Computerized physician order entry: lessons learned from the trenches

Implementation of computer physician order entry (CPOE) demands planning, teamwork, and a steep learning curve. The nurse-driven team at the hospital unit level is pivotal to a successful launch. This article describes the experience of one NICU in planning, building, training, and implementing CPOE. Pitfalls and lessons learned are described. Communication between the nurse team at the unit and the clinical informatics team needs to be ongoing. Self-paced training with realistic practice scenarios and one-on-one "view then practice" modules help ease the transition. Many issues are not apparent until after CPOE has been implemented, and it is vital to have a mechanism to fix problems quickly. We describe the experience of "going live" and the reality of day-to-day order entry.

Intervention to increase the proportion of acute myocardial infarction or coronary artery bypass graft patients receiving an order for aspirin at hospital discharge

BACKGROUND

Chronic aspirin therapy is recommended by the American College of Cardiology/American Heart Association (ACC/AHA) following acute myocardial infarction (AMI) and by the Society of Thoracic Surgeons(STS) following coronary artery bypass graft (CABG). Aspirin therapy at discharge following a hospitalization for AMI or CABG is a common pay for-performance indicator used by third-party payers and was included asa quality measure in the Centers for Medicare & Medicaid Services (CMS)/Premier Hospital Quality Incentive Demonstration initiated in 2003. A formal prescription for aspirin, such as required for other cardiovascular drugs,could serve as a reminder to all health care providers (doctors, nurses, and pharmacists) to include aspirin on a discharge medication list.

OBJECTIVE

To evaluate if an aspirin prescription placed in the patient chart shortly after hospital admission would increase compliance with guidelines for aspirin use at discharge in patients with AMI or CABG.

METHODS

This was a single-center prospective pre-intervention to post intervention comparison study in a 411-bed hospital. Patients admitted during the 3-month period from July through September 2008 with an AMI or undergoing CABG surgery served as the pre-intervention group, and patients admitted during the 3-month period from January through March 2009 were in the post-intervention group. The intervention included multiple educational sessions with clinical staff, conducted both prior to and during the pilot, and blank pre-printed aspirin prescriptions placed in the charts of patients for whom no contraindication to aspirin was present. The blank prescriptions were then completed by the attending physician (or physician extender), and the discharge nurse used the completed aspirin prescription, with other prescriptions and written orders, as a reference when creating the discharge medication list. The primary outcome measure was the percentage of patients who had aspirin documented on the discharge medication list. Differences in compliance rates in the comparison and pilot periods were assessed using the Pearson chi-square test.

RESULTS

A total of 458 patients were identified with a CABG procedure and/or an admitting diagnosis of AMI; 447 met inclusion criteria, and 11 were excluded (1 patient in each of the groups had a contraindication to aspirin due to bleeding, and 9 died during hospitalization). The intervention was associated with an increase in the proportion of patients with aspirin documented on the discharge medication list, 266 of 269 patients (98.9%)compared with 169 of 178 patients (94.9%, P = 0.012) in the pre-intervention group. In the subsample of patients with CABG, 54 of 59 (91.5%)patients in the pre-intervention group had aspirin documented on the discharge medication list compared with 100% of 66 patients in the postintervention group (P = 0.016). In the subsample of patients with AMI, aspirin was documented in 115 of 119 (96.6%) patients in the pre-intervention group versus 200 of 203 (98.5%) in the post-intervention group (P = 0.263).

CONCLUSION

A quality improvement initiative that included clinical staff education and placement of aspirin prescriptions in patient charts during the hospital stay was associated with an increase in the proportion of patients who had aspirin documented on the discharge medication list for the overall sample of patients with AMI or CABG and for patients with CABG alone but not for the quality measure for AMI patients.

Identifying medication errors in surgical prescription charts

BACKGROUND

Each year 200 million prescriptions for children and adolescents are issued in the UK, with a 1.5 per cent prevalence of errors.

AIM

To identify and quantify medication errors on surgical children's prescription charts over a four-month period at two hospital sites.

METHOD

Retrospective review of the prescription charts of 175 children at a children's hospital and a children's unit.

RESULTS

Errors totalled 301, the most common was overwriting of a prescription, the least common was incorrect dates. No resulting adverse events were recorded.

CONCLUSION

The prevalence of errors needs to be reduced to avoid serious adverse incidents. Computerised physician order entry systems are discussed as a potential solution.

QI efforts lead to success in VTE prophylaxis

VTE prophylaxis is not a complicated process. CPOE alerts can remind physicians to do risk assessments. Online "resource room" is available for staff education.

Effect of bar-code technology on the safety of medication administration

BACKGROUND

Serious medication errors are common in hospitals and often occur during order transcription or administration of medication. To help prevent such errors, technology has been developed to verify medications by incorporating bar-code verification technology within an electronic medication-administration system (bar-code eMAR).

METHODS

We conducted a before-and-after, quasi-experimental study in an academic medical center that was implementing the bar-code eMAR. We assessed rates of errors in order transcription and medication administration on units before and after implementation of the bar-code eMAR. Errors that involved early or late administration of medications were classified as timing errors and all others as nontiming errors. Two clinicians reviewed the errors to determine their potential to harm patients and classified those that could be harmful as potential adverse drug events.

RESULTS

We observed 14,041 medication administrations and reviewed 3082 order transcriptions. Observers noted 776 nontiming errors in medication administration on units that did not use the bar-code eMAR (an 11.5% error rate) versus 495 such errors on units that did use it (a 6.8% error rate)--a 41.4% relative reduction in errors (P<0.001). The rate of potential adverse drug events (other than those associated with timing errors) fell from 3.1% without the use of the bar-code eMAR to 1.6% with its use, representing a 50.8% relative reduction (P<0.001). The rate of timing errors in medication administration fell by 27.3% (P<0.001), but the rate of potential adverse drug events associated with timing errors did not change significantly. Transcription errors occurred at a rate of 6.1% on units that did not use the bar-code eMAR but were completely eliminated on units that did use it.

CONCLUSIONS

Use of the bar-code eMAR substantially reduced the rate of errors in order transcription and in medication administration as well as potential adverse drug events, although it did not eliminate such errors. Our data show that the bar-code eMAR is an important intervention to improve medication safety. (ClinicalTrials.gov number, NCT00243373.)

CPOE: a clear purpose plus top-notch technical support equals high physician adoption

As with any fundamental change, the transition to computerized physician order entry [CPOE] is not a risk-free endeavor, major questions hover around this facet of the arduous and controversial paper-to-electronic conversion currently preoccupying the healthcare industry: Could physician over-reliance on electronic prompts actually lead to an increase in some types of medical errors? Could automated workstations ultimately hinder safety and the delivery of quality care by diminishing face-to-face communication and nuanced discussions? In an ironic twist, could electronic solutions insidiously leach creativity, intuition and judgment from good medicine by keeping physicians tied to tools that consume their time but do not offer effective clinical decision support?

Electronic prescribing at the point of care: a time-motion study in the primary care setting

OBJECTIVE

To evaluate the impact of an ambulatory computerized provider order entry (CPOE ) system on the time efficiency of prescribers. Two primary aims were to compare prescribing time between (1) handwritten and electronic (e-) prescriptions and (2) e-prescriptions using differing hardware configurations.

DATA SOURCES/STUDY SETTING

Primary data on prescribers/staff were collected (2005-2007) at three primary care clinics in a community based, multispecialty health system.

STUDY DESIGN

This was a quasi-experimental, direct observation, time-motion study conducted in two phases. In phase 1 (n=69 subjects), each site used a unique combination of CPOE software/hardware (paper-based, desktops in prescriber offices or hallway workstations, or laptops). In phase 2 (n=77), all sites used CPOE software on desktops in examination rooms (at point of care).

DATA COLLECTION METHODS

Data were collected using TimerPro software on a Palm device.

PRINCIPAL FINDINGS

Average time to e-prescribe using CPOE in the examination room was 69 seconds/prescription-event (new/renewed combined)-25 seconds longer than to handwrite (99.5 percent confidence interval [CI] 12.38), and 24 seconds longer than to e-prescribe at offices/workstations (99.5 percent CI 8.39). Each calculates to 20 seconds longer per patient.

CONCLUSIONS

E-prescribing takes longer than handwriting. E-prescribing at the point of care takes longer than e-prescribing in offices/workstations. Improvements in safety and quality may be worth the investment of time.

Chronology of prescribing error during the hospital stay and prediction of pharmacist's alerts overriding: a prospective analysis

BACKGROUND

Drug prescribing errors are frequent in the hospital setting and pharmacists play an important role in detection of these errors. The objectives of this study are (1) to describe the drug prescribing errors rate during the patient's stay, (2) to find which characteristics for a prescribing error are the most predictive of their reproduction the next day despite pharmacist's alert (i.e. override the alert).

METHODS

We prospectively collected all medication order lines and prescribing errors during 18 days in 7 medical wards' using computerized physician order entry. We described and modelled the errors rate according to the chronology of hospital stay. We performed a classification and regression tree analysis to find which characteristics of alerts were predictive of their overriding (i.e. prescribing error repeated).

RESULTS

12 533 order lines were reviewed, 117 errors (errors rate 0.9%) were observed and 51% of these errors occurred on the first day of the hospital stay. The risk of a prescribing error decreased over time. 52% of the alerts were overridden (i.e error uncorrected by prescribers on the following day. Drug omissions were the most frequently taken into account by prescribers. The classification and regression tree analysis showed that overriding pharmacist's alerts is first related to the ward of the prescriber and then to either Anatomical Therapeutic Chemical class of the drug or the type of error.

CONCLUSIONS

Since 51% of prescribing errors occurred on the first day of stay, pharmacist should concentrate his analysis of drug prescriptions on this day. The difference of overriding behavior between wards and according drug Anatomical Therapeutic Chemical class or type of error could also guide the validation tasks and programming of electronic alerts.

Biomedical and health informatics for surgery

The application of biomedical and health informatics to surgery holds tremendous opportunities to enhance surgical care. Better use of information in surgical practice has the potential to streamline care, remove inefficiencies, and allow for improvements in surgical research. With greater EHR adoption, health care reform, and direct investment in HIT, an increasing opportunity exists for surgeons to access and use patient information more effectively. For this to happen, greater focus on the specific needs of surgeons is particularly important, alongside increasing the number of surgical informatics stakeholders.

Alerts in clinical information systems: building frameworks and prototypes

Alerts in Clinical Information Systems and CPOE are powerful tools for decision support. However, studies show that physicians override a large part of these alerts. Low specificity and high bandwidth of alerts lead to alert fatigue. Moreover, alerts seem to have usability issues as they are interrupting workflows and not always efficient to handle. This paper provides three different views on alerts: a system-based view, a human-computer interaction view and an organizational view. Based on this framework, we present a prototype of alert handling, which might ameliorate some of the problems with alerts.

Usability of clinician order entry systems in Singapore: an assessment of end-user satisfaction

OBJECTIVES

To gather end-user feedback and evaluate factors that influence end-user satisfaction of order entry systems used in the hospitals under National Healthcare Group, Singapore.

DESIGN

Questionnaires were sent to a randomly selected group of 100 doctors and nurses.

RESULTS & CONCLUSIONS

52 doctors and nurses responded to the survey. The users' satisfaction with the clinical systems was average. (Mean satisfaction score is 3.85 on a scale of 1 to 7). Users generally agree that the systems could help reduce patient care errors and improve delivery of quality care to patients. System reliability, intuitive navigational capabilities and ease of use are strongly and positively correlated with user satisfaction. System response time however, is found to be strongly but negatively correlated with user satisfaction with a correlation coefficient of -0.717 (p<0.001). These findings suggest that more efforts should be made to improve these aspects in order to improve user satisfaction. These elements should also form important considerations in all future clinical systems development.

Impact of a restrictive antimicrobial policy on the process and timing of antimicrobial administration

BACKGROUND

In-hospital antimicrobial approval policies are designed to curb the indiscriminant use of antimicrobials. These policies usually require written forms and/or direct requests to an Infectious Disease specialist (or surrogate) prior to release of the antimicrobial. We hypothesized that the approval processes at our institution results in delayed antimicrobial administration.

METHODS

We performed a retrospective cohort study examining the time from order to administration for 25 different antimicrobials ordered "stat." Antimicrobials were classified as restricted (required approval) or unrestricted. We compared these 2 classes to each other both during the daytime (8 AM to 10 PM), when approval is required for restricted antimicrobials, and at night when the first dose of all antimicrobials is exempted. We defined a delay in administration when the medication was given >1 hour from time of order. We separately examined delays of >2 hours.

RESULTS

A higher percentage of >1-hour delays occurred when the antimicrobial was restricted (odds ratio [OR] = 1.49; 95% confidence interval [CI] = 1.23-1.82). Similar results were seen for >2-hour delays (OR = 1.78; 95% CI = 1.39-2.21). During the exempt-from-restriction time period (10 PM to 8 AM), there was no difference between these 2 classes of antimicrobials. Results were unchanged by adjustment for service (medicine vs. surgery vs. other), patient characteristics (age, sex, race), or by weekday vs. weekend.

CONCLUSIONS

Statistically significant delays in stat antimicrobial administration occur in our institution when antimicrobials require preapproval. These findings illustrate the importance of considering clinical efficiency when restrictions are put in place for time-sensitive therapies such as antimicrobials.

Can F-MTI semantic-mined drug codes be used for adverse drug events detection when no CPOE is available?

BACKGROUND

Adverse Drug Events (ADEs) endanger the patients. Their detection and prevention is essential to improve the patients' safety. In the absence of computerized physician order entry (CPOE), discharge summaries are the only source of information about the drugs prescribed during a hospitalization. The French Multierminology Indexer (F-MTI) can help to extract drug-related information from those records.

METHODS

In first and second validation steps, the performance of the F-MTI tool is evaluated to extract ICD10 and ATC codes from free-text documents. In third step, potential ADE detection rules are used and the confidences of those rules are compared in several hospitals: using a CPOE vs. using semantic mining of free-text documents, diagnoses and lab results being available in both cases.

RESULTS

The F-MTI tool is able to extract ATC codes from documents. Moreover, the evaluation shows coherent and comparable results between the hospitals with CPOEs and the hospital with drugs information extracted from the reports for potential ADE detection.

CONCLUSION

semantic mining using F-MTI can help to identify previous cases of potential ADEs in absence of CPOE.

Optimizing prevention of hospital-acquired venous thromboembolism (VTE): prospective validation of a VTE risk assessment model

BACKGROUND

Hospital-acquired (HA) venous thromboembolism (VTE) is a common source of morbidity/mortality. Prophylactic measures are underutilized. Available risk assessment models/protocols are not prospectively validated.

OBJECTIVES

Improve VTE prophylaxis, reduce HA VTE, and prospectively validate a VTE risk-assessment model.

DESIGN

Observational design.

SETTING

Academic medical center.

PATIENTS

Adult inpatients on medical/surgical services.

INTERVENTIONS

A simple VTE risk assessment linked to a menu of preferred VTE prophylaxis methods, embedded in order sets. Education, audit/feedback, and concurrent identification of nonadherence.

MEASUREMENTS

Randomly sampled inpatient audits determined the percent of patients with "adequate" VTE prevention. HA VTE cases were identified concurrently via digital imaging system. Interobserver agreement for VTE risk level and judgment of adequate prophylaxis were calculated from 150 random audits.

RESULTS

Interobserver agreement with 5 observers was high (kappa score for VTE risk level = 0.81, and for judgment of "adequate" prophylaxis = 0.90). The percent of patients on adequate prophylaxis improved each of the 3 years (58%, 78%, and 93%; P < 0.001) and reached 98% in the last 6 months of 2007; 361 cases of HA VTE occurred over 3 years. Significant reductions for the risk of HA VTE (risk ratio [RR] = 0.69; 95% confidence interval [CI] = 0.47-0.79) and preventable HA VTE (RR = 0.14; 95% CI = 0.06-0.31) occurred. We detected no increase in heparin-induced thrombocytopenia (HIT) or prophylaxis-related bleeding using administrative data/chart review.

CONCLUSIONS

We prospectively validated a VTE risk-assessment/prevention protocol by demonstrating ease of use, good interobserver agreement, and effectiveness. Improved VTE prophylaxis resulted in a substantial reduction in HA VTE.

E-prescribing: where health information and patient care intersect

Medication therapy management can play a key role in the rising costs of healthcare, as it relates to both workflow efficiency and patient safety. E-prescribing allows for prescribing the most medically appropriate and cost effective prescription at the point of care and transmitting the prescription electronically to the patient's choice of pharmacy. E-prescribing can helpeliminate medication errors, improve patient safety, andreduce costs by making the medication therapy management process more efficient.

How to select end user clinical data entry devices. Rush University Medical Center develops tool to identify the quantity of devices needed for the implementation of a new EMR and CPOE system

Selecting the right types and quantities of computers to support data entry to an inpatient Electronic Medical Record (EMR) can be challenging. In addition to software and hardware considerations, many other variables affect the decision including staffing levels, hospital workflows, and floor plans. Rush University Medical Center (RUMC) developed a tool to help identify the quantity of devices needed in a Patient Care Unit (PCU). RUMC successfully used the tool in selecting the quantity of devices needed for the implementation of a new EMR and Computerized Provider Order Entry (CPOE) system. This case study describes the use of the tool to determine quantities of PCU devices, the advantages and disadvantages of different types of computing devices for bedside documentation and areas that require special considerations in the selection of devices.

Development of an asthma management system in a pediatric emergency department

Asthma is the leading chronic childhood disease with exacerbations resulting in urgent and emergency care visits. Guidelines adherence improves patient care but is suboptimal. A computerized guideline system can help improve compliance through automatic initiation and reminders to increase adherence. We designed a computerized management system for asthma care in the pediatric emergency department (ED). The system will be evaluated in two phases. The first phase evaluates a computerized diagnostic prompt using the ED's existing asthma protocol. The second phase evaluates a computerized asthma management system including temporal reminder elements for scoring and medication orders. The system was developed in conjunction with the pediatric ED multidisciplinary care team. The computerized system is entirely automatic and a prospective evaluation of the diagnostic component is ongoing.

Cognitive analysis of decision support for antibiotic prescribing at the point of ordering in a neonatal intensive care unit

Computerized decision support systems have been used to help ensure safe medication prescribing. However, the acceptance of these types of decision support has been reported to be low. It has been suggested that decreased acceptance may be due to lack of clinical relevance. Additionally, cognitive fit between the user interface and clinical task may impact the response of clinicians as they interact with the system. In order to better understand clinician responses to such decision support, we used cognitive task analysis methods to evaluate clinical alerts for antibiotic prescribing in a neonatal intensive care unit. Two methods were used: 1) a cognitive walkthrough; and 2) usability testing with a 'think-aloud' protocol. Data were analyzed for impact on cognitive effort according to categories of cognitive distance. We found that responses to alerts may be context specific and that lack of screen cues often increases cognitive effort required to use a system.

Decision making and physician prescribing characteristics: a pilot study of Japanese physicians

The factors that affect physicians' prescribing remain unclear. Although previous reports suggest that prescription decisions are associated with various clinical situation, most of these studies analyzed simulated patient models rather than actual clinical practice. Here, we retrospectively analyzed actual cases of statin prescription for hyperlipidemia at Tohoku University Hospital between Apr 1, 2004 and Mar 31, 2008. Twelve physicians (6 cardiologists, 3 nephrologist, and 3 diabetologist) made decisions on whether to prescribe statins to 187 patients in 788 visits. As expected, cardiologists started prescribing statins at significantly lower serum total cholesterol levels than other specialists (221.7mg/dL vs. 244.7mg/dL, P<0.05). Interestingly, the total cholesterol levels that triggered prescribing differed significantly among cardiologists (p<0.05). These results suggested that prescription decisions differed not only among specialties but also among individuals.