Enhancing Participant Safety through Electronically Generated Medication Order Sets in a Clinical Research Environment: A Medical Informatics Initiative

Abstract

While clinical medicine is often well supported by health system information technology infrastructure, clinical research may need to create strategies to use clinical‐medicine informational technology tools. The authors describe a medication‐safety initiative that was carried out in a National Institutes of Health Clinical and Translational Science Award (CTSA)‐sponsored clinical research environment. A web based, medical informatics application was designed and implemented that allowed research groups to build protocol specific, electronic medication templates that were subsequently used to create participant‐specific medication order sets for conductance of clinical research activities in the CTSA‐sponsored clinical research environment. The medical informatics initiative eliminated typewritten or handwritten medication orders, created research protocol‐specific templates meeting institutional order‐writing requirements, and formalized a rigorous review and approval process. Enhancing safety in medication ordering and prescribing practices in a clinical research environment provided the background for multidisciplinary cooperation in medical informatics. Clin Trans Sci 2010; Volume 3: 312–315

Introduction

With increasing use of information technology in the health care environment, medication use and prescribing‐process studies have begun to evaluate the impact of standardized, electronic medication‐ordering systems on reduction of medication errors and adverse drug events in various patient populations. ¹ , ² , ³ , ⁴ Computerized prescriber order entry (CPOE) has been adopted by many health care institutions. Until health care information technology transitions entirely away from paper‐based medication orders, the use of standard, preprinted medication order sets will continue to be used successfully in various health care environments to meet clinical and research goals. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰

Standardized order sets have been used to promote adherence to clinical guidelines and enhance medication safety in clinical care environments for patient‐controlled analgesia, ⁵ community‐acquired pneumonia, ⁶ pediatric sedation, ⁷ and chemotherapy. ¹ Safety features that have been incorporated into clinical care order sets include identification of adverse event risk criteria, determination of need for supportive measures, and renal insufficiency medication dose adjustments. ⁸ In order for a pharmacy to prepare and dispense investigational medications according to a clinical research protocol, medication orders for research participants must include participant identification, research protocol number, study visit information, and accurately written investigational medication orders. ⁹ , ¹⁰ In the rigorous environment of clinical research, standardized medication order sets are vitally needed to support participant safety through reduction and elimination of medication‐error sources such as illegible handwritten orders, misuse of abbreviations on medication orders, and inaccurately prescribed medication orders. ⁹ , ¹⁰

The Mayo Clinic Center for Translational Science Activities (CTSA) clinical research unit (CRU) is an integrative inpatient, outpatient, and mobile research program supported in part by a grant from the National Institutes of Health (NIH). In 2008, the CRU environment had 30 beds, averaged over 13,000 participant visits, and supported 258 research protocols. Two investigational drug service (IDS) pharmacies dispensed investigational medications and worked closely with the CRUs. An evaluation from 2007 showed that the IDS pharmacies dispensed more than 8,000 compounded sterile preparations and unit dose oral medications for 254 protocols.

Mayo Clinic is an 1,100 bed, tertiary care, academic health care system that also includes outpatient clinics. To date, hospital inpatient‐CPOE implementation has been completed in an incremental, multiphase rolling implementation. During the hospital inpatient transition period between paper‐chart orders and the CPOE environment, document‐imaging technology (Pyxis®Connect, Cardinal Health, Dublin, OH, USA) was adopted by the institution. This process mirrored the experiences of other institutions. ¹¹ , ¹² , ¹³ Briefly, scanners located in patient care areas transmitted an archivable image of medication orders to a medication order entry queue for a pharmacist to enter into a medication‐dispensing computer system. This document imaging technology was implemented in the IDS pharmacies to support the CRUs and was the means of transmitting images of nonstandardized medication orders to the pharmacies. The clinical research environment was not included in the original inpatient hospital scope of the CPOE project, and a timeline for anticipated CPOE implementation was unknown. In order to keep pace with the rapid electronic information technology changes in the clinical care environment, an urgent need for a medical informatics application was identified in the clinical research environment.

Methods

Initiative overview

Nonstandardized, typewritten, or handwritten medication orders were used in the clinical research environment prior to 2007. In 2007, the CRU informatics team began a formalized process of collecting input from users and institutional oversight groups that would culminate in a web based, medical informatics application used to create research protocol‐specific medication‐order templates that would ultimately generate participant‐specific medication order sets for clinical research activities in the CRU. Multidisciplinary round table discussions of existing institutional systems and processes were conducted and deliverables were determined based on a phased approach. Research requirements were identified based on existing order sets and practices at the institution. Key users, group members, and institutional committees were identified and included a CRU informatics specialist, CRU nursing staff, CRU nursing education specialist, IDS pharmacists, institutional medication safety pharmacists, members of research groups, and institutional oversight committees including Order Set Protocol Approval Group (OSPAG). Pilot testing was conducted before full release. Targeted application training for research groups, nursing, and IDS pharmacists was undertaken. Feedback from the users using the web‐based medication order sets was collected and in response, enhanced one‐on‐one sessions were provided to users to support creation of medication‐order templates using the web‐based medical informatics application.

Design objectives

Through discussions with multidisciplinary users, a list of design objectives was generated and used to build the web based, medical informatics application.

• 1Lean process: Involve key institutional groups to streamline the review, approval, and amendment process for efficient use of resources.
• 2User friendly application: Build an application that could be easily learned and used.
• 3Standardized medication order writing: Comply with institutional medication order‐writing standards and create new standards for research scenarios where previously clinical medication order‐writing standards were unable to address research needs.
• 4Formalized review process: Develop a standardized method for medication order review and approval by clinical research user groups and institutional oversight committees.
• 5Version control: Provide the ability to track edits.
• 6Controlled access: Implement role‐based access to the medical informatics application.

System description

The web based, medical informatics application was an addition to the suite of CRU applications that integrated participant and investigator demographic data, scheduling resources, and protocol information. Overall, the medical informatics application served three main roles. First, it allowed research group users to build a protocol‐specific medication‐order template. Second, the template was rigorously reviewed and approved in an electronic environment by oversight groups to meet practice and order‐writing standards. Third, the template was used to generate and print participant‐specific orders that were transmitted to the IDS pharmacies.

In order to build a protocol‐specific medication‐order template, a JSP‐generated user interface led the user to create a patient‐specific medication order set through separate input pages and allowed the user to specify order elements including, investigational medication information (date/time, dosage, route, frequency, and instructions), intravenous administration instructions (placement and solution), anthropometric measurement requirements (height, weight, serum creatinine, body mass index, and body surface area), specialized alerts, and nursing care instructions. All back‐end processing was done using java and java servlet technology. The output format for each patient specific, investigational medication order set was an Adobe® ColdFusion application software‐rendered PDF document. The process to render the PDF document was a stand‐alone process, which could be implemented using any technology that was capable of transforming data into a PDF document. ColdFusion was selected due to its built‐in support for PDF generation. A template/instance model was adopted with reusability in mind.

Once a medication‐order template was created, reviewed, and approved for a specific research protocol, it could be used for each participant enrolled on the protocol, thus avoiding duplication of effort and erroneous manual input. Pull‐down lists with standardized selection criteria were incorporated wherever possible to eliminate nonstandard and noninstitutional abbreviations.

A formalized review process was created to support electronic tracking and governing of templates. When a draft version of a medication‐order template was finished, it was submitted (in sequence) for approval to individuals who represent user groups (i.e., nursing, pharmacy, research group, and institutional oversight committee). The draft template was reviewed for accuracy, clarity, and standardized terminology and nomenclature. Each reviewer was required to select one of the following options: approve, exempt, or disapprove. An “approve” or “exempt” decision advanced the approval process, while a “disapprove” decision halted the process and required reassessment. Reviewers may have been exempted from the process and marked “exempt” if the research protocol did not require their practices’ involvement. The template was considered complete and ready for use after user and reviewer groups had indicated approval.

The life cycle of a medication‐order template was based on different stages and was indicated by a specific status. An “in‐progress” status indicated that the medication‐order template was being edited. A “review” status indicated that the draft version was being actively reviewed. A “complete” status indicated that the medication‐order template was approved by all groups. Based on feedback from the review and approval process, the medication‐order template might go iteratively through the “in‐progress” and “review” stages. Status‐based email was integrated into the application to communicate the progress of the medication‐order template.

Version control tracked author, content, and date of revision enabling the research community to archive and retrieve versions of a medication‐order template. In addition, versioning created the opportunity to generate a new medication‐order template from a previously created template. This allowed medication‐order templates that were in use in the CRU environment to be modified. An example of this function was the incorporation of protocol amendment‐driven changes into the medication‐order template.

When creating medication‐order templates, a rudimentary medication dictionary allowed research groups to specify investigational medications and commonly prescribed medications (e.g., Acetaminophen [Tylenol^®, Johnson & Johnson, Fort Washington, PA, USA.]). Once defined in the medication dictionary, the medications were available via drop‐down lists, which allowed for more consistent medication entries for building each protocol. Drop‐down lists included investigational medications, dosage, units of measure, and frequency of administration.

Finally, electronic signature capability was added. This functionality allowed prescribers to sign medication orders off‐campus and during CRU off‐hours.

Overview of the medication‐order template generation and approval process

Designated members of the research group built a unique medication‐order template based upon an Institutional Review Board approved research protocol ( Figure 1 ). A member of the research group, typically the lead study coordinator, attended a meeting with an IDS pharmacist and a member of OSPAG to create a draft of the protocol specific, medication‐order template. The draft was then electronically submitted for review and approval. Review and approval was conducted by IDS pharmacy, CRU nursing, the research protocol’s principal investigator, and OSPAG. Electronic submission was required by OSPAG for final approval and archive. If a change was required to comply with institutional medication order‐writing standards, then the medication‐order template was returned to the research group for rework and resubmission.

Figure 1.

Flow diagrams demonstrating the real‐time relationship between the electronic‐review process and the status of the medication‐order template as tracked by the medical informatics application.

Results

Although formal surveys or metrics were not performed prior to implementation of the electronic medical informatics initiative, a postimplementation convenience survey was administered to users. The survey reflected positive feedback and improvements to the system and to participant care. Of the 105 staff members surveyed, 59 (56%) responses were returned. Sixty‐five percent of the respondents indicated that intrateam communication was positively impacted, 71% indicated an improvement in participant safety, 65% thought the new process was more efficient than the old process, and 19% indicated no difference in efficiency. To date, more than 400 research medication‐order templates are electronically maintained.

Discussion

Medication order sets have been used in many clinical settings to facilitate and improve the quality of medication ordering and prescribing. Development of clearly written and well‐designed medication orders address important aspects of the medication‐ordering process including content, readability, style, and safety. ¹⁴ Clinical initiatives to enhance the process of medication ordering and prescribing have been shown to aid in reducing errors and omissions for inpatient and outpatient populations. ¹⁵ , ¹⁶

Clinical research encompasses a unique population of participants who are enrolled in tightly controlled, investigational protocols. Clinical research protocols have stringent inclusion and exclusion criteria and may include randomized and double‐blinded study group allocations. Failure to comply with a research protocol may result in protocol deviations and increased risk to participants. Therefore, in addition to the usual medication ordering and prescribing process, investigational medication orders for clinical research protocols often involve a higher level of scrutiny for safe and accurate medication dispensing. Thus, order‐writing requirements and a formal, rigorous review and approval process were built into the design of a medical informatics application to support research groups using the CTSA‐sponsored clinical research environment.

Conclusion

A multidisciplinary clinical research group at a large academic health care system successfully designed and implemented a medical informatics application with the ability to create medication‐order templates that ultimately allowed research users to generate participant‐specific medication order sets in an integrative inpatient, outpatient, and mobile clinical research program. The medical informatics application was incorporated into the processes and workflow of the CRU environment and was positively received by users.

Conflict of Interest

The authors have no financial support or personal connections that could be perceived to bias their work.