can result in a total of 12 concentration–dosing unit possibilities. Another medication to highlight is vecuronium, a neuromuscular blocking agent used to paralyze critically ill patients. Seven hospitals preferred a concentration of 1 mg/min, but others reported concentrations of 0.2, 0.4, and 0.5 mg/mL.

The most logical reason for this variability is that intrahospital standardization, rather than a national standard, has been used to create the various available concentration. If a national guideline were in existence, many hospitals would use it as a starting point in creating their standardization practices.

One major benefit of drug infusion standardization is to optimize drug libraries. These libraries not only need standardized concentrations, but they also need standardized dosing units. The 5 institutions that reported not having current policies were smaller (<300 beds) institutions and/or were located closer to rural areas. For hospitals that do not have standardized infusion concentrations, guidance to the pharmacy for preparation and nursing upon administration will improve safety of care. It will also provide the backbone that can be used to develop the information placed into smart pumps. With smart pump library standardization, the process of collecting data from multiple hospitals to track and trend utilization data on medications across a wider geographic area can begin. With the current variability in practice, this is not possible.

The largest limitation of this study is that it might not be possible to standardize all drugs to one concentration. With titratable medications and patients on fluid restriction, a higher concentration medication might be needed. However, these hypothetical recommendations are not to be used for implementation purposes but to illustrate the variability in the current standardized practices and the potential reduction in variability that could occur with standardization. In addition to the safety benefit that could come from statewide or national standardization, other benefits could include conservation of medications on shortages or the generation of a large enough market share for a certain concentration that could lead to a manufacturer developing a ready-to-administer formulation. For the drug library, having standardized concentrations and dosing units could lead to fewer updates of the smart pumps, which takes staff time and can lead to the introduction of errors.

Conclusion

Many institutions in North Carolina, while maintaining a standard infusion protocol, have a variety of standard IV infusion concentrations. Furthermore, some institutions do not have standard infusion concentrations and dosing units. The use of a single standardized list for continuous infusions would allow health care providers to deliver safer and more consistent patient care in a complex medical environment.

This information and proposed list of standards will serve as an initial starting point for the development of a statewide consortium of pharmacists. This group will review and discuss individual medication standardization to determine a preliminary standardization policy to improve the safety of continuous IV medication administration in adults within North Carolina institutions.