In 1998, the Institute of Medicine s (IOM s) National

Transcription

1 Computer Physician Order Entry: Benefits, Costs, and Issues Gilad J. Kuperman, MD, PhD, and Richard F. Gibson, MD, PhD Improving Patient Care Several analyses have detected substantial quality problems throughout the health care system. Information technology has consistently been identified as an important component of any approach for improvement. Computerized physician order entry (CPOE) is a promising technology that allows physicians to enter orders into a computer instead of handwriting them. Because CPOE fundamentally changes the process, it can substantially decrease the overuse, underuse, and misuse of health care services. Studies have documented that CPOE can decrease costs, shorten length of stay, decrease medical errors, and improve compliance with several types of guidelines. The costs of CPOE are substantial both in terms of technology and organizational process analysis and redesign, system implementation, and user training and support. Computerized physician order entry is a relatively new technology, and there is no consensus on the best approaches to many of the challenges it presents. This technology can yield many significant benefits and is an important platform for future changes to the health care system. Organizational leaders must advocate for CPOE as a critical tool in improving health care quality. Ann Intern Med. 2003;139: For author affiliations, see end of text. In 1998, the Institute of Medicine s (IOM s) National Roundtable on Health Care Quality noted that the quality of health care in the United States has substantial problems and that the problems will be addressed only through profound changes to the current health care system (1). The recent IOM report (2), Crossing the Quality Chasm, contained an unsettling list of studies documenting the prevalence of poor-quality care in the United States. Another IOM report (3), To Err Is Human, focused specifically on the current state of patient safety and ways to improve it. All three of these reports recognized the potential for information technology to be an important agent for change. Computer physician order entry (CPOE) allows physicians to enter orders directly into a computer rather than handwriting them (4 6). Because CPOE introduces supportive technology into one of the key processes in medicine it has the potential to address many quality problems. Computer physician order entry affords a health care organization opportunities to standardize practice; incorporate clinical decision support into daily practice; improve interdepartmental communication; facilitate patient transfers; and capture data for management, research, and quality monitoring. It provides physicians and other clinicians with an environment that is more appropriate than a paper-based setting to the complexities of today s medicine. Patients also can have the comfort of knowing that sophisticated technologies are being applied to help ensure their safety. Because implementing CPOE is an expensive and complex project that touches almost all aspects of the health care operation (4), only a few health care organizations in the United States have implemented it (4, 7, 8). However, such stakeholders as business, government (9), and payers (10) are creating pressures and incentives for organizations to adopt CPOE. In particular, the Leapfrog Group (11), a coalition of the country s largest employers who are eager to acknowledge and reward health-related quality improvement efforts, has identified CPOE as one of three important leaps that organizations can take to substantially improve patient safety. Many health care organizations are moving forward with CPOE efforts. In 2002, 67% of respondents to a survey conducted at the Health Information Management and Systems Society (HIMSS) Annual Conference said that their organizations were planning to implement CPOE; 21% said that they were currently implementing it, and 29% said that they were actively discussing it (12). As health care organizations decide whether and when to tackle the hurdles of implementing CPOE, they should keep in the forefront of their minds the ways in which this technology can be used to improve health care quality and efficiency. We examine the potential benefits, costs, and other important issues associated with CPOE. BENEFITS OF CPOE The National Roundtable on Health Care Quality (1) recognized that problems in health care quality fall into three categories: the underuse, overuse, and misuse of health care services. The Roundtable noted that reducing the overuse and misuse of health services leads to an increase in health care quality while decreasing costs. Reducing underuse increases quality but may increase costs. Computerized physician order entry may affect all three categories of health care quality problems, as well as inefficiencies in the health care system. Table 1 lists studies that have evaluated the effect of CPOE on various aspects of health care. Several studies have shown that CPOE can reduce un- Improving Patient Care is a special section within Annals supported in part by the U.S. Department of Health and Human Services (HHS) Agency for Healthcare Research and Quality (AHRQ). The opinions expressed in this article are those of the authors and do not represent the position or endorsement of AHRQ or HHS American College of Physicians 31 Downloaded From: by a Penn State University Hershey User on 03/06/2016

2 Improving Patient Care Computer Physician Order Entry and Quality of Care Table 1. Studies Showing the Effect of Computer Physician Order Entry on Health Care Variables* Study (Reference) Outcome Category Outcome Variables Features of CPOE Relevant to Outcome Variable Setting Participants, n Design Length of Trial Dexter et al. (13) Teich et al. (14) Overhage et al. (15) Overhage et al. (16) Tierney et al. (17) Bates et al. (18) Bates et al. (19) Tierney et al. (20) Rates of use of preventive care measures Compliance with formulary and prophylactic heparin use guidelines; dosing appropriateness Compliance with drug monitoring and preventive care guidelines Rates of use of preventive care measures Laboratory test Laboratory test Laboratory test Laboratory test Increased rates for pneumococcal and influenza vaccine, prophylactic heparin, and aspirin at discharge Increased frequency of use of s H 2 -blocker of choice; increased rate of prophylactic heparin; decreased rates of excessively high dosing; increased appropriateness of frequency for use of ondansetron Overall, compliance with guidelines was greater in the intervention group (46.3% vs. 21.9%; P 0.001) No significant difference between control and intervention groups In the intervention group, physicians ordered 14% fewer tests (P 0.005) and charges for tests were 13% lower (both P 0.05) 4.5% fewer laboratory tests ordered in the intervention group (not statistically significant); minimal difference for radiologic tests Test cancelled in 70% of 939 reminders; effect intervention decreased because many tests were sent to the laboratory without computer orders The number of tests decreased significantly in both groups, but more in the intervention group (16.8% in the intervention group and 10.9% in the control group) Reminders indicating that patient is eligible for preventive care measures Reminder to change H 2 -blocker order to comply with s formulary; reminder to order prophylactic heparin when patient is placed on bed rest; medication dosing and frequency guidance with pick lists 87 different reminders designed to decrease errors of omission Reminders indicating that patient is eligible for preventive care measures On-screen display of the charge for the outpatient diagnostic test being ordered and the total charge of tests for that patient for that day On-screen display of the charges for laboratory and radiologic tests at the time of computer Computerized reminders that a test seemed to be redundant (that is, the same test had already been ordered recently) On-screen display of previous results when1of8 diagnostic tests was ordered General medicine inpatients at a teaching General medicine inpatients at a teaching General medicine inpatients at a teaching Academic primary care medical practice Academic primary care medical practice 6731 patients; 202 physicians Specific numbers of patients and physicians not given 2181 inpatients; 87 physicians 1929 patients; 78 housestaff physicians 8392 patients; 121 physicians (22 faculty and 99 housestaff) 7080 patients for study of clinical laboratory tests; patients for study of radiologic tests Time-series patients 5946 patients; 111 physicians (97 housestaff and 14 faculty) 18 mo 4-wk study periods before and after interventions 30 wk 6mo 26 wk 4 mo for clinical laboratory tests; 7 mo for radiologic tests 15 wk 16 wk Continued on following page 32 1 July 2003 Annals of Internal Medicine Volume 139 Number 1

3 Computer Physician Order Entry and Quality of Care Improving Patient Care Table 1 Continued Study (Reference) Outcome Category Outcome Variables Features of CPOE Relevant to Outcome Variable Setting Participants, n Design Length of Trial Tierney et al. (21) Sanders and Miller (22) Harpole et al. (23) Bates et al. (24) Bates et al. (25) Shojania et al. (26) Evans et al. (27) Laboratory test Radiologic test Radiologic test Medication errors Medication errors Vancomycin patterns Antibiotic patterns, adverse drug events, and costs Charges for study tests were 8.8% lower in the intervention group (P 0.05) 60% agreement with recommendations; increase in use of brain MRI without contrast Cancellation rates in response to critiques were very low: 3% in phase 1 and 4% in phase 2; users accepted suggestions for alternative studies more often: 38% in phase 1 and 55% in phase 2 55% reduction in serious medication errors (P 0.01); decrease in preventable adverse drug events of 17% (P 0.2) Overall, 81% reduction in nonmissing dose medication errors; much of the reduction was achieved with minimal decision support Intervention physicians wrote 32% fewer orders for vancomycin (P 0.04); duration of vancomycin ordered by intervention physicians was 36% lower than that of control physicians (P 0.05) Significant decreases in adverse drug events due to antibiotics (4 vs. 28; P 0.02), antibiotic susceptibility mismatches (12 vs. 206; P 0.01), excess drug dosages, antibiotic costs, and total costs On-screen display of probability that the result would be positive when 1 of 8 diagnostic tests was ordered (based on locally created predictive models) Decision support screen to ensure compliance with guidelines for use of brain MRI Automated critiques presented at time of if, given the patient s symptoms, an abdominal radiograph (KUB) was unlikely to provide useful information or if an alternative study was more appropriate; phase 2 had more strongly worded messages Entire CPOE system Entire CPOE system, with increasing amounts of decision support Display at time of of guidelines for use of vancomycin Antibiotic advisor made patientspecific suggestions (about choice of antibiotic, dose, and frequency of use) by using laboratory and other computerized data Academic primary care medical practice Academic tertiary care Selected units at a teaching Selected units (2 general medical and 1 intensive care unit) at a teaching 12-bed shock trauma respiratory intensive care unit at a university teaching 9496 patients; 112 physicians Specific numbers of patients and physicians not given 681 patients; 67 providers in phase 1; 236 providers in phase 2 Time-series Prospective cohort 6mo 9-wk control period; 8 wk intervention period Two phases (9 wk and 19 wk, respectively) 6711 admissions Time-series 6-mo preintervention period; 9-mo postintervention period 1817 admissions over the 4 phases 1798 patients; 396 physicians Time-series 4 periods: 51, 68, 49, and 52 d over a 5-y period 9mo 1781 patients Time-series 2-year preintervention period; 1-year intervention period Continued on following page 1 July 2003 Annals of Internal Medicine Volume 139 Number 1 33

4 Improving Patient Care Computer Physician Order Entry and Quality of Care Table 1 Continued Study (Reference) Outcome Category Outcome Variables Features of CPOE Relevant to Outcome Variable Setting Participants, n Design Length of Trial Chertow et al. (28) Mekhjian et al. (29) Tierney et al. (30) Dosing appropriateness Multiple process variables Total charges and length of stay Appropriateness of prescriptions greater in the intervention group (67% vs. 54%; P 0.001) Significant decreases in turnaround times, elimination of transcription errors, improvements in order countersignature, and decrease in length of stay at 1 of the 2 settings Charges in the intervention group were $887 (12.7%) lower than in the control group; mean length of stay was 0.89 day shorter (P 0.11) Dosing decision support for patients with renal dysfunction A CPOE system and an electronic medication administration record Entire CPOE system Inpatients at 2 academic medical centers General medicine inpatients at a teaching patients Time-series 4 alternating 2-mo periods Specific numbers of patients and physicians not given 5219 patients; 68 physicians Time-series mo for each period 18 mo * CPOE computer physician order entry; MRI magnetic resonance imaging. deruse. Dexter and colleagues (13) found that, for eligible patients, reminders in CPOE systems increased rates for prophylactic aspirin for coronary artery disease, pneumococcal vaccine, influenza vaccine, and prophylactic subcutaneous heparin. Teich and colleagues (14) also found an increase in the use of prophylactic subcutaneous heparin with CPOE. Overhage and colleagues (15) demonstrated that compliance with the monitoring of drug levels doubled when automated reminders were implemented. The effect of CPOE on care may depend greatly on the subtle details of the application design. In one evaluation, a decision support feature did not increase compliance with guidelines (16); however, success was realized after the design of the feature was enhanced (13). Inappropriate admissions accounted for much of the overuse documented by the IOM report (2); CPOE applications may not be able to affect this problem. However, overuse of diagnostic procedures (31, 32) and antibiotics (33) has been well documented and can be addressed by CPOE. As part of the interaction with the user, CPOE can present cost data (17, 18), previous results (19, 20), and information about the likelihood of finding an abnormal result (21, 34), all of which have been shown to reduce the overuse of diagnostic tests. In one, the Centers for Disease Control and Prevention s guidelines for vancomycin use were embedded in a CPOE application; 28% (P 0.02) fewer patients were prescribed vancomycin or had their vancomycin order renewed as compared with CPOE without the automated guidelines (26). Of two CPOEbased interventions designed to reduce the use of radiology examinations, one was successful (22), while another had minimal impact (23). In summary, more study is needed to understand which quality problems are best suited to remedy by computerized intervention and which system designs are most successful in achieving behavioral change. Misuse of medications is an important concern that leads to various safety problems, including serious medication errors and preventable adverse drug events, as highlighted in the IOM s recent report (3). Two studies that examined the root cause of serious medication errors (35, 36) showed that about half were errors such as an incorrect choice of medication for the patient s condition, an incorrect dose of the medication given the patient s physiologic state (for example, renal insufficiency or advanced age), and prescribing a medication to which the patient was known to be allergic. Computer physician order entry systems have the potential to address these problems, and studies have shown that CPOE systems can reduce medication errors (24, 25) and associated costs and injuries (27). The features of CPOE that promote the safe use of medications include patient-specific dosing suggestions (14, 28, 37), reminders to monitor drug levels (15), reminders to choose an appropriate drug (14, 15, 26, 27), checking for drug allergy and drug drug interactions, standardized order sets (that is, departmentally sanctioned predefined collections of orders that promote safety and efficiency, especially in complex situations), increased legibility, automated communication to ancillary departments, and easy access to patient data and reference information while (37). A CPOE application with an integrated antibiotic advisor has been shown to reduce the 34 1 July 2003 Annals of Internal Medicine Volume 139 Number 1

5 Computer Physician Order Entry and Quality of Care Improving Patient Care rates of excessive drug dosages and antibiotic-susceptibility mismatches (27). Additional medication-related benefits can be realized if CPOE is well integrated with other components of the medication process. The benefits of interfacing CPOE to a pharmacy application have been described (38), and interfacing a CPOE application and an electronic medical administration record at one institution eliminated transcription errors (29). Although the IOM report (3) focused on the misuse of medications in the inpatient setting, new data suggest that the magnitude of the problem is similar in the outpatient setting (39 41). Emerging technologies that automate prescribing and electronically link the provider, the pharmacy, and the patient may address many medication-related problems in the ambulatory environment. Of note, CPOE applications are complex and are intended for use in complex health care environments. Although CPOE has the potential to reduce many sources of error, organizations must take care as CPOE is implemented so that new kinds of errors will not be introduced (25, 42). Most evaluations of CPOE have examined its effect on specific components of the care process, such as the use of diagnostic testing or medication error rates. A few studies have examined the effect of CPOE on overall costs and length of stay. Evans and colleagues (27) demonstrated reduced costs ($ vs. $35 283; P 0.005) and length of stay (10.0 days vs days; P 0.003) in patients in an intensive care unit when the suggestions of a computerized antibiotic advisor incorporated into the process were followed. A randomized, at one (30) indicated that inpatient costs were 12.7% lower (P 0.02) and the length of stay was 0.89 day shorter (P 0.11) for patients on general medicine wards that used CPOE with decision support. In a time-series design, Mekhjian and colleagues (29) documented a decreased length of stay at one of two study s from 3.91 days to 3.71 days (P 0.02); the length of stay at the other in the study did not change, and overall costs at the two study s did not change. Researchers are studying ways that CPOE can help physicians plan a therapeutic regimen (43, 44). Such planning systems hold special promise for improving quality and safety; however, they involve complex modeling of diagnosis and therapeutics (45). To be successful, planning programs must be integrated tightly into the workflow of the application and must not require excessive additional data entry by the physician. To date, attempts at integrating complex clinical guidelines into CPOE have met with limited success (46, 47), largely because published guidelines, although suitable in printed form for use by practicing clinicians, have been difficult to translate into explicit workflow-oriented algorithms that can be embedded in clinical information systems. In addition to improving the quality of care, CPOE applications have been documented to improve efficiency. Communications to nursing and consulting services and such ancillary departments as the laboratory, radiology, pharmacy, and dietary departments can be instantaneous and automated. Although the result can be a savings in time (29, 48), care needs to be taken because it is possible to introduce error when constructing such automated interfaces (49). COSTS OF CPOE Whereas the benefits of CPOE have been well documented, the associated costs and implementation issues are less well described. Such issues may seem daunting, even to organizations experienced with complex clinical information systems. The fact that CPOE requires an analysis and redesign of important clinical processes, as well as an implementation of technology, complicates the matter. We consider three major areas of costs associated with CPOE: technical costs, costs of process redesign, and cost of implementation and support. The technical costs of CPOE include hardware, software, technical support, and integration with existing systems. Computerized physician order entry requires the to have a robust information infrastructure. The computer network must extend to every location and be fast, secure, and reliable. The number of workstations must be sufficient. Many institutions that implement CPOE use a combination of desktop personal computers and wireless mobile laptop devices. Many organizations are considering the use of wireless handheld computers, but such devices are not yet in widespread use. The organization needs to enable physicians to interact with CPOE from such off-site locations as their homes and offices. The needs a well-functioning help desk so that technical problems are resolved rapidly and do not interfere with clinical work. Electronic interfaces must be developed between CPOE and other clinically relevant applications, such as patient registration, laboratory, pharmacy, radiology, and nursing documentation. The financial costs of a CPOE initiative vary widely depending on the size of the institution and the existing hardware and software systems. The costs of implementation of CPOE at a 500-bed, in the absence of a need for major network upgrades, have been estimated to be almost $ , with ongoing annual maintenance costs of $ (50). These estimates include hardware, software, network, workstation, information systems personnel, and non information systems personnel (for example, physician leadership) (50). Smaller s that have an existing satisfactory platform of well-integrated administrative, ancillary (laboratory, pharmacy, and radiology departments), and clinical (for example, nursing documentation and results review) applications would face lower costs. Larger s, academic medical centers, and integrated delivery networks that require substantial modifications to their network infrastructure, new server platforms 1 July 2003 Annals of Internal Medicine Volume 139 Number 1 35

6 Improving Patient Care Computer Physician Order Entry and Quality of Care Table 2. Studies Documenting Clinician Satisfaction with and Perception of Computer Physician Order Entry Systems* Study (Reference) Outcome Category Outcome Variables Features of CPOE Relevant to Outcome Variable Setting Participants Design Murff and Kannry (52) Lee et al. (53) Weiner et al. (54) Satisfaction with CPOE Satisfaction with CPOE Satisfaction with or perception of CPOE Overall satisfaction: 3.67 (scale, 0 to 9) for a commercial system and 7.21 for VA system; strongest correlate with satisfaction was ability to perform tasks in straightforward manner On scale of 1 to 7: medical house officers, 5.55; surgical house officers, 4.45; and nurses, 4.84; correlates of satisfaction were impressions of effect on productivity, ease of use, and speed Only 29% to 34% of physicians felt quality was better with CPOE; significant percentage of physicians felt that CPOE decreased time with patients; 34% to 42% reported high satisfaction Entire CPOE application Entire CPOE application Entire CPOE application Commercially available inpatient CPOE application and VA CPOE application Academic medical center using internally developed application Academic medical center using modified commercial application 144 house officers for commercially available system; 132 house officers for VA system 200 housestaff physicians; 200 nurses 271 house officers, attending physicians, fellows, and nurses Survey; 63% and 64% response rates for commercial and VA systems, respectively Survey; 56%, 32%, and 47% response rate from medical housestaff, surgical housestaff, and nurses, respectively Survey; 85% response rate overall * CPOE computer physician order entry; VA Veterans Affairs. to support CPOE, extensive integration of CPOE with existing systems, large-scale deployments of new clinical workstations, and large-scale training efforts might be faced with project costs in the tens of millions of dollars. Computerized physician order entry affects the work of several constituencies and several key processes. Most evident, physicians must enter their orders into the computer instead of writing them on paper. Physicians often resist this new way of working (51) and may recoil at the constraints of computer compared with paper. However, if the intended benefits are well conveyed, the system is well designed, and training is sufficient, acceptance of the applications can be very good (Table 2) (52, 53, 55). The ability to execute tasks in a straight-forward manner (52) and the response time of the application (53) have been correlates of satisfaction. Different CPOE applications have different levels of user satisfaction (52, 54). More research is needed to determine the factors that lead to high physician satisfaction with CPOE. Other constituencies affected by CPOE include nursing staff, ward assistants, pharmacists, and staff in the laboratory and radiology departments. Important processes that are affected by the adoption of CPOE include the medication management process and coordination of diagnostic procedures. The organization must form multidisciplinary teams to understand how these processes currently are carried out and how they will take place once CPOE has been implemented. The requisite analysis and planning are time consuming and require careful forethought but are critical to the success of CPOE implementation. Of note, the analysis and planning phase can be more complex than the technical implementation of the hardware and software. In addition, the analysis and redesign offer opportunities for process improvement. Such process improvements are an important and desired outcome of CPOE implementation; however, too much change at one time can be difficult for an organization to absorb. The software license fee represents a small part of the total cost of a CPOE project (6, 50). A large portion of the cost is dedicated to training and support activities. Clinicians need to be trained before implementation. Physicians and nurses usually can be trained in several hours; however, substantial support (sometimes described as help at the elbow ) is needed for the clinician users 24 hours a day for up to 2 weeks after implementation. When CPOE was implemented at the Brigham and Women s Hospital, Boston, Massachusetts, 1 support person was available per shift for every 30 acute-care beds and every 10 intensive care unit beds (56). Nurse staffing must allow for adequate clinical coverage during the training phase. Emerging computer-based training approaches have the potential to decrease training and support costs July 2003 Annals of Internal Medicine Volume 139 Number 1

7 Computer Physician Order Entry and Quality of Care Improving Patient Care OTHER ISSUES TO CONSIDER Several other important issues must be considered before an organization embarks on a CPOE implementation. 1. Computerized physician order entry is a complex undertaking and should not be the first computerized clinical system attempted by an organization. A CPOE application is more likely to be accepted if the existing clinical systems are well received. Organizations considering CPOE should have in place clinical systems, such as a comprehensive results review application and online access to electronic textbooks, the medical literature, and guidelines. 2. The organization must address the issues of detailed workflow and unintended consequences that accompany any new system installation. Implementing CPOE is a large enough project that the organization should be wary of doing other major administrative or clinical information system projects concurrently. 3. The return on investment for a CPOE project may be difficult to calculate because baseline costs of key processes are hard to determine; several benefits are not easily amenable to measurement (for example, improved interdepartmental communication and strategic positioning); and many organizations do not currently measure rates of medication errors and adverse drug events. Although measurable improvement may be detected in well-defined areas, such as the use of expensive diagnostic and therapeutic procedures and compliance with specific guidelines (15), CPOE should be viewed as supportive technology for such organizational initiatives as quality improvement, patient safety, and cost reduction. In addition, it is important that CPOE be viewed as part of an organizational strategy to achieve the previously mentioned objectives rather than as an information technology initiative. 4. Many physicians express concern that with CPOE takes longer than with paper. Some studies have confirmed this aspect of CPOE (Table 3) (57, 58), but at least one study has shown that electronic can be time-neutral compared with a paper process (59). Features of CPOE that can reduce the time burden to physicians include the use of predefined collections of orders for complex conditions (for example, initial management of the patient after bypass graft surgery), access to CPOE from locations other than the or office, adequate training, easy access to patient and reference data, and progressive familiarity with the application. Continual system refinement can also improve efficiency over time. The presence of alerts and reminders that prevent errors and in an information-rich environment may also make computerized a satisfying experience. 5. Although CPOE applications have existed since the 1970s, an organization embarking on CPOE today has to create its own road map to success. Vendor offerings are evolving rapidly (6), and purchasers must take care to understand the details of the software. Only a few institutions have described their approaches to a successful implementation (50, 60 62). More research is needed to create and evaluate models of CPOE implementation and to understand the specific challenges that exist for institutions of different sizes and different staffing models. Implementation failures have been reported (61), and little has been written about the use of CPOE in organizations with private medical staffs. Recently, a large organization had to Table 3. Studies of Effect of Computer Physician Order Entry on Physician Time Spent Ordering* Study (Reference) Outcome Category Outcome Variables Features of CPOE Relevant to Outcome Variable Setting Participants Design Length of Trial Shu et al. (57) Bates et al. (58) Overhage et al. (59) Effect of CPOE on time spent Effect of CPOE on time spent Effect of CPOE on time spent Percentage of time spent increased from 2.1% to 9.0%; 2.0% time recovered, so 5.0% net increase in time spent with CPOE Percentage of time spent increased from 5.3% to 10.5%; 2.7% of time recovered on medical service Overall, 6.2% increase in time spent (not statistically significant); experienced users were timeneutral with paperbased Entire CPOE system Entire CPOE system CPOE system as part of an outpatient electronic medical record Medical service at a teaching Medical and surgical services at a teaching 11 community health care and commercial practices 43 interns pre-cpoe; 29 interns post-cpoe 29 house officers pre-cpoe; 33 house officers post-cpoe Time-series Time-series 34 physicians Two 2-mo periods Not stated 18 mo * CPOE computer physician order entry. 1 July 2003 Annals of Internal Medicine Volume 139 Number 1 37

8 Improving Patient Care Computer Physician Order Entry and Quality of Care suspend its CPOE implementation at least in part because of complaints from the private staff (63). The organization must plan carefully and pay attention to detail to maximize the likelihood that expected benefits are realized. 6. Computerized physician order entry offers the institution the chance to implement automated decision support as part of the process (6). Decision support can be as simple as offering a pick list to reduce data entry error (for example, offering a choice list of reasonable doses for a given medication) or offering predefined order sets to standardize care. Examples of more complex decision support include alerts for drug drug and drug allergy interactions (37) and advice about antibiotic selection (26, 27). Although many years may be needed for an organization to evolve a sophisticated set of clinical decision support features, even a basic CPOE application can reduce medication error rates and confer other benefits (25). 7. Some organizations require all physicians to enter all orders into a CPOE system. At other organizations, the use of CPOE is encouraged but remains optional. At least one organization has reported good adoption of CPOE when use was optional (64). Obviously, the benefits realized will be proportional to the rate with which CPOE is adopted. In deciding whether to make CPOE optional or mandatory, the organization must consider the extent of the quality problems it is trying to fix and how confident it is that it can support physicians through the change process. No organization has reported that implementing CPOE is easy (61). However, the difficulty does not last forever, and once the pain of change has been tolerated, the organization has a powerful tool for the improvement of care. CONCLUSION Health care is in need of fundamental change. Quality problems in medicine are caused not by inadequate knowledge but by the current health care system s inability to consistently and accurately apply that knowledge. Computerized physician order entry can improve quality by standardizing processes and providing physicians guidance as they care for patients. When an organization decides to implement CPOE, it should plan well and be well girded for the experience. Institutions that have a vision of highquality care should include CPOE among their initiatives. From Partners HealthCare System and Harvard Medical School, Boston, Massachusetts, and the Providence Health System, Portland, Oregon. Acknowledgments: The authors thank Joan Ash, PhD, who organized a conference in 2001 to discuss factors that are important in the implementation of CPOE and stimulated the authors to consider the ideas presented in this paper. They also thank Anne Kittler for assistance in preparing the manuscript. Potential Financial Conflicts of Interest: None disclosed. Requests for Single Reprints: Gilad J. Kuperman, MD, PhD, Partners HealthCare System, 93 Worcester Street, 2nd Floor, Wellesley, MA 02481; , gkuperman@partners.org. Current author addresses are available at References 1. Chassin MR, Galvin RW. The urgent need to improve health care quality. Institute of Medicine National Roundtable on Health Care Quality. JAMA. 1998;280: [PMID: ] 2. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington: National Academy Pr; Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Pr; American Hospital Association. AHA Guide to Computerized Order Entry Applications. Washington, DC: American Hospital Association; Sittig DF, Stead WW. Computer-based physician order entry: the state of the art. J Am Med Inform Assoc. 1994;1: [PMID: ] 6. Metzger J, Turisco F. Computerized Order Entry: A Look at the Vendor Marketplace and Getting Started. Oakland CA: California Healthcare Foundation and First Consulting Group; December Ash JS, Gorman PN, Hersh WR. Physician order entry in U.S. s. Proc AMIA Symp. 1998: [PMID: ] 8. A Primer on Physician Order Entry. Oakland, CA: California Healthcare Foundation; California Senate Bill 1875, session; chaptered 9/28/2000. Accessed at on 12 May Empire says incentives spreading Leapfrog standards. Accessed at healthcare.com/article.cms?articleld on 12 May The Leapfrog Group for Patient Safety: Rewarding Higher Standards. Accessed at on 12 May HIMSS (Healthcare Information and Management Systems Society) 2002 Hot Topic Survey. Accessed at on 6 May Dexter PR, Perkins S, Overhage JM, Maharry K, Kohler RB, McDonald CJ. A computerized reminder system to increase the use of preventive care for ized patients. N Engl J Med. 2001;345: [PMID: ] 14. Teich JM, Merchia PR, Schmiz JL, Kuperman GJ, Spurr CD, Bates DW. Effects of computerized physician order entry on prescribing practices. Arch Intern Med. 2000;160: [PMID: ] 15. Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A randomized of corollary orders to prevent errors of omission. J Am Med Inform Assoc. 1997;4: [PMID: ] 16. Overhage JM, Tierney WM, McDonald CJ. Computer reminders to implement preventive care guidelines for ized patients. Arch Intern Med. 1996; 156: [PMID: ] 17. Tierney WM, Miller ME, McDonald CJ. The effect on test of informing physicians of the charges for outpatient diagnostic tests. N Engl J Med. 1990;322: [PMID: ] 18. Bates DW, Kuperman GJ, Jha A, Teich JM, Orav EJ, Ma luf N, et al. Does the computerized display of charges affect inpatient ancillary test utilization? Arch Intern Med. 1997;157: [PMID: ] 19. Bates DW, Kuperman GJ, Rittenberg E, Teich JM, Fiskio J, Ma luf N, et al. A randomized of a computer-based intervention to reduce utilization of redundant laboratory tests. Am J Med. 1999;106: [PMID: ] 20. Tierney WM, McDonald CJ, Martin DK, Rogers MP. Computerized display of past test results. Effect on outpatient testing. Ann Intern Med. 1987;107: [PMID: ] 21. Tierney WM, McDonald CJ, Hui SL, Martin DK. Computer predictions of abnormal test results. Effects on outpatient testing. JAMA. 1988;259: [PMID: ] 22. Sanders DL, Miller RA. The effects on clinician patterns of a computerized decision support system for neuroradiology imaging studies. Proc AMIA Symp. 2001: [PMID: ] 23. Harpole LH, Khorasani R, Fiskio J, Kuperman GJ, Bates DW. Automated evidence-based critiquing of orders for abdominal radiographs: impact on utilization and appropriateness. J Am Med Inform Assoc. 1997;4: [PMID: ] 24. Bates DW, Leape LL, Cullen DJ, Laird N, Petersen LA, Teich JM, et al. Effect of computerized physician order entry and a team intervention on preven July 2003 Annals of Internal Medicine Volume 139 Number 1

9 Computer Physician Order Entry and Quality of Care Improving Patient Care tion of serious medication errors. JAMA. 1998;280: [PMID: ] 25. Bates DW, Teich JM, Lee J, Seger D, Kuperman GJ, Ma Luf N, et al. The impact of computerized physician order entry on medication error prevention. J Am Med Inform Assoc. 1999;6: [PMID: ] 26. Shojania KG, Yokoe D, Platt R, Fiskio J, Ma luf N, Bates DW. Reducing vancomycin use utilizing a computer guideline: results of a randomized. J Am Med Inform Assoc. 1998;5: [PMID: ] 27. Evans RS, Pestotnik SL, Classen DC, Clemmer TP, Weaver LK, Orme JF Jr, et al. A computer-assisted management program for antibiotics and other antiinfective agents. N Engl J Med. 1998;338: [PMID: ] 28. Chertow GM, Lee J, Kuperman GJ, Burdick E, Horsky J, Seger DL, et al. Guided medication dosing for inpatients with renal insufficiency. JAMA. 2001; 286: [PMID: ] 29. Mekhjian HS, Kumar RR, Kuehn L, Bentley TD, Teater P, Thomas A, et al. Immediate benefits realized following implementation of physician order entry at an academic medical center. J Am Med Inform Assoc. 2002;9: [PMID: ] 30. Tierney WM, Miller ME, Overhage JM, McDonald CJ. Physician inpatient order writing on microcomputer workstations. Effects on resource utilization. JAMA. 1993;269: [PMID: ] 31. Canas F, Tanasijevic MJ, Ma luf N, Bates DW. Evaluating the appropriateness of digoxin level monitoring. Arch Intern Med. 1999;159: [PMID: ] 32. Schoenenberger RA, Tanasijevic MJ, Jha A, Bates DW. Appropriateness of antiepileptic drug level monitoring. JAMA. 1995;274: [PMID: ] 33. Gonzales R, Steiner JF, Lum A, Barrett PH Jr. Decreasing antibiotic use in ambulatory practice: impact of a multidimensional intervention on the treatment of uncomplicated acute bronchitis in adults. JAMA. 1999;281: [PMID: ] 34. Solomon DH, Shmerling RH, Schur PH, Lew R, Fiskio J, Bates DW. A computer based intervention to reduce unnecessary serologic testing. J Rheumatol. 1999;26: [PMID: ] 35. Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA. 1995;274: [PMID: ] 36. Leape LL, Bates DW, Cullen DJ, Cooper J, Demonaco HJ, Gallivan T, et al. Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA. 1995;274: [PMID: ] 37. Kuperman GJ, Teich JM, Gandhi TK, Bates DW. Patient safety and computerized medication at Brigham and Women s Hospital. Jt Comm J Qual Improv. 2001;27: [PMID: ] 38. Kuperman GJ, Cooley T, Tremblay J, Teich JM, Churchill W. Decision support for medication use in an inpatient physician order entry application and a pharmacy application. Medinfo. 1998;9 Pt 1: [PMID: ] 39. Gandhi TK, Burstin HR, Cook EF, Puopolo AL, Haas JS, Brennan TA, et al. Drug complications in outpatients. J Gen Intern Med. 2000;15: [PMID: ] 40. Gurwitz JH, Field TS, Harrold LR, Rothschild J, Debellis K, Seger AC, et al. Incidence and preventability of adverse drug events among older persons in the ambulatory setting. JAMA. 2003;289: [PMID: ] 41. Gandhi TK, Weingart SN, Borus J, Seger AC, Peterson J, Burdick E, et al. Adverse drug events in ambulatory care. N Engl J Med. 2003;348: [PMID: ] 42. McNutt RA, Abrams R, Arons DC. Patient safety efforts should focus on medical errors. JAMA. 2002;287: [PMID: ] 43. Chin HL, Wallace P. Embedding guidelines into direct physician order entry: simple methods, powerful results. Proc AMIA Symp. 1999: [PMID: ] 44. Hickam DH, Shortliffe EH, Bischoff MB, Scott AC, Jacobs CD. The treatment advice of a computer-based cancer chemotherapy protocol advisor. Ann Intern Med. 1985;103: [PMID: ] 45. Kassirer JP. Diagnostic reasoning. Ann Intern Med. 1989;110: [PMID: ] 46. Tierney WM, Overhage JM, Takesue BY, Harris LE, Murray MD, Vargo DL, et al. Computerizing guidelines to improve care and patient outcomes: the example of heart failure. J Am Med Inform Assoc. 1995;2: [PMID: ] 47. Tierney WM, Overhage JM, McDonald CJ. Computerizing guidelines: factors for success. Proc AMIA Annu Fall Symp. 1996: [PMID: ] 48. Lehman ML, Brill JH, Skarulis PC, Keller D, Lee C. Physician Order Entry impact on drug turn-around times. Proc AMIA Symp. 2001: [PMID: ] 49. Valenstein P, Meier F. Outpatient order accuracy. A College of American Pathologists Q-Probes study of requisition order entry accuracy in 660 institutions. Arch Pathol Lab Med. 1999;123: [PMID: ] 50. Computerized Physician Order Entry: Costs, Benefits, and Challenges A Case Study Approach. Long Beach, CA: First Consulting Group; January Massaro TA. Introducing physician order entry at a major academic medical center: I. Impact on organizational culture and behavior. Acad Med. 1993;68: [PMID: ] 52. Murff HJ, Kannry J. Physician satisfaction with two order entry systems. J Am Med Inform Assoc. 2001;8: [PMID: ] 53. Lee F, Teich JM, Spurr CD, Bates DW. Implementation of physician order entry: user satisfaction and self-reported usage patterns. J Am Med Inform Assoc. 1996;3: [PMID: ] 54. Weiner M, Gress T, Thiemann DR, Jenckes M, Reel SL, Mandell SF, et al. Contrasting views of physicians and nurses about an inpatient computer-based provider order-entry system. J Am Med Inform Assoc. 1999;6: [PMID: ] 55. Tierney WM, Overhage JM, McDonald CJ, Wolinsky FD. Medical students and housestaff s opinions of computerized order-writing. Acad Med. 1994; 69: [PMID: ] 56. Teich JM, Glaser JP, Beckley RF, Aranow M, Bates DW, Kuperman GJ, et al. The Brigham integrated computing system (BICS): advanced clinical systems in an academic environment. Int J Med Inf. 1999;54: [PMID: ] 57. Shu K, Boyle D, Spurr C, Horsky J, Heiman H, O Connor P, et al. Comparison of time spent writing orders on paper with computerized physician order entry. Medinfo. 2001;10: [PMID: ] 58. Bates DW, Boyle DL, Teich JM. Impact of computerized physician order entry on physician time. Proc Annu Symp Comput Appl Med Care. 1994:996. [PMID: ] 59. Overhage JM, Perkins S, Tierney WM, McDonald CJ. Controlled of direct physician order entry: effects on physicians time utilization in ambulatory primary care internal medicine practices. J Am Med Inform Assoc. 2001;8: [PMID: ] 60. Ahmad A, Teater P, Bentley TD, Kuehn L, Kumar RR, Thomas A, et al. Key attributes of a successful physician order entry system implementation in a multi- environment. J Am Med Inform Assoc. 2002;9: [PMID: ] 61. Ash JS, Stavri PZ, Kuperman GJ. A consensus statement on considerations for a successful CPOE implementation. J Am Med Inform Assoc. 2003;10: [PMID: ] 62. Payne TH. The transition to automated practitioner order entry in a teaching : the VA Puget Sound experience. Proc AMIA Symp. 1999: [PMID: ] 63. Hospital Heeds Doctors, Suspends Use of Software. Los Angeles Times. 22 January Accessed at la-headlines-pe-california on 6 May Davis DC, Moriyama R, Tiwanak G, Morse L, Saito C. Clinical performance improvement with an advanced clinical information system at Queen s Medical Center. 5th Annual Nicholas E. Davis Award. Proceedings of the CPR Recognition Symposium. Washington, DC: Computer-Based Patient Record Institute; 1999: July 2003 Annals of Internal Medicine Volume 139 Number 1 39

10 Current Author Addresses: Dr. Kuperman: Partners HealthCare System, Inc., 93 Worcester Street, 2nd Floor, Wellesley, MA Dr. Gibson: Providence Health System, SW 68th Parkway, Portland, OR E-40 Annals of Internal Medicine Volume Number