Parenteral Nutrition Safe Practices: Results of the 2003 American Society for Parenteral and Enteral Nutrition Survey*  
From the * Beth Israel Medical Center and Albert
Einstein College of Medicine, New York, New York; Correspondence: Gordon S. Sacks, PharmD, BCNSP, 777 Highland Avenue, University of Wisconsin–Madison, Madison, WI 53705. Electronic mail may be sent to gssacks{at}pharmacy.wisc.edu.
Background: The American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) recently published a revision of its "Safe Practices for Parenteral Nutrition" guidelines. Because there is a paucity of published scientific evidence to support good practices related to ordering, compounding, and administering parenteral nutrition (PN), a survey was performed in the process of the revision to gain insight into the discrepancies between reported practices and previous guidelines. Methods: A web-based survey consisting of 45 questions was conducted (n = 651) June 1–30, 2003. Respondents were queried about primary practice setting, professional background, processes for writing PN orders, computer order entry of PN orders, problems with PN orders, and adverse events related to PN. Results: There were 651 survey responses, 90% of which were from hospital-based practitioners. Almost 75% of responders processed between 0 and 20 PN orders per day. Overall, physicians (78%) were responsible for writing PN orders, but dietitians and pharmacists had significant involvement. PN base components were most often ordered as percentage final concentration after admixture (eg, 20% dextrose), which is inconsistent with safe practice guidelines of ordering by total amount per day (eg, 200 g/day). There was no consistent method for ordering PN electrolytes. Approximately 45% of responders reported adverse events directly related to PN that required intervention. Of these events, 25% caused temporary or permanent harm, and 4.8% resulted in a near-death event or death. Conclusions: Although the survey found consistency in PN practices for many areas queried, significant variation exists in the manner by which PN is ordered and labeled. As part of its mission to serve as the preeminent, interdisciplinary nutrition society dedicated to patientcentered, clinical practice worldwide through advocacy, education, and research in specialized nutrition support, the American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) published "Safe Practices for Parenteral Nutrition."1 This document provides guidelines with supporting evidence to foster quality parenteral nutrition (PN) therapy. The intent is for the principles provided in the document to become incorporated into healthcare organization practice for the purpose of minimizing the risks of PN complications. "Safe Practices for Parenteral Nutrition" was extensively revised, adding a new section on the PN ordering process and significantly expanding the section devoted to PN administration.2 Unfortunately, there is little, if any, published evidence to support good practices in these areas. Although data from randomized clinical trials of nutrition support are ideal for developing clinical practice guidelines, this type of information is not widely available. Several factors inherently limit the use of prospective randomized clinical trials in the evaluation of nutrition support.3 Those most likely to benefit from the treatment (eg, severely malnourished patients) cannot be randomized to an unfed control group due to ethical dilemmas. Other limitations include outcome results influenced by clinical variables independent of nutrition support and inability to recruit large numbers of eligible individuals from 1 medical center, contributing to the enrollment of marginal candidates for nutrition support. Because clinical guidelines cannot be based solely on prospective randomized trials, the A.S.P.E.N. Task Force for Revision of Safe Practices for Parenteral Nutrition conducted the 2003 Survey of PN Practices,4 focusing on policies and procedures relating to ordering, compounding, and administering PN and quality oversight of this process. The final document published as "Safe Practices for Parenteral Nutrition" was based primarily on the recommendations of experts in the field and was evidence-based for as much as the literature provided evidence to support these recommendations. The results from this survey were used as a basis for the revised "Safe Practices for Parenteral Nutrition" to enhance the quality and efficacy of nutrition support. Following is a summary and analysis of the responses to the survey.
Development of the Survey Instrument One of the survey objectives was to identify common practices related to ordering and compounding, administration of PN, and quality oversight of this process. A questionnaire based upon the existing A.S.P.E.N. "Safe Practices for Parenteral Nutrition" was developed to obtain an overview of the variance and consistency with current practices from a variety of healthcare settings. It was designed to include both hospital- and non–hospital-based PN practices. The survey instrument was not tested or validated before its distribution, but it was reviewed by a multidisciplinary panel of nutrition support practitioners and revised before becoming available for participant responses. The survey was administered electronically through the A.S.P.E.N. website and announced to the membership via society journals and A.S.P.E.N. list servers. Announcements inviting participation were also sent to selected professional groups, including the American College of Clinical Pharmacy, the American Society of Health-System Pharmacists, the National Home Infusion Association, and others. Participation in the survey was completely voluntary. The survey instrument consisted of 45 questions with multiple-choice and free-text responses. It was organized into 5 sections: demographics of the respondent, writing PN orders, computer order entry of PN orders, problems with PN orders, and adverse events related to PN. Questions in the demographic section focused on information such as professional background (ie, MD, RN, RD, RPh, other) and primary practice setting (ie, hospital, homecare, etc). The orderwriting section was designed to identify the discipline responsible for writing PN orders, whether or not standard PN order forms were used, and the manner in which PN components were ordered (ie, dextrose in percent final concentration vs g/day, electrolytes in mEq/L vs per day, etc). The computer order entry section was designed to quantitate the use of computerized order entry systems and automated compounding devices. The final 2 sections, problems and adverse events related to PN, were developed to capture the type and frequency of harm associated with the compounding and administration of PN formulations.
Data Collection
Data Analysis
Demographics The demographic characteristics of the respondents (n = 651) are presented in Table I. The majority of respondents were registered dietitians (55%), followed by registered pharmacists (32%). Hospitals (teaching, nonteaching, and government) were the most common primary practice setting (90%). The remaining 10% of respondents worked in the home infusion setting. Twenty-six percent of respondents indicated the average daily census for their organization was 400+ patients, whereas 22% serviced only 100–199 patients per day. Only 15% of respondents were preparing >20 adult PN formulations or >5 pediatric PN formulations per day. A once-daily nutrient infusion system was preferred by the majority of organizations (76%).
Approximately 54% (337 of 620) of those responding to the question indicated that their organization used a performance improvement (PI) process for various components of the PN system. Specific components identified as part of the PI process are listed in Table II. The appropriate use of PN and accuracy of PN orders were the most frequently used markers for PI; costs of PN compounding and PN wastage rates were monitored by <30% of respondents that used a PI process. Indicators reported in free-text answers but not included as choices in the survey instrument included (1) appropriate dose of nutrient, (2) appropriate monitoring of PN, (3) whether recommendations were followed, and (4) efficacy of PN.
Writing Parenteral Nutrition Orders Approximately 48% of organizations ordered the PN base components in percentage of final concentration after admixture (eg, 20% dextrose) and 33% as the volume of percentage of original concentration (eg, 200 mL of 50% dextrose). Several respondents reported using a combination of methods to order the individual macronutrients, as listed in Table III. Of note, there was no standardization between home infusion agencies and hospitals.
No consistent method was identified for ordering of electrolyte components or quantities. Use of the electrolyte salts (eg, sodium chloride, potassium chloride) was favored over the use of the individual electrolyte cations or anions (eg, sodium, potassium, chloride, acetate). However, the quantity of electrolyte ordered was expressed in a variety of ways. The number of mEq/L (or per 100 mL for neonates), mEq per day, and mEq per total volume were the 3 most common methods. Table IV includes a list of all the methods used. Phosphorus was ordered in mmol of the salt (eg, potassium phosphate or sodium phosphate) by 57% of respondents and in mEq of the cation by 26% of respondents. Approximately 17% used both methods of nomenclature. Some amino acid products are manufactured with electrolytes as buffers (ie, chloride, acetate) or contain electrolytes as components (ie, sodium, phosphorus). When queried as to whether these electrolytes were considered in the total electrolyte delivery, 71% did not include these in the total amount during the writing of PN orders. The pharmacy was allowed to adjust certain electrolyte additives such as acetate or chloride in 62% of respondent's organizations.
The overwhelming majority of organizations (90%) routinely ordered a standard volume of a commercially available multiple trace element product vs each trace element ordered separately. Approximately 1 of 5 respondents (22%) adjusted the dose of trace elements according to patient clinical condition (ie, gastrointestinal losses, liver failure). A total of 88% of respondents reported using a standardized PN order form. Improved interpretation of PN orders and reduced errors were perceived by 83% to be a result of using standardized PN order forms. Educational information was frequently included either on or with PN order forms, as those writing PN orders were felt to lack knowledge of PN preparation. Respondents from the home infusion setting reported that the primary reason for not using a standardized PN order form (32%) was the inability to standardize the ordering process because orders were received from multiple organizations. Only 54% of respondents reported using standard order forms for laboratory and patient care orders related to PN. There was a significant difference found between the use of standard PN order forms and how often PN orders required clarification. PN orders written with the use of standard order forms seldom required clarification, whereas PN orders required frequent clarification if standard order forms were not used (p = .015). There was a trend (p = .085) between the use of standard order forms and the reasons that PN orders required clarification (eg, incorrect or unstable macronutrient, inappropriate dosage according to disease state). Duration of administration of IV fat emulsion (hang time), whether administered separately from PN or as a component of a total nutrient admixture (TNA), was not consistent. Most respondents allowed fat emulsion to hang for a minimum of 12 hours and a maximum of 24 hours when infused separately from PN. The patient population appeared to influence the practice for maximum hanging time, with a 12-hour maximum hang time for adults but extending to a 24-hour hang time for pediatric/neonatal patients. Responses for minimum and maximum hang time of fat emulsion as part of a TNA were variable.
Computerized Prescriber Order Entry (CPOE) and Outsourcing The majority (85%) of organizations compounded PN formulations in their own facility and did not outsource this responsibility. When organizations did outsource PN compounding, 95% responded that a pharmacist reviewed the order before delivery to the compounding facility.
Problems With PN Orders
Adverse Events Related to PN Formulations
The final question asked institution-based respondents about their policies for allowing a patient to bring in or continue a PN formulation compounded at an outside facility. Approximately 43% of respondents reported that this practice was allowed. The most common reasons for barring this practice were safety and liability issues. A summary of responses is shown in Table VI.
The purpose of this survey was to identify current policies, procedures, and quality oversight being used by healthcare organizations for ordering and compounding PN. In addition, this information would serve as a foundation for revising previously published safe practice guidelines for PN. It was interesting to note that the majority of respondents' institutions prepare <20 PN orders daily. No previously published data have been analyzed to determine the correlation between overall institution size and the number of PNs written. As research increasingly supports the use of enteral nutrition over PN for patients with a functional gastrointestinal tract, one might expect a lack of correlation between the size of an institution and the use of PN if, for instance, larger academic institutions change behavior more quickly in response to published clinical research findings. Providers in the home infusion setting made up only 10% of the respondents. The low response rate from this group of practitioners serves to underscore the difficulty in reaching such a widely dispersed group and the need to increase recruitment of home infusion providers into societies such as A.S.P.E.N. The low rate of response from this group makes comparisons with practices in hospitals and home infusion difficult. PI programs were used by over half of respondents, but less than one-third included quality control of PN compounding as a component of this process. Errors occurring in the compounding process of sterile preparations have come under scrutiny due to adverse events associated with microbial contamination. Over the past 20 years, there have been reports of patient morbidity from improperly prepared PN formulations.5–7 Common factors in these cases involved lack of policy and procedures and lack of process validation for sterile compounding. According to a national survey of quality-assurance activities for pharmacy-compounded sterile preparations conducted in 2002, only 39% of respondents (n = 182) reported having such a program.8 The need for consistent evaluation of aseptic technique and in-process or end-product testing of compounded PN formulations has been recognized by the pharmacy profession with the publication of the United States Pharmacopeia (USP) Chapter 797. This chapter of the USP and the National Formulary outlines the practice standards for compounding sterile preparations and is enforceable by regulatory agencies.9 Thus, pharmacies compounding PN may be inspected for compliance with these standards by state boards of pharmacy, the US Food and Drug Administration (FDA), and accreditation organizations such as the Joint Commission on Accreditation of Healthcare Organizations (JCAHO). Failure of pharmacies to develop and implement these practice standards may result in the FDA ultimately restricting the compounding of sterile products to a select group of specially trained and certified professionals. The responses to questions regarding responsibility for PN order writing yielded several notable findings. Of particular interest was that nutrition support services were involved in PN order writing 31% of the time, despite the general perception that nutrition support services are a thing of the past. Of further interest is the significant role that physician assistants and advance practice nurse practitioners play in the orderwriting process. Although there are certifying examinations and organizations devoted to nurses in nutrition support, there are none for physician assistants. It was not surprising that physicians were most often identified as being responsible for writing PN orders. It was clear, however, that the training level of the physicians varied greatly, and it seemed that resident physicians were most often the ones responsible. Pharmacy and, to a lesser extent, dietitians were also involved in order writing, approximately 40% of the time each. The oversight of the order-writing process was most often the responsibility of the pharmacy (71%), a nutrition support service or committee (51%), or a medical staff committee (51%). It is likely that there was significant overlap of responses to the latter 2 options, given that both contained a committee. Several problems with the ordering process for PN were identified. It is of particular concern to the task force that so much variability exists in the manner in which PN components are ordered. This is seen as a potential source of error when orders are communicated from inpatient to outpatient settings, and vice versa, and may be the largest area of change in practice resulting from the revised Safe Practices Guidelines.2 There is a significant lack of concordance between the ordering methodology for base components, electrolytes, and other additive components and the recommendations contained in the revised Safe Practices Guidelines. The task force recommended that all PN components be ordered in amounts per day to avoid any misinterpretation. This means that the most commonly reported methods for ordering base components (ie, dextrose, fat, amino acids), such as percentage of final concentration (eg, 25% dextrose) and volume of percentage of original concentration (eg, 250 mL of 70% dextrose) should no longer be used. Although it may be understood among physicians who are writing PN orders that percentage of final concentration of dextrose is just another way to describe g of dextrose per day, this can often be misinterpreted by the pharmacists, dietitians, or nurses involved in the PN process. This disparity was just as prominent in the ordering of electrolytes. Although most respondents reported ordering in the electrolyte salt (eg, NaCl or Na acetate), only 39% of respondents indicated ordering in mEq (or mmol) per day. There was a great deal of variability in policy for infusion of fat emulsion. When fat emulsion is administered as a separate infusion from the dextrose and amino acids, minimum hanging time allowed most often varied from 4 to 12 hours, with 12 hours being most frequent. There were 16% of respondents lacking any minimum hanging time policy in this group, and approximately 6% of respondents indicated a minimum hanging time of 24 hours. Some respondents indicated that fat emulsion infusion rates were limited by a maximum rate instead of a minimum infusion time. Maximum hanging time for separately infused fat emulsion was most often reported as either 12 or 24 hours. When fat emulsion was included in a TNA, minimum hanging times were reflective of minimum allowable cycle times and were most often 8, 12, or 24 hours. Maximum allowed infusion time reported was 24 hours in nearly all respondents, suggesting a nearuniversal acceptance of 24 hours as a maximal infusion time for TNA. This was probably the only practice that achieved this level of concordance in the entire survey. Most respondents indicated that a standard order form was used when PN was ordered. Home infusion respondents most frequently indicated a lack of standardization, citing different referral sources as the reason standard PN order forms were not used. These data should be assessed with caution as respondents indicated a degree of confusion between the use of standardized PN order forms and the use of standard predesigned PN formulations. The significant difference between the use of standard PN order forms and how often PN orders required clarification was one of the many reasons that our task force advocated the use of standard PN order forms in the revised Safe Practice guidelines. One of the most common reasons necessitating a pharmacist contacting the prescriber was illegible writing. Approximately 25% of respondents answering this question (n = 496) identified that PN orders needed to be clarified "very often" or "often" as a result of illegible writing. This is dramatically higher than the rate recently published by USP using their MEDMARX system as the data collection source for medication errors.10 This system is an Internet-accessible program that has been used by >775 subscribing hospitals and health systems to report medication errors. Between the years 1999 and 2003, analysis of MEDMARX data found that 3% of all medication errors reported were associated with unclear or illegible handwriting. The higher frequency of illegible orders captured by our survey may be due to the complex nature of PN orders. Orders for PN are often messy and hard to read due to corrections of math errors or improper selection of components. CPOE may serve as one strategy to reduce errors and improve PN safety. One study reported a substantial reduction in the frequency of PN orders requiring clarification once CPOE was instituted in a neonatal intensive care unit, a pediatric intensive care unit, and a general pediatrics unit.11 Similar to our survey, 25% of PN orders written on the paper-based form required clarification. Only 1.6% of the orders generated with CPOE needed clarification. The most concerning statistics revealed by the survey were the frequency of errors and the severity of harm attributed to these errors. Almost two-thirds of respondents reported observing between 1 and 5 errors per month related to PN. Furthermore, about twothirds of respondents estimated that their organizations prepared 0–10 PN formulations per day. If this were truly representative data, extrapolating this information out to 1 month, one could propose an error rate ranging from 0.3% to 1.6% (assuming a range of 1–5 errors per 10 PNs per day for 30 days). To our knowledge, there are no published data on error rates associated with PN considered as "one" medication. As a comparison, MEDMARX reported a medication error rate of 3.6% for insulin and 2.1% for potassium chloride, both of which are high-alert medications, as well as additives that can be found in PN formulations. Interestingly, electrolytes, such as potassium chloride, were identified in our survey as the PN component most often associated with errors. Insulin and dextrose were the second most common ingredients of PN formulations linked to errors. Because there are up to 8 separate electrolyte salts that can be added to PN formulations, it follows that most errors are associated with these components. However, it is difficult to understand why more errors occur with insulin vs amino acids, vitamins, trace elements, or histamine2-receptor antagonists. Unfortunately, errors with insulin and potassium chloride are associated with the most harm (error categories E–I, Table VII).
Interestingly, there was a difference in adverse event reporting, including severity, according to the discipline. Adverse events requiring an increase in patient monitoring were reported most frequently by nurses, followed by physicians, pharmacists, and dietitians (p = .001). This observation appears intuitive at first because nurses are at the patient's bedside during PN administration. However, this finding may be related to a small number of nurse respondents (n = 25) compared with pharmacists (n = 186) and dietitians (n = 232). Dietitians were found to be unaware, more so than other disciplines, of any adverse event occurring as a result of PN administration. A new section was added to the revised Safe Practice guidelines emphasizing the need for monitoring the patient's response to PN therapy to ensure safe administration. The task force was also interested in identifying how often institutions allowed the administration of PN formulations compounded at outside facilities once a patient was admitted to an inpatient setting. Patients often request that their PN formulations compounded for home administration be infused in the hospital because a charge has already been incurred. The majority (57%) of institutions did not allow this practice. Despite significant limitations, the results of this survey point to a number of relevant issues related to delivery and quality assurance practices in the provision of PN. Because the number of practitioners involved in the provision of PN nationally is unknown, it cannot be assessed whether the respondents of the survey represent an adequate sample of the provider population. The proportion of respondents appeared not to be representative of the A.S.P.E.N. membership (eg, physician respondents 7% vs A.S.P.E.N. physician membership of 18% in 2003). Further, physicians and nurses composed very small fractions of the total response, possibly invalidating any statistical comparison of responses by discipline. The methodology used in developing this survey instrument may also limit the interpretation and applicability of the findings. The results reported here are derived from a survey that was designed to obtain a subjective overview of policies and procedures related to ordering and compounding PN, as well as the quality oversight of this process. The population surveyed was limited due to budgetary and time constraints and included persons who were on A.S.P.E.N. e-mailing lists or read A.S.P.E.N. journals in which the survey was advertised or who heard about the survey by word of mouth. Further, it is not known whether any given institution was represented by more than 1 respondent. Despite these limitations, the task force found the results useful in identifying practices pertinent to the revision of the Safe Practices guidelines. The task force believes that implementation of these guidelines by organizations will ultimately result in the safer provision of PN. The authors thank A.S.P.E.N. for their support in the collection and management of survey data.
Results presented in part at Nutrition Week Meeting of the American Society for Parenteral and Enteral Nutrition, Orlando, Florida, February, 2005. Received for publication August 25, 2005. Accepted for publication February 6, 2006.
Journal of Parenteral and Enteral Nutrition, Vol. 30, No. 3,
259-265 (2006) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
METHODOLOGY
RESULTS
2. The a priori
level of significance was set at 
0.05. 
