Integration of electronic health records with automated compounding for parenteral nutrition: A narrative review

Abstract

The integration of electronic health record systems has transformed healthcare by enhancing workflow efficiency, improving data accessibility, and enhancing patient safety. King Fahad Medical City transitioned from a fragmented, paper‐based parenteral nutrition management system to a fully integrated Epic electronic health record system, replacing the standalone Baxter ABACUS software. This review outlines a structured approach to integrating automated compounding devices with Epic, incorporating standardized data exchange protocols, interoperability frameworks, development of parenteral nutrition order sets, barcode‐enabled verification, and staff training. The clinical decision support system within Epic optimized parenteral nutrition prescription accuracy through real‐time alerts, dose limits, and automated calculations. A dedicated parenteral nutrition navigator tool further simplified access to patient data, supporting informed clinical decision‐making. This manuscript focuses on the implementation process rather than outcome evaluation. A qualitative approach, using process mapping, training assessments, and end‐user feedback, was employed to guide and monitor the integration progress. System refinement and multidisciplinary collaboration resolved challenges, including barcode recognition failures and coding mismatches. This experience underscores the importance of interoperability, decision support tools, and continuous quality improvement in optimizing parenteral nutrition therapy within electronic health record systems, offering a scalable model for healthcare institutions seeking digital transformation.

BACKGROUND

Over the past three decades, the healthcare industry has experienced a significant shift from paper‐based documentation to electronic health records, driven by advances in information technology. ¹ , ² The National Alliance for Health Information Technology defines electronic health records as digital records of health‐related information that conform to national interoperability standards and are accessible across healthcare organizations. ³ Electronic health records improve legibility, streamline data sharing, facilitate structured data analysis, and reduce storage and retrieval burdens. ⁴ , ⁵ These transformations have enhanced care delivery, patient safety, and healthcare accessibility. ⁶ , ⁷ , ⁸ , ⁹

Electronic health record systems incorporate several integrated tools to support clinical workflows and decision‐making. Among these tools, clinical decision support systems are embedded within the electronic health record and provide real‐time, patient‐specific recommendations to guide clinicians at the point of care. Clinical decision support systems function through alerts, reminders, clinical calculators, and evidence‐based order sets, all of which are accessed directly through the electronic health record interface. ⁷ , ¹⁰ , ¹¹ , ¹² This integration ensures a seamless data flow and timely support for clinicians, particularly in complex therapies such as parenteral nutrition, in which precision is crucial. ¹³

Parenteral nutrition is a high‐risk therapy that involves the daily prescription of individualized macronutrients and micronutrients, often compounded using automated compounding devices, such as the Medimix Multi from Germany. ¹³ , ¹⁴ Because of its complexity, parenteral nutrition is classified as a high‐alert medication by the Institute for Safe Medication Practices. ¹⁵ Errors are commonly reported during administration (35%) and transcription (39%) stages. ¹⁵ , ¹⁶ To mitigate such risks, the American Society for Parenteral and Enteral Nutrition (ASPEN) discourages handwritten orders and promotes the adoption of computerized provider order entry and clinical decision support systems tools. ¹⁷

Studies show that computerized provider order entry significantly reduces parenteral nutrition–related error rates from 15.6 to 2.7 errors per 1000 parenteral nutrition orders and helps patients achieve macronutrient goals more effectively. ¹⁵ , ¹⁸ , ¹⁹ It also standardizes care, reduces variability, and minimizes pharmacists’ interventions during verification. ¹⁸ , ²⁰ Clinical decision support systems with automated alerts have been shown to reduce entry errors by over 50%. ²¹ In light of these advantages, King Fahad Medical City implemented Epic, an electronic health record system, in February 2020 to support the digital transformation of hospital workflows. This study presents a step‐by‐step account of integrating Epic with automated compounding devices, focusing on the technical and clinical implementation aspects rather than evaluating clinical outcomes.

PRIOR TRANSITION TO THE ELECTRONIC HEALTH RECORD EPIC SYSTEM

Before transitioning to Epic, King Fahad Medical City relied on the Baxter ABACUS software to manage parenteral nutrition orders. This pharmacy‐based Windows application supported the preparation and management of parenteral nutrition formulations by enabling pharmacists to enter, verify, review, and cosign parenteral nutrition orders across adult, pediatric, and neonatal populations. The system accommodated both 3‐in‐1 and 2‐in‐1 parenteral nutrition orders, as well as ready‐made parenteral nutrition preparations. The order template included five key sections: patient information, diagnoses, allergies, order recommendations, and nutrition requirements. Clinicians could input order values for macronutrients and micronutrients, with the software automatically computing total daily requirements. Additionally, ABACUS displayed relevant order details, including osmolarity, administration parameters, energy contribution, ion content, and previous parenteral nutrition history. The software also featured two types of dose alerts: vendor‐established and user‐defined limits. However, vendor‐established limits could fail to flag potentially fatal doses, highlighting a critical safety limitation. ²²

Despite its use, ABACUS operated independently of the hospital's core electronic systems, posing significant challenges to the safety of parenteral nutrition and workflow efficiency. Healthcare providers were required to consult three separate, unintegrated systems: paper‐based patient charts, the Cortex system for laboratory values, and ABACUS for order entry. This fragmentation hindered access to real‐time laboratory data, medication history, progress notes, and essential dosing limits, increasing the potential for transcription errors and delays. The ABACUS interface, which communicated with the automatic compounding device (Baxter), used linear barcodes to facilitate the preparation of parenteral nutrition solutions. However, given the overall lack of integration between ABACUS, the paper charts, and the CorTex system, the process for entering parenteral nutrition was complex, potentially increasing the risk of errors.

ROADMAP OF INTEGRATION

Integrating the automated compounding devices with the Epic electronic health record involves technical and clinical aspects. The technical focus encompasses data exchange, interoperability, and flow optimization, whereas the clinical focus involves the development of parenteral nutrition models, which require expert staff, including pharmacists, physicians, and dietitians. With Epic's adoption, Baxter's parenteral nutrition compounder was replaced by the Medimix Multi (IMF) for parenteral nutrition preparation.

TECHNICAL ASPECTS

The National Alliance for Health Information Technology Standards

Interoperability enables healthcare systems to exchange and interpret data using standardized protocols. The National Alliance for Health Information Technology defines interoperability as the ability of information technology systems to communicate, exchange, and use data accurately and consistently. Achieving interoperability requires common standards (documents approved by organizations that provide guidelines for consistent use), interoperable technologies, and governance frameworks to ensure secure data exchange.

Levels of interoperability

The first level of interoperability is the technical level, which allows data transmission through Ethernet or wireless networks using operating systems such as Windows or Linux. ²³ For example, at King Fahad Medical City, Ethernet is used for networking, and Windows is used to integrate the parenteral nutrition compounder with Epic. The second level is the syntactic level, which ensures that the sender (Epic) and the receiver (parenteral nutrition compounder) can successfully exchange and interpret messages. ²³ This level includes consistency in data structure, format, and messaging standards.

Application programming interfaces

Application programming interfaces are sets of rules that allow software programs to communicate and exchange data within and across systems. ²⁴ Standards such as those from the National Alliance for Health Information Technology support this exchange. Epic uses application programming interfaces to facilitate interoperability with the parenteral nutrition compounder. ²⁵ To access Epic application programming interfaces, an application programming interface key is required, which serves as a secure credential for the compounder software to communicate with Epic. The application programming interface key is obtained through registration on the Unified Support Configuration and Discovery Interface portal.

Epic standards and integration process

Epic supports widely used interoperability standards, including Health Level 7 version 2 for messaging between healthcare systems and Fast Healthcare Interoperability Resources for integrating electronic health record data with external devices and applications. ²⁶

The integration process consists of five phases

• Design: determining data types, workflows, exchange methods (eg, Health Level 7 version 2), and security mechanisms.

• Development: applications are built using Epic's technical specifications, with unique client identification documents assigned for governance.

• Testing: functionality, performance, and security are validated in simulated environments before deployment.

• Connection: linking Epic with the automated compounding device.

• Data exchange: secure transfer of patient and prescription data following Epic's interoperability standards.

Security mechanisms

To protect sensitive health data, Epic employs OAuth 2.0, an open‐standard protocol that provides secure authentication and authorization for accessing system resources. This ensures that only approved applications and users can access patient data, and encryption protects the information throughout the process.

CLINICAL ASPECT

Team's participation

Computerized provider order entry is an integrated feature of electronic health records that allows healthcare providers to electronically enter orders for laboratory tests, imaging studies, consultations, diagnostic procedures, and medications, including parenteral nutrition prescriptions. ²⁷ Numerous parenteral nutrition order sets have been proposed, but only those that have been thoroughly reviewed and approved by professional clinical pharmacists and physicians are implemented. Proposed parenteral nutrition order sets adhere to ASPEN guidelines for prescribing, order review, compounding, labeling, dispensing, and administration, thereby minimizing the risk of errors and promoting safe practice. ¹⁷ A multidisciplinary team was formed to develop the Epic parenteral nutrition order template, comprising clinical and operational pharmacists, a pediatric gastroenterologist, a nurse practitioner, information technology specialists, Epic analysts, and representatives from an automated compounding device company. The clinical review team establishes parenteral nutrition policies, optimizes workflows, and reviews order activity, whereas Epic analysts manage system design, testing, and maintenance. Computerized provider order entry–integrated decision support tools enhance safety by customizing dosing regimens, calculating energy contributions, assessing osmolarity and mixture compatibility, and providing real‐time alerts.

Within Epic, parenteral nutrition orders are fully integrated with computerized provider order entry, directly connecting to pharmacy automated compounding devices without requiring an interface. Epic‐automated compounding device integration follows a structured, stepwise approach to ensure seamless interoperability and clinical efficiency.

First step: Data gathering

Establishing an electronic parenteral nutrition template requires comprehensive data collection for seamless system integration. A team of pharmacists, gastroenterologists, nurse practitioners, dietitians, and health informaticians ensured the accuracy of the data. Pharmacy staff gathered details on macronutrient and micronutrient concentrations, preparation methods, barcodes, administration instructions, energy contribution, osmolarity, and storage conditions, aligning them with the hospital formulary. Dextrose (70% default, 50% alternative), amino acids (Primene 10% for <2 years, Aminoplasmal 15% for older children and/or adults), and lipid emulsions SMOFlipid (20%) were configured in Epic. Electrolytes, multivitamins, trace elements, and additives like copper, chromium, selenium, and levocarnitine were standardized for compatibility with automated compounding devices. Multichamber parenteral nutrition bags (Olimel N9E for adults, Numeta G16E for neonates and/or pediatrics) were integrated into Epic. Pharmacy teams reviewed and validated drug‐specific data to ensure accuracy. The components and concentrations of parenteral solutions are shown in Table 1.

Table 1.

Components of parenteral solutions and concentrations.

Medication name	Original size of container	Original type of container	Machine name	Machine size of container	Machine type of container	Specific gravity
Water for injection, sterile IV solution	1000	Bag	Water for injection 1000 ml Baxter	1000	Bag	1
Parenteral amino acid 15% combination number 4 IV solution	500	Bottle	Aminoven 15% Baxter	500	Bottle	1.05
Parenteral amino acid 10% combination number 11 (pedi) IV solution	250	Bottle	Aminoven infant 10%	250	Bottle	1.03
Calcium gluconate 100 mg/ml (10%) IV solution	10	Plastic bottle	Calcium gluconate 10% Baxter	1000	Bag	1.05
Sodium chloride 2.5 mmol/ml IV solution	50	Vial	Sodium chloride 2.5 mmol/ml 14.61% Baxter	1000	Bag	1.09
Potassium chloride 2 mmol/ml IV solution	20	Plastic bottle	Potassium chloride 15% 2 mEq/ml Baxter	1000	Bag	1.09
Sodium glycerophosphate 1 mmol/ml IV solution	20	Plastic bottle	Sodium glycerophosphate Baxter	1000	Bag	1
Sodium acetate 2 mmol/ml IV solution	20	Plastic bottle	Sodium acetate 2 mmol/ml	1000	Bag	1.08
Trace elements pedi Zn–Cu–Mn–Se–F–I 521 mcg–53.7 mcg–3.6 mcg/ml IV	20	Plastic bottle	Trace elements (pedi)	500	Bag	1.001
Zn sulfate 1 mg/ml IV solution	10	Amp	Zn sulfate 1 mg/ml	250	Bag	1.01
Magnesium sulfate 2 mmol/ml (50%) injection solution	50	Vial	Magnesium sulfate 50% Baxter	50	Vial	1.004
MVI, pedi number 1 with vit K 80 mg–400 unit–200 mcg/5 ml IV solution	5	Vial	MVI ‐ pedi/NEO Baxter	250	Bag	1.0102
Levocarnitine 200 mg/ml IV solution	10	Amp	Levocarnitine Baxter	250	Bag	1
Se 40 mcg/ml IV solution	10	Vial	Se 10 mcg/ml	100	Bag	1
Water for injection, sterile IV solution	1000	Bag	Water for injection 1000 ml Baxter	1000	Bag	1
Water for injection, sterile IV solution	2000	Bag	Dextrose 70% Baxter	2000	Bag	1.26

Abbreviations: Amp, ampoule; Cu, copper; F, fluorine; I, iodine; IV, intravenous; Mn, manganese; MVI, multivitamin infusion; NEO, neonate; pedi, pediatric; Se, selenium; vit, vitamin; Zn, zinc.

Second step: Build order sets components

Order sets, a form of clinical decision support system, are preestablished groups of related orders pertaining to parenteral nutrition components. The parenteral nutrition template features an intuitive single‐window interface that simplifies organization and navigation. By standardizing parenteral nutrition order prescriptions, templates promote consistency of care and reduce variations in care. When a parenteral nutrition formulation is ordered, a pop‐up screen displays various parenteral nutrition formulations, allowing the selection of an order set aligned with the treatment plan. This safe presentation of product selection menus minimizes the likelihood of medication selection errors (Figure S1).

Prescribers select 3‐in‐1 or 2‐in‐1 parenteral nutrition order sets for adults and pediatrics. The 3‐in‐1 parenteral nutrition combines dextrose, amino acids, lipid emulsion, and micronutrients in one entry, whereas 2‐in‐1 parenteral nutrition includes dextrose and amino acids in one bag with lipid emulsion provided in a separate linked order. To ensure timely processing, parenteral nutrition orders must be submitted by 1 pm. A best practice alert will notify prescribers after the cutoff, delaying preparation until the next business day.

Following 1 pm, a best practice alert is triggered, and prescribers can then view warning messages explaining that parenteral nutrition orders are not permitted after working hours. They must select “Accept” to acknowledge the best practice alert. The warning message states, “parenteral nutrition orders placed after 1:00 pm will not start until tomorrow. Consider using a nutrition bridge until parenteral nutrition starts tomorrow. If you feel this parenteral nutrition is needed today, please call the pharmacy.”

Accepting the best practice alert message opens the “SmartSet” functionality, which provides a group of clinical orders including a parenteral nutrition template and dextrose 5% solution, in which the parenteral nutrition template indicates that the order is disabled (not orderable), whereas dextrose 5% is provided as an alternative nutrition support until the next day (Figure S2B).

The parenteral nutrition order entry system features an integrated input and output window into a single screen, where physicians and/or clinical pharmacists can enter mandatory parenteral nutrition component values in the input window (Figure S2 and Figure S2M). The output window simultaneously displays the calculated total amount of each parenteral nutrition component, energy contribution, glucose infusion rate, osmolarity, and a calcium phosphate curve figure.

The input window includes various fields for data entry, such as patient weight, total fluid intake, frequency, administer over (infusion rate), taper up and down option, duration, starting infusion time, infusion site, parenteral nutrition indication, and the amounts of dextrose, amino acids, lipids, electrolytes, multivitamins, trace elements, and other additives (Figure S2C). First, clinicians should ensure accurate weight selection for adult and pediatric patients. For adult patients, prescribers can choose between actual body weight, ideal body weight, or adjusted body weight, following ASPEN guidelines for individuals with obesity or for individuals who are nonobese. For pediatric patients (<2 years of age), weight‐based dosing is used for the administration of macronutrients and micronutrients. Pediatric patients (<2 years of age) use weight‐based dosing for the administration of macronutrients and micronutrients. Weight options in pediatric order sets include “previous weight,” “recorded weight,” and “order‐specific weight,” which are entered manually by the prescriber (Figure S2 and Figure S2M). The previous weight option is presented if the weight varies throughout the hospitalization. In contrast, the recorded weight option can reflect the current weight, and order‐specific weight is an option when the prescriber wants to enter the intended weight. Second, the system uses a unified unit of mL for total fluid intake dosing, avoiding variations like cc, ml, mLs, or mls. The authors recommend avoiding ml/kg dosing in pediatric order sets for total fluid intake if the patient's weight is expected to change during hospitalization, especially for long‐term parenteral nutrition therapy aimed at promoting growth and weight gain. This is to prevent concurrent changes in electrolyte doses that accompany weight fluctuations. To prevent confusion and potential errors, trailing zeros at the end of parenteral nutrition solution dosages should not be used (eg, 1900 ml instead of 1900.0 ml). The frequency option determines the parenteral nutrition infusion mode, which can be either cyclic or continuous (Figure S2C).

Parenteral nutrition infusion can be continuous (24 h) or cyclic (16–20 h), requiring tapering to prevent complications like hypoglycemia. Infusion rate, site (central, peripheral, or umbilical), and osmolarity limits (900–1100 mOsm/L) are set, with a clinical decision support system providing alerts to ensure safe peripheral administration. The start time is fixed at 5 pm per institutional policy. Parenteral nutrition dosing varies between adult and pediatric templates. For adults, dextrose, amino acids, and lipids are presented in grams per day, whereas for pediatrics, they are expressed as a percentage concentration per liter. Standardization refers to using fixed concentrations and formulations of dextrose (70% default, 50% alternative), amino acids (10% for <2 years, 15% for older children and adults), and lipid emulsions (20%), all aligned with ASPEN guidelines and configured for compatibility within Epic. Electrolytes and additives are prescribed daily, including sodium, potassium, calcium, magnesium, copper, chromium, insulin, heparin, levocarnitine, vitamin K, and zinc sulfate. Clinical decision support system alerts enforce safe dosing limits and prevent incompatibility issues, ensuring accurate parenteral nutrition prescribing and preparation.

Integrated within each order set is a sidebar (output window) that provides a summary of parenteral nutrition orders (Figure S3B). This summary encompasses order details, macronutrients, micronutrients, energy contribution, mixture compatibility, and a calcium phosphate solubility curve (Figure S3D). Furthermore, a parenteral nutrition navigator tool was established to facilitate parenteral nutrition workflow, enabling prescribers to view their parenteral nutrition order history in a single window, alongside consultation orders, fluid and energy management, patient medications, laboratory results, comprehensive and daily assessments, plans, and progress notes. Additionally, the parenteral nutrition navigator conveniently displays the last four parenteral nutrition prescriptions, allowing users to easily track the history of parenteral nutrition orders. This feature provides a quick reference for reviewing past parenteral nutrition orders and identifying trends or patterns (Figure S3C).

Parenteral nutrition navigator tool

A parenteral nutrition navigator tool was established to facilitate parenteral nutrition workflow, enabling prescribers to view their parenteral nutrition order history in a single window, alongside consultation orders, fluid and energy management, patient medications, laboratory results, comprehensive and daily assessments, plans, and progress notes. The parenteral nutrition navigator conveniently displays the last four parenteral nutrition prescriptions, enabling users to track their parenteral nutrition order history. This feature provides a quick reference for reviewing past parenteral nutrition orders and identifying trends or patterns (Figure S3C).

Calcium and phosphate solubility curve

Another feature is the calcium and phosphate solubility curve for inorganic phosphate products (Figure S3D). The curve illustrates the degree of solubility of calcium and phosphate with the entered doses, considering the percentage of amino acids to enhance solubility and prevent precipitation. The curve is not warranted with sodium glycerophosphate; however, its use requires dosing limits for calcium and phosphate, with a maximum dose of 48 mmol and 120 mmol of calcium and phosphate, respectively.

Types of parenteral nutrition order sets

2‐in‐1 parenteral nutrition order sets

The 2‐in‐1 parenteral nutrition order set is similar to the 3‐in‐1 but links fat emulsion separately for stability. Required fields include total volume, tapering, frequency, and infusion site, with optional fields for indication, administration, electrolytes, and additives. For adults, 15% amino acids and 70% dextrose are required. Electrolyte options include sodium chloride, acetate, glycerophosphate, potassium chloride, acetate, phosphate, calcium gluconate, and magnesium sulfate. Neonatal (≤2 years) sets use 10% amino acids, whereas older pediatrics use 15%. Dosing units are measured in milliliters (ml), milliliters per kilogram (ml/kg), or milliliters per hour (ml/h). The fat emulsion is linked to 2‐in‐1 orders for simultaneous submission, with a 12‐h hang time for stability. The infusion rate defaults to 20 h (pediatrics) and 24 h (neonates) per King Fahad Medical City policy. Dosing (g or g/kg) and frequency are mandatory fields.

3‐in‐1 parenteral nutrition order sets

Total nutrient admixtures comprise amino acids, dextrose, and lipid emulsion. The 2‐in‐1 inpatient parenteral nutrition order set is designed similarly to the 3‐in‐1 order set, except that the fat emulsion order is linked to the 2‐in‐1 order set because of stability and administration concerns (Figure S4B). With the ease of administering only one container, adult 3‐in‐1 parenteral nutrition is used for patients receiving home health nursing care. The order set for an adult 3‐in‐1 parenteral nutrition consists of fields that the provider is required to fill. Mandatory fields include the total volume, hour‐based tapering, frequency (cyclic or continuous), infusion site (central, peripheral, or umbilical), and selection of lipid emulsion type (Figure S4B). Indication, administration instruction, electrolytes, and additives are considered optional fields (Figure S4B). The total volume of parenteral nutrition can be entered manually or selected from predetermined options (1400 ml, 1800 ml, 2400 ml, or 3000 ml). Lipid concentration is fixed at 20%, whereas amino acids and dextrose are set at 15% and 70%, respectively. For adults, amino acid and dextrose dosing are expressed in grams per day, whereas for pediatrics, dosing is expressed as a percentage concentration per liter. Various electrolyte options are available, including sodium chloride, acetate, or glycerophosphate; potassium chloride, acetate, or phosphate; calcium gluconate; and magnesium sulfate. Additives include multivitamins and trace elements (Figure S4B).

The neonatal and pediatric 3‐in‐1 order set mirrors the adult version in terms of mandatory and nonmandatory fields. Unique aspects of the neonatal and pediatric (<2 years) order set include the use of 10% amino acids and weight‐based dosing with options for reported, dosing, or order‐specific weight, which is manually entered by the prescriber. Clinicians can incorporate additive medications such as levocarnitine, metoclopramide, and heparin.

Home and home healthcare parenteral nutrition order sets

At home, the parenteral nutrition model allows patients or caregivers who are trained and competent to administer their 2‐in‐1 parenteral nutrition solution to order a prescription for a period of 1 week (up to 14 days in cases of travel). Multivitamins are packaged separately to extend the stability of the preparation. Upon administration, patients or caregivers add multivitamins to the 2‐in‐1 parenteral nutrition solution. Prescribers can order home parenteral nutrition using the same inpatient order sets for 2‐in‐1 formulations available through the Epic system. Typically, 3‐in‐1 formulations are not used in this population because of concerns about shorter stability. The order duration is selected based on the total number of bags each patient needs. In contrast, in the home healthcare parenteral nutrition model, home healthcare nurses visit the patient's home to administer a 3‐in‐1 parenteral nutrition admixture. Orders are prescribed daily using the 3‐in‐1 inpatient order sets (Figure S4B). Upon preparation, the pharmacy team adds multivitamins to the admixture.

Multichamber bag order sets

The Olimel N9E and Numeta G16E order sets are the two ready‐made multichamber parenteral nutrition bags available in the King Fahad Medical City formulary. Both sets have obligatory fields, such as tapering and frequency, and a sidebar that provides details of projected energy and fluid management, along with daily totals of calories in kcalories and volume in milliliters. According to the manufacturer's recommendations, electrolytes and additives can be added (Figure S4A).

Third step: Verification page layout

Following the transmission of the parenteral nutrition order to the Epic pharmacy, the pharmacist can access the prescription in the pharmacy queue. The verification page layout presents the parenteral nutrition prescriptions, allowing the pharmacist to review and approve the order for label generation (Figure S5A and S5B).

Fourth step: Label generation and design

A barcode readability issue arose when the automated compounding device scanner failed to recognize two‐dimensional (2D) matrix barcodes (e.g. Data Matrix/QR), which encode data in both horizontal and vertical directions, thereby preventing the preparation of parenteral nutrition. To resolve this, the Epic team generated two labels after verifying the parenteral nutrition order. The production label, used by nurses and physicians, includes patient details, infusion site and rate, ingredient doses, expiration date, and administration instructions, with a 2D barcode for bar code medication administration but not for automated compounding device use. The mixture label, designed for pharmacy staff, displays parenteral nutrition ingredient doses and dispensing volumes and features a one‐dimensional (1D) barcode (a linear barcode in which data are encoded along a single axis) that the automated compounding device scanner reads to initiate parenteral nutrition preparation. This 1D barcode integration enables direct data exchange between Epic and an automated compounding device, eliminating the need for interface software.

Another challenge we encountered was the identification code mismatch for items used for preparation. The codes designated for items during compounding and those entered through the electronic system are different, which means that scanning a specific item during compounding may result in an inaccurate designation for that item. An example was the manual addition of levocarnitine to the mixture label. The label, designed for pharmacy staff, displays parenteral nutrition ingredient doses and dispensing volumes and features a 1D barcode that the automated compounding device scanner reads to initiate parenteral nutrition preparation. This 1D barcode integration enables direct data exchange between Epic and an automated compounding device, eliminating the need for interface software.

Another challenge we encountered was the mismatch in identification document codes for items used in preparation. The codes designated for items during compounding and those entered through the electronic system are different, which means that scanning a specific item during compounding may result in an inaccurate designation for that item. An example was the manual addition of levocarnitine bags because of a lack of proper coding. We resolved this issue by creating a mapping draft that included all codes for each ingredient and addressing any discrepancies in code integration between Epic and the automated compounding system.

Fifth step: Electronic parenteral nutrition referral form (consultation form)

In the Epic system, a referral form, also known as a consultation form, serves as a communication tool to convey patient information, the reason for the referral, and relevant clinical details (eg, infusion line location, catheter placement confirmation with x‐ray if central line used, or fluid restriction status) to the consulted parenteral nutrition specialist (Figure S5C). This information is crucial for the parenteral nutrition specialist to make informed decisions about the patient's care and ensure they receive the appropriate treatment. Through computerized provider order entry, referring clinicians can electronically submit the referral form using the “Pharmacy to dose parenteral nutrition” order set, sending the information (via push message) to the clinical pharmacist. The consulted clinical pharmacist reviews the referral details to gain a comprehensive understanding of the patient's condition and the referring clinician's concerns. Open communication is needed between the consulted and referring clinicians to address any questions or concerns.

Sixth step: Staff training

Before implementing the new automated compounding device with the parenteral nutrition model, the pharmacy team underwent vendor‐led training to ensure a smooth transition. The program covered operation, maintenance, troubleshooting, and compatibility with parenteral nutrition solutions. Staff received hands‐on training and participated in simulations to apply their knowledge. The training lasted 1 month, with three staff members from each group being trained on a weekly basis. Competency checklists ensured readiness before beginning operations, equipping the team to use the automated compounding device while maintaining high standards for parenteral nutrition preparation effectively.

Monitoring was incorporated into this step to evaluate success. Key activities included structured process mapping, daily feedback from frontline users, beginning operating readiness checklists, and post‐training evaluations. Training sessions were conducted for pharmacists and clinical users, and feedback loops with Epic analysts allowed for iterative improvements.

METHODS

To monitor the success of the implementation, a qualitative, process‐oriented approach was employed throughout the transition. Key activities included structured process mapping, daily feedback from frontline users, beginning operating readiness checklists, and post‐training evaluations. Training sessions were conducted for pharmacists and clinical users, and feedback loops with Epic analysts allowed for iterative improvements.

DISCUSSION

Several key aspects are essential for developing and integrating parenteral nutrition through a newly incorporated healthcare system at any medical facility. Pursuing a stepwise approach, such as data gathering, order set and label design, application, and amplification, has facilitated implementation. A survey of ASPEN members was conducted in 2012 to evaluate the quality and efficiency of commercially available electronic health record systems. In fact, there was some variation in reported ratings, with the highest‐rated vendor receiving an overall satisfaction score of 90.3% and the lowest‐rated vendor receiving an overall satisfaction score of 57.8%. ²⁸

Designing the system required input from providers, nurses, and pharmacists to ensure order, the medication administration record, and verification layouts met practice standards. Ongoing adjustments, clear communication between informaticists and users, and regular training were essential for quality improvement. A key benefit of Epic is consolidating parenteral nutrition orders within a single system, eliminating the previous fragmentation of orders and laboratory values. The parenteral nutrition consultation tool enables the quick retrieval of patient information for parenteral nutrition, thereby enhancing workflow. Implementation challenges included coding mismatches, data mapping discrepancies, and 2D barcode recognition failures, all of which were resolved through multidisciplinary collaboration, Epic analyst support, and iterative system refinements.

Although formal outcome data were not included in this manuscript, we acknowledge the importance of measuring implementation success through error tracking and workflow metrics. We recommend that future evaluations incorporate preintegration and postintegration outcome assessments and resource analyses to support evidence‐based decision‐making. Continuous quality and safety measures remain crucial for maintaining long‐term system performance and ensuring optimal patient care.

Lessons learned

Integrating automated compounding devices with the Epic electronic health record system highlighted several key lessons in the management of parenteral nutrition. Early multidisciplinary collaboration among pharmacists, physicians, dietitians, information technology specialists, and Epic analysts was essential to align system design with clinical workflows. Continuous system refinements, including enhancements to barcode recognition, alert parameters, and user interface design, contributed to improved efficiency and usability. The integration process highlighted the importance of tailoring clinical decision support tools to meet patient‐specific needs and institutional guidelines. To ensure progress throughout the implementation, success was monitored using qualitative methods, including process mapping, beginning operating readiness checklists, structured training evaluations, and regular feedback from end‐users and stakeholders. These approaches provided insight into workflow optimization and highlighted areas needing iterative improvement.

The implementation involved notable investments, including infrastructure upgrades, staff training programs, and vendor coordination. The perceived benefits included enhanced workflow standardization, reduced transcription errors, minimized manual data entry, and improved access to consolidated patient information. These operational gains contributed to more streamlined parenteral nutrition management processes. Future evaluations should incorporate preimplementation and postimplementation outcome metrics, such as error rates and efficiency benchmarks, along with economic assessments, to provide a more comprehensive evaluation of impact and support data‐driven decision‐making in other institutions.

Finally, comprehensive staff training and proactive change management proved vital for successful implementation and long‐term adoption. Lessons from this experience may serve as a roadmap for institutions planning similar transitions toward integrated, electronic health record–based parenteral nutrition management systems.

CONCLUSION

Integrating automated compounding devices with the Epic electronic health record at King Fahad Medical City resulted in operational improvements in parenteral nutrition safety, efficiency, and workflow standardization. Standardized data exchange, clinical decision support, and barcode verification facilitated a more reliable prescribing and compounding process, aligning with the ASPEN guidelines. The project involved key investments in infrastructure, staff training, and cross‐functional coordination. Initial implementation challenges, such as barcode recognition issues and code mismatches, were addressed through system refinement and multidisciplinary collaboration. Although this manuscript does not include formal outcome data, we acknowledge the importance of quantifying implementation success. Future evaluations should incorporate outcome metrics such as error rates and workflow efficiency benchmarks, along with economic assessments, to guide evidence‐based adoption. This model offers a replicable framework for institutions seeking to integrate electronic health records with parenteral nutrition, highlighting the importance of strategic planning, stakeholder engagement, and continuous quality improvement in health informatics transformation.

AUTHOR CONTRIBUTIONS

Nora Albanyan contributed to the conception and design of the research. Ahmed Ali Alrashed, Jude Howaidi, Aljawharah Binrokan, and Razan Orfali contributed to the research design. Yahya Mohzari, Ahmed AlRassan, and Dena Alahaideb contributed to the data acquisition. All authors contributed to the manuscript and the Supporting Figures.

CONFLICT OF INTEREST STATEMENT

None declared.

Supporting information

Supporting information.

Supporting information.

Supporting information.

Supporting information.

Supporting information.

Supporting information.

Albanyan N, Alrashed AA, Howaidi J, et al. Integration of electronic health records with automated compounding for parenteral nutrition: a narrative review. J Parenter Enteral Nutr. 2026;50:161‐169. 10.1002/jpen.70045

DATA AVAILABILITY STATEMENT

Data sharing is not applicable.