American Journal of Health-System Pharmacy

doi: 10.1093/ajhp/64.suppl_9.24pmid: N/A

Learning objectives After studying these articles, the reader should be able to Explain the relationship between technological improvements and a culture of safety. List some ways pharmacists can play a leadership role in improving medication safety. Summarize some of the recommendations of the most recent IOM report. List recent medication safety technologies and their advantages. List some process improvements that impact medication safety. Propose some ways pharmacists can carve out more time for medication therapy management. Self-assessment questions Philip Schneider illuminates the concept of a culture of safety by comparing it with holding people accountable for assuming personal responsibilities for what they do in high-risk situations like Skydiving. Freeway driving. Scuba diving. Flying. Schneider notes that one of the surprising aspects of the IOM report is its focus on Centralized technology. Spontaneous reporting. Patient involvement. Software. Schneider observes that in a majority of hospitals, _____ are responsible for patient education programs. Pharmacists. Nurses. Physicians. Technicians. The second major recommendation in the IOM report is to reduce medication errors through Improved scheduling processes. Increased accountability. Building more verification steps into the process. Technology. Without a _____ , it is not possible to improve safety by purchasing such things as smart pumps. System of regulations. Certified IT specialist. Decentralized pharmacy. Culture of safety. Schneider points out that the most common step where errors that result in injury occur is in the _____ step. Prescribing. Transcription. Dispensing. Drug administration. The closer we get to _____ , the less likely an error is to be discovered. A reliance on technology. The patient. The end of a shift. A reliance on unit dose packaging. A pharmacy-coordinated unit dose and drug administration system documented an error rate of less than 10%. 5%. 3%. 1%. Schneider asserts that one way to assure that nurses witness the dose being administered is to Remove medications from their packaging before being administered to the patient. Adopt a BCMA system. Adopt spontaneous monitoring. Improve scheduling. The patient safety innovation that is being adopted at the slowest rate is CPOE systems. Smart pumps. BCMA systems. Decentralized automated dispensing. Schneider projects that at least _____of the hospitals will not meet the guideline of having a plan to implement CPOE by 2008. A quarter. A third. Half. Two thirds. Schneider observes that many of the new technologies save a great deal of time. True. False. The 2006 report, Preventing Medication Errors, shows that estimated very conservatively, there is approximately one medication error per patient per _____ . Day. Every three days. Week. Month. The IOM report suggests that access to comprehensive reference information is adequate but that pharmacists are not relying on it to a great enough extent. True. False. Bates calls it remarkable that there is still not A standard way of representing drugs. A system for receiving all prescriptions electronically. Financial incentives for doing things that improve medication safety. More spontaneous reporting of errors and near-misses. A 1998 study published in JAMA reported that the cost of each preventable ADE was about $2,500.00. $4,000.00. $6,000.00. $8,000.00. The 2001 “Nephros” study showed that in a control group for renal dosing, the prescribed dose was correct only _____ of the time. 38%. 42%. 54%. 65%. In contrast to the implementation of a commercial CPOE application at the University of Pittsburgh that was fully introduced over a six-day period and which was accompanied by an increased mortality rate, implementation of CPOE at Brigham and Women’s Hospital was introduced over a _____ period. Six-week. Three-month. Six-month. Three-year. In explaining the importance of safety with i.v. infusion systems, Bates points out that with “smart pumps,” administration errors leave no one between the _____ and the _____ to intercept the error. Nurse / patient. Pharmacist / nurse. Physician / patient. Physician / nurse. The study published in Critical Care Medicine in 2005 revealed that nurse overrides of important warnings and alerts with smart pumps were _____ extensive than investigators anticipated. Much more. More. Less. Much less. Bates argues that _____ holds “significant promise” to improve infusion safety and quality of care. Decision support. Comprehensive reference information. Centralized technology. Standardization. Bates cites a recent study on bar coding that showed the dispensing error rate fell by almost _____ after bar-code implementation. A quarter. A third. Half. Three quarters. The McLeod Regional Medical Center added _____ to its Medication Safety Committee. Accountants. Orderlies. Technicians. Patients. Nicol reports that in contrast to 70 reported errors per year, McLeod’s new non-punitive reporting system now averages more than _____ per month. 70. 100. 125. 150. Instead of reporting the number of ADEs per 1,000 doses, leaders at McLeod felt it would be more meaningful to clinicians if they converted this statistic to the number of patients per day, which in their case was 35 patients per day. Implementation of medication safety efforts reduced this number to less than _____ event(s) per day. 20. 10. 5. 1. Witkowski reports that improved medication safety efforts at Alliance Community Hospital have resulted in spending less than _____ of pharmacist time per dose of medication administered. 1 second. 5 seconds. 30 seconds. 1 minute. Pharmacists at Alliance Community Hospital now spend most of their time on Medication therapy management. Verifying the accuracy of newly filled orders and checking unit dose carts. Drug distribution. Technician training. Witkowski says that medications at Alliance are bar-code scanned only _____ times between the time they enter the building and the time they are administered to the patient. Three. Five. Nine. 17. Automation costs at Alliance now average _____ per withdrawal transaction. 26 cents. 34 cents. 53 cents. 62 cents. When a pharmacist at Alliance enters an order, the average amount of time required before that medication is available on the floor is only about 20 minutes. 10 minutes. 5 minutes. 3 minutes or less. AJHP continuing education AJHP CE Process Continuing education (CE) credit for this AJHP supplement is free for both members and nonmembers. Successful completion of a CE test is required to obtain CE credit. The CE test can only be taken online through ASHP’s CE Testing Center. After successful completion (score of 70% or better), you will be able to immediately print your own CE statement for your records. You will have two opportunities to pass the CE test, and you may stop and return to the test at any time before submitting your final answers. ASHP will keep a record of the credits you have earned from this and other CE activities, and you will be able to view your entire transcript through the CE Testing Center. Instructions for the CE Testing Center Go to http://ce.ashp.org and select “Enter CE Testing Center” ASHP members and registered customers: Type in either your 8-digit ASHP ID or your user name, and password. Click on “Register for Tests”. Follow instructions to select the desired supplement or, an easy way to search, click on “Edit”; select “Find on this Page”; type in a word or phrase in the title. Nonmembers: Click on the “Become a Registered User” link at the bottom of the page. Fill out the information requested and create your password. Go back to http://ce.ashp.org and select “Enter CE Testing Center” You will automatically be logged into the center. To view the complete list of available AJHP CE articles and supplements go to www.ashp.org, click on “Continuing Education” on the left menu, and then click on “AJHP CE Publications” under “Self- Study Programs”. While many of these are free to nonmembers, some are not. To have access to all the CE AJHP has to offer for free, consider becoming a member of ASHP. To find out more about the benefits of being an ASHP member, go to www.ashp.org and click on “Member Center” on the left menu. Supplement: Preventing medication errors: The Institute of Medicare report ACPE #: 204-000-07-005-H04P CE credit: 2.0 hours (0.2 CEUs) Expiration date: July 15, 2010 The American Society of Health- System Pharmacists is accredited by the Accreditation Council for Pharmacy Education as a provider of continuing pharmacy education Questions? Call ASHP Processing Center: 866-279-0681 (toll free) +1-240-646-7082 (international callers) Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.

Cohen, Michael, R.

doi: 10.2146/ajhp070189pmid: 17617507

When the Institute of Medicine (IOM) report, To Err is Human,1 was published in 2000, it led to considerable excitement and media coverage, dramatically increasing the nation’s focus on medical errors and their prevention. Soon after the report was published, I had the unique opportunity of standing with President Bill Clinton at the podium during a news conference in the White House Rose Garden when, in referring to the report, he said, “I think we should look at this as a very positive event in the progress of American health care…This is a good day for America, not only because of this report, but because of the response to this report.”2 No one can deny that the response to the To Err is Human report was to make medical errors a leading public health issue and lay the groundwork for improvements in patient safety. But has all this attention actually resulted in real improvement, particularly in the area of medication safety? A new landmark IOM report released in July 2006, Preventing Medication Errors, clearly indicates that we still have much work to do.3 The report, mandated by Congress in legislation that provides prescription drug coverage (Part D) in the Medicare Prescription Drug, Improvement, and Modernization Act (MMA), calls the frequency of medication errors and related injuries “a serious concern.” It suggests numerous error-prevention strategies that will require our undivided attention, such as improved pharmacy leaflets and medication-related information for consumers; better communication of clinical information about the patient to those who need it; collaboration among industry, FDA, and patient safety organizations to address problems with drug naming, labeling, and packaging; improved error reporting and dissemination of knowledge; support for more involvement by educators and oversight agencies such as state boards of pharmacy; and implementation of patient safety technologies. As a follow-up to publication of Preventing Medication Errors, Cardinal Health sponsored a “Dimensions in Leadership” program on December 3, 2006, at the ASHP Midyear Clinical Meeting to provide a forum to review the report’s findings and provide a means for attendees to share thoughts on related matters. The proceedings published here from Preventing Medication Errors—The IOM Report 2006 Symposium document the event. A distinguished panel of experts participated, including two members of the IOM Committee that coauthored the report, David Bates and Michael Cohen; two nationally known practitioners who have demonstrated leadership in implementing many of the error prevention strategies suggested in the report, Natasha Nicol and Paul Witkowski, who are also directors of pharmacy at a tertiary-care and a community hospital, respectively; and Philip Schneider, an educator who throughout his career has made medication safety issues a focus. The panel reviewed the report and also examined how safety strategies discussed in the report are being successfully employed, such as bedside scanning of medication doses, use of automated medication dispensing systems to control drug access and allow redeployment of clinical pharmacists, smart infusion pumps with drug libraries that prevent harm from accidental overinfusion, computerized prescriber order entry systems with decision support, and active surveillance systems to prevent adverse drug events. Incorporating and integrating such technologies across the entire medication management spectrum should be a strategic goal of every organization. Although we still have a great deal to accomplish, there are unquestionable signs that progress is being made. In the coming years, I look forward to working with you to advance the recommendations made in this latest IOM report to better serve the health needs of consumers and the practice needs of providers and, most importantly, to prevent medication errors. Your participation is crucial for success. Let’s make it another good day. Footnotes Based on the proceedings of a symposium held December 3, 2006, during the ASHP Midyear Clinical Meeting in Anaheim, CA, and supported by an educational grant from Cardinal Health. ISMP received an honorarium for participating in the symposium and writing this article. ISMP and Dr. Cohen have no financial relationships to report. Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.

Schneider, Philip, J.

doi: 10.2146/ajhp070188pmid: 17617508

Abstract Purpose. To summarize key points from the most recent Institute of Medicine (IOM) report, Preventing Medication Errors, and their relevance to health-system pharmacists. Summary. Creating a culture of safety is an important antecedent to making changes needed to reduce medication errors. The patient can play an important, but often unrealized, role in preventing medication errors. There are considerable opportunities to improve the application of technology to prevent medication errors. The National Hospital Pharmacy surveys conducted ty the American Society of Health- System Pharmacists have demonstrated the slow rate of adoption of technologies. Pharmacists can play a leadership role in improving both the patients’ role and the use of technology to improve medication-use safety. Conclusion. There are major opportunities for pharmacists to re-think how involved patients are in their care in the institutional setting, by seeing patients as building another check into the medication-use system and by recommitting ourselves to getting patients involved. There are also opportunities to help with patient-safety technology decisions. These extremely expensive technologies almost always involve people changing what they do and their implementation and use often involve other people besides pharmacists, yet the potential is great for new technologies to reduce medication errors. Errors, medication, Hospitals, Institute of Medicine, Patient information, Pharmacists, hospital, Pharmacy, institutional, hospital, Quality assurance, Risk management, Technology, Toxicity Most of you flew to this meeting on an airplane and, like me, ignored the fiight attendant asking you to pay attention and listen to a few safety warnings that were being offered. In spite of having listened, I had to be reminded to turn off my cell phone before departure and I had to be reminded to put my seat back in its upright position when we were landing. The point of this is that there are places with a culture of safety where people are held accountable for assuming personal responsibilities for what they do in high-risk situations like flying. Does that exist within healthcare? When patients are admitted in your hospital, do they receive the same kind of advice about the potential hazards that could happen to them that they might when they board an aircraft? The Institute of Medicine’s (IOM’s) report suggests that the answer to this question is “no.”1 Patient education One of the surprising aspects of the IOM report is its focus on patient involvement, particularly in the inpatient setting. What are the implications of recommendations about the role that hospitalized patients can and should play? How can a patient who is admitted into an acute care hospital—sick, maybe even unconscious—be expected to be more involved in his or her care? The IOM encourages consumers to ask what drugs they are being given. 1 Patients often view themselves as passive witnesses to the care that they receive in the hospital, rather than being involved with their care. Some of the recommendations made in Preventing Medication Errors encourage consumers to be involved more in their care.1 These recommendations include asking what drugs you are being given, not taking a drug without being told to, having a surrogate monitor if you cannot ask, asking about post-surgery consequences prior to surgery, and getting a complete medication list prior to discharge. If we think about the medication-use system in the hospital and intercepting adverse drug events (ADEs), the nurse is often viewed as the last step. But the patient is really the last step in the process. So, it makes sense to have patients more involved in the medications they receive. It adds an additional double-check to the medication-use system and is an important strategic opportunity for us to improve patient safety. Patient involvement in the hospital consists of medication histories, self-medication programs, medication administration records, and education programs. There is literature showing that pharmacists can take medication histories better than physicians and nurses, yet because of workforce issues pharmacists rarely perform medication histories anymore. 2 In fact, we no longer ask this question in the ASHP National Hospital Pharmacy surveys. Instead, we ask about medication reconciliation, because that has been identified as a practice that is important to identify changes in drug therapy when patients are transferred from one point of care to the next. Medication histories are coming back, probably because of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) standards. Self-medication programs were developed to teach patients how to take their medications before they were discharged.2 These types of programs have a history in healthcare, but the number of institutions that have them is now relatively small.2 There is an opportunity to get patients more involved in their care once again. One way is to provide medication administration records to patients so that they can follow along to be sure that the medicine that they are receiving is correct. Some clinicians are afraid to do this, because they are afraid that patients might catch them making a mistake. A better way to look at this is as building another double-check into the medication-use system. In many hospitals, entire departments have been devoted to patient education activities.2 Based on data from the ASHP National Hospital Pharmacy surveys, pharmacy often is not involved in these programs.2 Pharmacists may be members of a patient education committee or serve in an advisory capacity, but in the overwhelming majority of hospitals nurses are responsible for patient education programs.2 When pharmacy directors are asked what determines whether a patient gets counseled about medications, the most frequent reason is a physician order followed by a patient request.2 Some less frequent ways that patients are selected for patient education programs are focusing on specific targeted medicines, patients with a history of noncompliance, patients on newly prescribed medicines that they have not taken before, patients with certain diseases or taking medications with drug interactions, and patients who are taking a large number of medicines that might cause confusion. A very small number of institutions have a policy to counsel all patients about their medications. Most, however, use ways to select patients who might be most likely to benefit from patient education programs and perhaps target those medicines or those patients. One of the things to consider after reading this report is the extent to which pharmacists should be involved with patient education programs. Even for institutions that have such programs, patient education is not being provided for all patients. Technology The second major recommendation in the IOM report is to increase the use of technology to reduce medication errors.1 The ASHP National Hospital Pharmacy surveys have tracked the adoption of technologies such as computerized prescriber order entry (CPOE), bar-code medication administration (BCMA), and smart pumps and demonstrated the slowness of change.3 One of the important recommendations for healthcare providers is that a plan be in place to implement CPOE by 2008 and that this system be in place by 2010, with a caveat that it be well-designed.1 Decision support that has information about patients and their medicines is what makes electronic transfer of information safer. Monitoring for errors, finding out where errors are common, and trying to learn from those mistakes are important. It is not possible to improve safety simply by purchasing a technology without having a culture of safety. Many of the bad experiences with technology have been the result of not having a culture of safety, not addressing the technology/human interface well, or attempting to implement a system in too short of a time period.4 Failures happen at all steps in the medication-use system, such as prescribing (39%), transcription/ verification (12%), dispensing (11%), and nurse administration (38%).5 One of the ways to focus attention on reducing medication errors might be to think about where the mistakes are most common. Interceptability or the discoverability of an error might be another way to establish priorities. Failure mode effects analysis is based on three principles: how often does an event happen, how likely is it to go undiscovered, and how likely is it to cause serious injury if it happens? The closer we get to the patient, the less likely an error is to be discovered.6 Potential ADEs were intercepted upon physician ordering (48%), transcription/verification (33%), pharmacy dispensing (34%), and nurse administration (2%).6 One of the ways to increase discoverability at the point of care is to get the patients more involved. The other is to apply technology, such as bar-code medication administration (BCMA), as an additional check before a medication is administered to a patient. Technology adoption How quickly are these technologies being adopted? The adoption of CPOE systems has been a challenge—fewer than 10% of hospitals have implemented this technology (Figure 11).3 Figure 1. Open in new tabDownload slide Adoption of CPOE in U.S. hospitals.4 Figure 1. Open in new tabDownload slide Adoption of CPOE in U.S. hospitals.4 What about safe drug distribution systems? Virtually all hospitals have adopted unit dose and i.v. admixture systems.7 But now the emphasis has changed to improving the efficiency and even further adding to the accuracy of these systems, through either robotic centralized systems or automated dispensing cabinets that store medications closer to patients. Automated devices more accurately and efficiently compound sterile preparations in the pharmacy.8 For drug administration programs, the lowest error rates that have ever been recorded was with a pharmacy-coordinated unit dose and drug administration system.9 Error rates of less than 1% were documented, but this innovation was not widely adopted.9 More recently, BCMA systems have been recommended, and the FDA has mandated bar codes be printed on all unit dose packages.10 With i.v. medications, the current technology being recommended is the infusion pump with decision support (smart pump) that alerts or prevents nurses from administering doses that have the potential to harm patients.1 Errors can be minimized at the bedside. Safe practices at the bedside include: 1) verifying patient’s name by verbal questioning; 2) checking pre-medication orders; 3) checking the dose against the order itself; 4) removing medications from their packaging right before they are administered to the patient; and 5) witnessing the dose being administered (Figure 22). This does not happen all the time because nurses take shortcuts while administering medications. 11 One of the ways to keep these double-checks in the system is a BCMA system. The adoption of this technology has been a little bit faster than CPOE systems and is now more than 10% (Figure 3 3).3 Figure 3. Open in new tabDownload slide Rapid Adoption of Bar-Code Medication Administration.4 Figure 3. Open in new tabDownload slide Rapid Adoption of Bar-Code Medication Administration.4 Figure 2. Open in new tabDownload slide Increases in Safe Medication Administration Practices.4 Figure 2. Open in new tabDownload slide Increases in Safe Medication Administration Practices.4 Smart pumps came along much more recently than CPOE and BCMA systems.3 When implemented in a hospital, they are used throughout the hospital, rather than just in areas or for certain high-risk patients (Figure 44).3 About one third of hospitals have adopted the use of smart pumps as a patient safety technology (Figure 5 5).3 The uptake of smart pump technology appears to be happening at a little bit quicker rate than CPOE or BCMA systems.3 Figure 5. Open in new tabDownload slide Smart Pump are Similarly Diffused in all but the Smallest Hospitals.4 Figure 5. Open in new tabDownload slide Smart Pump are Similarly Diffused in all but the Smallest Hospitals.4 Figure 4. Open in new tabDownload slide Smart Pump Implementation Throughout Hospital.4 Figure 4. Open in new tabDownload slide Smart Pump Implementation Throughout Hospital.4 With all of these technologies, the challenge for hospitals is deciding which one to implement first. There are two ways to think about this. First, where in the medication-use system are mistakes most common? The answer is in the drug ordering and drug administration steps. Second, where is an event least likely to be intercepted? The answer is during the drug administration step. Directors of pharmacy can play a role in helping to decide what to do first. Rogers found that the diffusion of innovation is a slow process.12 It takes 15–20 years before an innovation is widely adopted, and different innovations are adopted at different rates. The information from the ASHP National Hospital Pharmacy surveys shows that some patient safety innovations are being adopted at a faster rate than others.3 Smart pumps are being adopted at the fastest rate, followed by BCMA systems. CPOE systems are being adopted at the slowest rate.3 None of these technologies is a bad idea—they are all recommended for improving patient safety. Reengineering the medicationuse system Another aspect of the diffusion of innovations is how the speed of adoption of a single innovation can be increased. One of the ways to increase the rate of change is to have either a published recommendation that becomes a standard of practice or an enforceable regulation. For example, medication reconciliation has been rapidly adopted, because JCAHO requires it.13 It remains to be seen to what extent the new IOM report will increase the speed of adoption of patient safety technologies. Not every innovation is sustainable. An example of an innovation that was not sustainable was the pharmacy-coordinated unit dose and drug administration program, despite the evidence that it was a safer system.13 Thus, there is a need to be discriminating in considering innovations and determining whether or not they will ever be widely adopted. Are CPOE systems an example of this? Is CPOE simply too hard to do? The recommendation from the IOM report acknowledges the difficulty, but it may be an innovation that does take 15–20 years to achieve wide-spread adoption.12 Trend toward decentralization One of the trends in pharmacy is the recognition that unit dose and i.v. admixture programs, which are centrally based and labor intensive, create waiting times and delays that sometimes compromise patient care and nurse satisfaction. There is a trend to make use of technologies to decentralize drug dispensing systems so that medications are closer to patients.14 There are still strong advocates for a centralized program, because there are fewer work-arounds and medications can be provided in ready-to-use form more often. Despite this view, decentralized automated dispensing technologies are being adopted more quickly than centralized technology. 14 When directors of pharmacy are asked, “What do you have now, and where would you like to be in your envisioned future?,” there is an interest in moving toward a more decentralized drug distribution system (Figure 66).14 There appears to be a fundamental shift in drug distribution in healthcare today, from centralized systems to decentralized systems.14 This has implications for nursing and pharmacy workload and time management. Directors of pharmacy were also asked for their opinions about plans to adopt new technologies in patient safety. They were asked if the technology was already in place, or whether there was a plan to implement it within a year, one to three years, or more than three years, or if there were no plans to implement it at all. It will be interesting to see whether those numbers change after this new IOM report has been published. At least half of the hospitals will not meet the guideline of having a plan to implement CPOE by 2008, unless priorities are changed as a result of the recommendation (Figure 77).15 Speeding the diffusion of innovations It is important to understand the diffusion of innovation, particularly the length of time that is required to adopt innovation. The IOM and JCAHO make good recommendations that are well thought out and important. The hard part is making the change. One role for leaders in pharmacy is to accept this, realize everything cannot be done at once, and help make the tough choices about what to do first. It is important to realize two things: not every change is the same relative to its uptake and its feasibility, and it takes more time than people might think to make a change. These are very important principles to keep in mind. Even though many of these technologies affect activities that are primarily the responsibility of other people, namely, physicians who prescribe and nurses who administer medications, pharmacists have the expertise and ability not only to help evaluate which of these technologies should be adopted first and in what order, but also to help design the decision support that is necessary for these to work. Rogers noted some of the ways to speed the diffusion of innovations.12 First, find sound innovations. Some institutions have had bad experiences, because they have purchased off-the-shelf technologies that might not have been configured in a way that really met the needs of the institution, or were not usable by the people that they were trying to help. “Role modeling” is one of the important strategies that can be used by thought leaders within an institution to drive behavior. Second, find the innovators, support them, and give them time to be trainers. This second step can be very important for the successful implementation of a technology that has the potential to improve patient safety but may not, if not introduced properly. Third, invest in early adopter activity and make success stories visible to the rest of the staff. For example, it might be useful to regularly tell the staff, “Look at the number of errors that we’ve prevented” or “Look at how many preventable events would have been prevented, if you would have used the decision support.” This is a very powerful tool for making the benefits of using a technology observable. Fourth, trust and enable reinvention. One example is refining decision support information. If nine out of ten alerts do not really help the staff, they will not pay much attention to them. This will slow or even stop the adoption of an innovation, unless the decision support information is continually refined based on the experience of the staff and the information provided to them focuses on things that really add benefit. Thus, fifth, create slack for change. There needs to be time, not only for reviewing experience and making improvement, but for learning how to use these systems, rather than a “big bang” theory of expecting a system to be used right “off the shelf.” Finally, lead by example. Conclusion There are major opportunities for pharmacists to rethink how involved patients are in their care in the institutional setting, by seeing patients as another check in the medication-use system, and by recommitting ourselves to getting patients involved. There are also opportunities to help with patient-safety technology decisions. These extremely expensive technologies almost always involve people changing what they do, and their implementation and use often involve other people besides pharmacists, yet the potential is great for new technologies to reduce medication errors. Figure 6. Open in new tabDownload slide Trend Toward Decentralization of Medications.7 Figure 6. Open in new tabDownload slide Trend Toward Decentralization of Medications.7 Figure 7. Open in new tabDownload slide Medication-Use Safety Technology Plans.8 Figure 7. Open in new tabDownload slide Medication-Use Safety Technology Plans.8 Footnotes Based on the proceedings of a symposium held December 3, 2006, during the ASHP Midyear Clinical Meeting in Anaheim, CA, and supported by an educational grant from Cardinal Health. Mr. Schneider received an honorarium for participating in the symposium and writing this article. Mr. Schneider has no financial relationships to report. References 1 Institute of Medicine. Preventing medication errors: quality chasm series. Washington, DC: National Academy Press. 2006 . 2 Covington TR and Pfeiffer FG. The pharmacist-acquired medication history. Am J Hosp Pharm . 1972 ; 29 : 692 –5. PubMed 3 Pedersen CA, Schneider PJ, Scheckelhoff DJ. ASHP national survey of pharmacy practice in hospital settings: Monitoring and Patient Education—2006. Am J Health-Syst Pharm . 2007 ; 64 : 507 –20. Crossref Search ADS PubMed 4 Bates DW. Computerized physician order entry and medication errors: finding a balance. J Biomed Inform . 2005 ; 38 : 259 –61. Crossref Search ADS PubMed 5 Bates DW, Cullen DJ, Laird N et al. Incidence of adverse drug events and potential adverse drug events—implications for prevention. JAMA . 1995 ; 274 : 29 –34. Crossref Search ADS PubMed 6 Leape LL, Bates DW, Cullen KJ et al. Systems analysis of adverse drug events. JAMA . 1995 ; 274 : 35 –43. Crossref Search ADS PubMed 7 Crawford SY and Myers CE. ASHP national survey of hospital-based pharmaceuticals. Am J Hosp Pharm . 1993 ; 50 : 1371 –1404. PubMed 8 Johnson R, Coles BJ, Tribble DA. Accuracy of three automated compounding systems determined by end-product laboratory testing and comparison with manual preparation. Am J Health-Syst Pharm . 1998 ; 55 : 1503 –7. Crossref Search ADS PubMed 9 Shultz S, White SJ, Latiolais CJ. Medication errors reduced by unit dose. Hospitals - JAHA . 1973 ; 47 (6): 106 –12. 10 Traynor K. FDA to require bar coding of most pharmaceuticals by mid-2006. Am J Health-Syst Pharm . 2004 ; 61 : 644 –5. Crossref Search ADS PubMed 11 Schneider PJ, Pedersen CA, Montanya KR et al. Improving the safety of medication administration using an interactive CD-ROM program. Am J Health-Syst Pharm . 2006 ; 63 : 59 –64. Crossref Search ADS PubMed 12 Rogers EM. Diffusion of Innovations. New York; 1995 : The Free Press. 13 Thompson CA. JCAHO views medication reconciliation as adverse-event prevention. Am J Health-Syst Pharm . 2005 ; 62 : 1528 ,1530,1532. Crossref Search ADS PubMed 14 Pedersen CA, Schneider PJ, Scheckelhoff DJ. ASHP national survey of pharmacy practice in hospital settings: Dispensing and administration—2005. Am J Health- Syst Pharm . 2006 ; 63 : 327 –45. Crossref Search ADS PubMed 15 Bussard B, McAlearney AS, Pedersen CA et al. Adoption of technology to improve medication safety: Perspectives of Pharmacy Directors. J Pt Safety . 2006 ; 2 :2(4): 217 –24. Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.

Nicol,, Natasha

doi: 10.2146/ajhp070191pmid: 17617509

McLeod Regional Medical Center is a 476-bed tertiary-care hospital with a patient volume of 25,000 discharges annually. In 2001, we committed ourselves to improving medication safety at our institution and began a series of organizational, technological, and process improvements. We started with a vision: “We are perfecting the medication delivery system to be safe for every patient, every time, while making it easy for caregivers to do the right thing, and impossible to do the wrong thing.” Some people took issue with the word “perfecting,” but we decided we had to stick with the word “perfecting” in our vision, because we could not accept the idea that it was allright to harm anyone. Establishing a culture of safety is essential. This requires years of effort and leadership that are immersed, not just involved, in bringing about change. “No blame” is key; however, no blame does not mean no responsibility. If people follow the guidelines, use the system as intended, and events still occur, individuals are not necessarily to blame. If people exceed the limits, then responsibility certainly plays a part. Kotter1 helped committee members stay on track and move beyond “be more careful” to focus on systemic issues. The next step was to redesign the Adverse Drug Event Committee and rename it to the Medication Safety Committee. The new committee was much more interdisciplinary and included physicians, pharmacists, nurses, dietitians, respiratory therapists, educators, and representatives from anesthesia, surgery, children’s services, marketing, and information services (IS). The active involvement of IS was particularly important. Another significant change was the addition of patients to the committee. Clinicians cannot separate themselves from being healthcare workers who come to the defense of the healthcare system. Patients have a completely different perspective and can identify opportunities for improvement that otherwise never would be discovered. In the past, perhaps five people would attend an ADE Committee meeting and discuss perhaps eight reported errors. No one reported significant errors because they would be blamed for the error and possibly fired. As a result, committee members might spend an entire meeting discussing seven pharmacy mis-fills and late delivery of a non-essential suppository. With the new medication safety improvement efforts, it was very energizing to remake the committee to focus on the system. Process and automation developments Each step in the medication-use process represents an opportunity for error; therefore, eliminating steps presents an important opportunity to improve medication safety. To establish a baseline, we met with pharmacy, nursing, anesthesia, and anyone else involved with medications and identified 17 major steps in our medication-use process. We then worked to identify ways to streamline and simplify the system. A major change was to decentralize our pharmacy operations. Our medication safety technologies included the Siemens Pharmacy System2 and the Pyxis Connect physician order management system3 that allows orders to be scanned to the pharmacy instead of being sent in paper form. The Pyxis MedStation 30004 automated medication management system allowed nurses in patient care areas to access medications from patient-profile-based automated dispensing machines (ADMs). Physicians began using handheld personal computers (PCs) as a precursor to computerized prescriber order management (CPOM). Pharmacy and Therapeutics (P&T) allowances were made for automatic intravenous (i.v.)-to-oral switches and renal dosing. This empowered pharmacists to make changes without having to call a physician, which helped to streamline the ordering process and increased pharmacists’ job satisfaction. We also hired pharmacy specialists, including pharmacists trained in oncology, internal medicine, and critical care. A pediatric and neonatal intensive care satellite was opened. Having this satellite pharmacy was considered extremely important because of the tighter variances required for neonatal safety and resulted in a dramatic decrease in the number of ADEs. The Siemens Med Administration Check system5 for bar coding was implemented in 2003 following extensive research, literature searches, and numerous site visits. To paraphrase Carol Haraden from the Institute for Healthcare Improvement, “Steal shamelessly, do not reinvent anything.” It was also important to learn about common work-arounds, in order to develop and implement safe and practical processes. We found that people often worry most about bar coding drugs in the pharmacy; however, we found this was not difficult. In 2003 we had to bar code about 50% of the drugs. Now that manufacturers provide more readable bar codes on their unit dose packaging, we pharmacy bar code only 19–20% of drugs. A bigger challenge is that bar-code implementation requires significant changes to the order entry system and processes. For example, pharmacists must enter drug orders with specific administration times, in order for a nurse to electronically chart drug administration appropriately. In September 2006 we implemented the Alaris System,6 a modular i.v. medication safety system that incorporates smart pumps and patient respiratory monitoring on a common user interface. A significant investment of pharmacy time was required to develop the customized safety software data sets, but the return in regards to patient safety is huge. The dose-error reduction software (DERS) allows maximum– minimum limits (i.e., “guardrails”) to be preestablished to avoid over- or underdosing. The inclusion of smart pumps, continuous pulse oximetry, and capnography on a single platform allows simultaneous monitoring of respiratory status of intubated or nonintubated patients receiving i.v. infusions. Medication reconciliation The need for improved medication reconciliation was shown by a project in which a pharmacy resident pulled medication histories, talked to patients, called pharmacies, did all the checking, and compared that information to what was produced in the hospital by nurses. In the first 15 patients, there were 75 errors in the hospital’s medication lists. This suggests strongly that medication reconciliation needs to be led by pharmacy, and any related logistical challenges need to be resolved. Relating specific events that happened at our institution had the greatest impact on motivating people to change. For example, a hypotensive patient was admitted through our emergency department (ED) and the physician wrote, “Continue home meds.” It turned out that the patient was on four different antihypertensives ordered by various physicians. The patient was stabilized and discharged, and the staff wrote, “Continue home meds” on discharge. The patient returned to the ED, and the staff again wrote, “Continue home meds.” The patient did not survive. Reconciling a patient’s medications is important; telling such stories makes a difference. As for medication safety, we began with a clear vision for medication reconciliation: “Every patient in the region will have a complete and accurate list of current medication available to all caregivers, regardless of setting, in order to facilitate a safe medication delivery system.” Medication reconciliation goals and items to be included on admission, transfer, and discharge are shown in Table 11. Table 1. Medication Reconciliationa aOTC = over-the-counter medications, MD = physician, D/C = discontinue, OR = operating room, RN = nurse. Goals aintain accurate medication lists throughout hospital stay Reduce duplication of efforts Decrease transcription Focus on ease of process Verify patient identity (use bar codes). Automate whenever possible Admission Reconciliation (serves as order sheet to be verified by a pharmacist) List of current home meds, OTC, herbals, samples Source of information Option for MD to continue, D/C, or change drug Retail pharmacy of choice Transfer Reconciliation New orders required at any transfer (level of care, OR, Cath Lab) No “continue all meds” allowed Nurse prints electronic Transfer Med Summary Order Form Contains all current hospital meds and home meds as reference Physician indicates whether to continue, stop, or change orders All previous orders invalid without form Discharge Reconciliation Automated for MD or RN Patient receives Universal Medication Form Instructions on meds and general care Instructions on maintaining accurate list of meds Instructed to share Universal Medication Form with All doctors All pharmacists Any other healthcare provider Copy sent to physician, retail pharmacy aOTC = over-the-counter medications, MD = physician, D/C = discontinue, OR = operating room, RN = nurse. Goals aintain accurate medication lists throughout hospital stay Reduce duplication of efforts Decrease transcription Focus on ease of process Verify patient identity (use bar codes). Automate whenever possible Admission Reconciliation (serves as order sheet to be verified by a pharmacist) List of current home meds, OTC, herbals, samples Source of information Option for MD to continue, D/C, or change drug Retail pharmacy of choice Transfer Reconciliation New orders required at any transfer (level of care, OR, Cath Lab) No “continue all meds” allowed Nurse prints electronic Transfer Med Summary Order Form Contains all current hospital meds and home meds as reference Physician indicates whether to continue, stop, or change orders All previous orders invalid without form Discharge Reconciliation Automated for MD or RN Patient receives Universal Medication Form Instructions on meds and general care Instructions on maintaining accurate list of meds Instructed to share Universal Medication Form with All doctors All pharmacists Any other healthcare provider Copy sent to physician, retail pharmacy Open in new tab Table 1. Medication Reconciliationa aOTC = over-the-counter medications, MD = physician, D/C = discontinue, OR = operating room, RN = nurse. Goals aintain accurate medication lists throughout hospital stay Reduce duplication of efforts Decrease transcription Focus on ease of process Verify patient identity (use bar codes). Automate whenever possible Admission Reconciliation (serves as order sheet to be verified by a pharmacist) List of current home meds, OTC, herbals, samples Source of information Option for MD to continue, D/C, or change drug Retail pharmacy of choice Transfer Reconciliation New orders required at any transfer (level of care, OR, Cath Lab) No “continue all meds” allowed Nurse prints electronic Transfer Med Summary Order Form Contains all current hospital meds and home meds as reference Physician indicates whether to continue, stop, or change orders All previous orders invalid without form Discharge Reconciliation Automated for MD or RN Patient receives Universal Medication Form Instructions on meds and general care Instructions on maintaining accurate list of meds Instructed to share Universal Medication Form with All doctors All pharmacists Any other healthcare provider Copy sent to physician, retail pharmacy aOTC = over-the-counter medications, MD = physician, D/C = discontinue, OR = operating room, RN = nurse. Goals aintain accurate medication lists throughout hospital stay Reduce duplication of efforts Decrease transcription Focus on ease of process Verify patient identity (use bar codes). Automate whenever possible Admission Reconciliation (serves as order sheet to be verified by a pharmacist) List of current home meds, OTC, herbals, samples Source of information Option for MD to continue, D/C, or change drug Retail pharmacy of choice Transfer Reconciliation New orders required at any transfer (level of care, OR, Cath Lab) No “continue all meds” allowed Nurse prints electronic Transfer Med Summary Order Form Contains all current hospital meds and home meds as reference Physician indicates whether to continue, stop, or change orders All previous orders invalid without form Discharge Reconciliation Automated for MD or RN Patient receives Universal Medication Form Instructions on meds and general care Instructions on maintaining accurate list of meds Instructed to share Universal Medication Form with All doctors All pharmacists Any other healthcare provider Copy sent to physician, retail pharmacy Open in new tab To help streamline medication reconciliation, McLeod’s IS Department created numerous electronic reporting forms. The nurse inputs the prescription information from the patient, which electronically populates a form that may be printed. The medication reconciliation form also is used as an order form, so a physician does not have to duplicate efforts and rewrite or transcribe again. The form includes the retail pharmacy of choice to provide at least an initial contact for someone to verify prescriptions that may have been filled. Measurement To validate the effectiveness of medication safety improvement efforts, key measurements must be tracked from start to finish. A non-punitive reporting system was used to track and trend reported errors. In 2001 reported errors averaged 70 per year; we now average more than 150 per month. This does not mean that patient safety has decreased; rather, that people are now more comfortable reporting errors, including serious errors. Medication turn-around times were measured before and after automation. The number of bar-coding “saves” and CPOM and handheld PC utilization were tracked and trended. We used the IHI Trigger Tool for Measuring ADEs7 with prospective or retrospective chart reviews to identify harmful events and determine the rate of harm. Typically rate of harm is expressed as number of events per 1,000 doses. However, in 2001 we felt that discussing overall safety in terms of “3.5 harmful events per 1,000 doses” would have little meaning to clinicians, so we converted this to number of patients per day. We dispense about 10,000 doses a day hospital-wide. So instead of saying 3.5 harmful events per 1,000 doses, we told our employees that ADEs occurred in 35 patients per day, which has much greater impact. Results As Kotter discussed, an important aspect of achieving change is to celebrate the successes along the way.1 If there is no opportunity to reflect and to celebrate achievements, people will lose interest and the safety effort will lose momentum. Table 22 and Figures 11 and 2 2 show medication safety achievements celebrated at McLeod by the end of 2006. Figure 2. Open in new tabDownload slide Medication Administration Check (Bar Coding) “Saves” per 1000 Doses (December 2004 to present = 3.06 saves per 1000 doses). Figure 2. Open in new tabDownload slide Medication Administration Check (Bar Coding) “Saves” per 1000 Doses (December 2004 to present = 3.06 saves per 1000 doses). Figure 1. Open in new tabDownload slide Rate of Harm Presented as “Harmful Events Per Day.” Figure 1. Open in new tabDownload slide Rate of Harm Presented as “Harmful Events Per Day.” Table 2. Results of Medication Safety Improvement Efforts 2001–2006a aADEs = adverse drug events, PCs = personal computers, CPOM = computerized prescriber order management. Steps in medication administration process decreased from 17 to 5 >95% of medications available through automation 90% decrease in harmful events: from 35 per day in 2001 to a running rate of 0.09 per day for the last six months of 2006 ADEs in neonatal patients decreased from a high of 4.78/1000 doses to a rate of 0.3/1000 doses for the last 12 months Medication turnaround time decreased from 92 to 17 minutes More than 85% of medical records are now online 130 physicians use handheld PCs 45 physicians use CPOM 265 orders sets are available online aADEs = adverse drug events, PCs = personal computers, CPOM = computerized prescriber order management. Steps in medication administration process decreased from 17 to 5 >95% of medications available through automation 90% decrease in harmful events: from 35 per day in 2001 to a running rate of 0.09 per day for the last six months of 2006 ADEs in neonatal patients decreased from a high of 4.78/1000 doses to a rate of 0.3/1000 doses for the last 12 months Medication turnaround time decreased from 92 to 17 minutes More than 85% of medical records are now online 130 physicians use handheld PCs 45 physicians use CPOM 265 orders sets are available online Open in new tab Table 2. Results of Medication Safety Improvement Efforts 2001–2006a aADEs = adverse drug events, PCs = personal computers, CPOM = computerized prescriber order management. Steps in medication administration process decreased from 17 to 5 >95% of medications available through automation 90% decrease in harmful events: from 35 per day in 2001 to a running rate of 0.09 per day for the last six months of 2006 ADEs in neonatal patients decreased from a high of 4.78/1000 doses to a rate of 0.3/1000 doses for the last 12 months Medication turnaround time decreased from 92 to 17 minutes More than 85% of medical records are now online 130 physicians use handheld PCs 45 physicians use CPOM 265 orders sets are available online aADEs = adverse drug events, PCs = personal computers, CPOM = computerized prescriber order management. Steps in medication administration process decreased from 17 to 5 >95% of medications available through automation 90% decrease in harmful events: from 35 per day in 2001 to a running rate of 0.09 per day for the last six months of 2006 ADEs in neonatal patients decreased from a high of 4.78/1000 doses to a rate of 0.3/1000 doses for the last 12 months Medication turnaround time decreased from 92 to 17 minutes More than 85% of medical records are now online 130 physicians use handheld PCs 45 physicians use CPOM 265 orders sets are available online Open in new tab Summary Essential elements of medication safety improvement efforts at McLeod Regional Medical Center included an emphasis on a culture of safety, a clear vision, a highly interdisciplinary team, and leadership that was immersed in the change process. Automation and process improvements included the use of a pharmacy information system, physician order management system, decentralized drug dispensing and distribution with patient-profile-based ADMs, and pocket PCs as a precursor to CPOM. Creation of a neonatal intensive care satellite pharmacy helped to dramatically reduce ADEs in this vulnerable population. Trigger tools were used to calculate the rate of harm, and a nonpunitive reporting system was used to track reported errors. Turn-around times, bar-coding “saves,” and utilization of CPOM and handheld PCs were also tracked. Electronic reporting forms facilitated accurate medication reconciliation at key points in patients’ hospitalization. At this tertiary-care hospital these efforts dramatically decreased the number of major steps in the medication-use process and medication turn-around time, while reducing the rate of harm from more than 35 harmful events per day in 2001 to less than one harmful event per day in each of the last six months of 2006. As these results make clear, measurable improvements in medication safety are eminently achievable, as hospitals work to implement the recommendations of the latest IOM report.8 Footnotes Based on the proceedings of a symposium held December 3, 2006, during the ASHP Midyear Clinical Meeting in Anaheim, CA, and supported by an educational grant from Cardinal Health. Dr. Nicol received an honorarium for participating in the symposium and writing this article. Dr. Nicol is an employee of Cardinal Health. References 1 Kotter JP. Leading Change. Boston, MA: Harvard Business School Press; 1996 . 2 Siemens Pharmacy, Siemens Medical Solutions USA, Inc., Malvern, PA. 3 Pyxis Connect, Cardinal Health, Inc., San Diego, CA. 4 Pyxis MedStation 3000, Cardinal Health, Inc., San Diego, CA. 5 Siemens Med Administration Check, Siemens Medical Solutions USA, Inc., Malvern, PA. 6 Alaris System, Cardinal Health, Inc., San Diego, CA. 7 Rozich JD, Haraden CR, Resar RK. Adverse drug event trigger tool: A practical methodology for measuring medication related harm. Qual Saf Health Care . 2003 ; 12 (3): 194 –200. www.ihi.org/IHI/Topics/PatientSafety/MedicationSystems/Literature/AdverseDrugEventTriggerTool.htm (accessed 2007 Jan 25). Crossref Search ADS PubMed 8 Institute of Medicine. Preventing medication errors: quality chasm series. Washington, DC: National Academy Press; 2006 . Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.

Witkowski,, Paul

doi: 10.2146/ajhp070192pmid: 17617510

This case study will focus on the novel ways that automation enhances medication safety—not how it can directly help avert errors, but how it can make time available for the staff to focus on medication therapy management and work directly with physicians and nurses to improve medication safety and quality of care. Alliance Community Hospital is an independent, nonteaching, 245- bed community hospital that also services 75 long-term/subacute beds. In January of 2006 we moved into a new facility. We have eight full-time equivalent (FTE) pharmacists and 7.5 FTE certified pharmacy technicians (CPhTs). When I started as Director of Pharmacy six years ago, we had no pharmacists scheduled to work in patient care areas. Now we have two full-time pharmacists who work in patient care areas every weekday without increasing our number of FTEs. The key has been the use of automation and maximizing the use of our pharmacy technician staff. Pharmacy technicians About four years ago, I informed my technicians that I wanted them all to become certified, which made them very nervous. We spent a year working with them, and they all passed the national certification exam. What I had not anticipated was how much better they would feel about themselves and how much more confident they would be in their capabilities by becoming certified pharmacy technicians. The end result has been the evolution of a technician staff willing to accept more decision-making responsibilities and to work independently with more complex tasks. Our entire drug distribution and dispensing system is highly automated. We order and receive drugs electronically. We have an automated just-in-time inventory refill service and use a bar-code pick-check- delivery technology (Pyxis PARx) with an electronic audit trail. We use automated dispensing machines (ADMs) with a patient-specific profile interface for nurses to access the drugs directly in patient care areas. We implemented this cartless drug distribution system four years ago and implemented remote medication order review by our pharmacists the moment we went cartless. All medications are pharmacist-verified using bar-code scanning (Table 11). For drug procurement and distribution, we now spend less than one second of pharmacist time per dose of medication administered. Our pharmacists have only a very small role in checking medications based on what is required to meet legal and safety standards. As a result, our pharmacists can now focus most of their time on medication therapy management. Table 1. Drug Distribution: Pharmacist and Technician Tasksa CPhT Tasks RPh Taska aCPhT = certified pharmacy technician, RPh = registered pharmacist, ADMs = automated dispensing machines. Drug receiving All meds pharmacist-verified by scanning Medication selection and processing for ADM restocking Controlled substance storage and distribution Med refill in ADMs CPhT Tasks RPh Taska aCPhT = certified pharmacy technician, RPh = registered pharmacist, ADMs = automated dispensing machines. Drug receiving All meds pharmacist-verified by scanning Medication selection and processing for ADM restocking Controlled substance storage and distribution Med refill in ADMs Open in new tab Table 1. Drug Distribution: Pharmacist and Technician Tasksa CPhT Tasks RPh Taska aCPhT = certified pharmacy technician, RPh = registered pharmacist, ADMs = automated dispensing machines. Drug receiving All meds pharmacist-verified by scanning Medication selection and processing for ADM restocking Controlled substance storage and distribution Med refill in ADMs CPhT Tasks RPh Taska aCPhT = certified pharmacy technician, RPh = registered pharmacist, ADMs = automated dispensing machines. Drug receiving All meds pharmacist-verified by scanning Medication selection and processing for ADM restocking Controlled substance storage and distribution Med refill in ADMs Open in new tab ADMs are restocked completely by technicians. They determine restocking needs and print out a bar-coded label for each medication required. The pharmacists’ only involvement is to verify that the medication and quantity are correct and consistent with the bar-coded label. Scanning their name badge and the item to verify that the bar-coded package is correct also creates an electronic audit trail. The checked medications are then taken to the patient care areas and placed in the ADMs by our technicians, again using a bar-code scan verification. Automation Medications are bar-code scanned five times between the time they enter the building and the time they are administered to the patient: 1) upon receipt from the wholesaler; 2) when the medication is selected by our technicians for refill into our dispensing cabinets; 3) in the pharmacist verification step; 4) when the drug is scanned into the ADM to open the drug’s precise refill location (e.g., correct ADM, drawer, and pocket); and 5) at the bedside, when the nurse scans for the “5 rights” (right person, medication, dose, time, and route) and administration is documented in the electronic medication administration record (eMAR). We have had bar-code medication administration (BCMA) and the eMAR for two years. After receiving a drug from a wholesaler (first bar-code scan), a complete electronic audit trail for each package tells us who filled the medication and when (second bar-code scan), who checked it and when (third bar-code scan), and who loaded it into the correct ADM and when (fourth bar-code scan). This very tight system of checking also makes internal diversion of drugs much more difficult. A major question is, “How much does all of this cost?” Automation systems appear to carry a substantial price tag, and pharmacy directors need to successfully cost-justify them in order to obtain approval for purchase and implementation. When I summed all of our withdrawal transactions over the last three months and the total cost of automation systems, our automation costs averaged just 34 cents per withdrawal transaction. The components of our automated system virtually assures no lost charges, no outdated drugs, no free floorstock, no missing doses, no narcotic shift counts, tight control on every item including plain intravenous (i.v.) solutions, no calls for replacement doses, commercially premixed i.v. fluids stocked directly into the ADM, and drugs reordered directly from the whole-saler to the machine. That buys me a tremendous amount of financial and logistical efficiency—things that chief financial officers and hospital administrators care about. Almost all (92%) of our medications are now taken from ADMs. For some i.v. medications we use frozen premixes. To store these drugs on the nursing units we have refrigerators attached to all our ADMs. We handle these refrigerated premixed i.v. medications just like oral dosage forms. They are loaded and stocked in the machines, as opposed to being individually processed, pulled, labeled, checked, and dispensed out of the pharmacy. Turnaround time Our turnaround time from when the pharmacist enters the order to when the medication is available on the floor is only two to three minutes, and the eMAR is updated instantly. For example, if a physician writes a prescription for one gram of i.v. ceftriaxone, the pharmacist in the patient care area reviews the order, has the opportunity to talk about it with the physician, and then enters the order into the system electronically. When the pharmacist hits the “enter” button, within two to three minutes’ time the nurse can go to the local ADM and take the drug out of the integrated refrigerator. It does not even need a label from the pharmacy, because it has a bar code. The nurse takes the drug to the bedside and scans the bar code, the drug name appears on the eMAR, the system verifies the five rights, the nurse administers the dose, and the administration is automatically posted in detail on the patient’s eMAR. Our physicians have been impressed with our drug distribution system. When they prescribe a drug, the pharmacist reviews and enters the order right on the floor. The nurse commonly is taking the drug into the room while the physician is still there. Sometimes the ink on the order sheet is hardly even dry, and the medication is going into the room to be administered and electronically charted. That is very impressive to everyone involved. Medication therapy management Freeing our pharmacists from most of the tasks involved in drug procurement and distribution has allowed them to spend much more time on medication therapy management (Figure 11). Before implementing automation and delegating most of the drug distribution tasks to pharmacy technicians, the majority of the pharmacists’ time was spent on verifying the accuracy of newly filled orders and checking unit dose carts. Those tasks are labor-intensive and very expensive when you consider the amount of pharmacist time they consume. Now that we have automated such checking, much more pharmacist time is available for medication therapy management without having to increase our number of pharmacists. Pharmacist time spent in medication therapy management has thus been maximized. Figure 1. Open in new tabDownload slide Pharmacist Task Distribution. Figure 1. Open in new tabDownload slide Pharmacist Task Distribution. Pharmacists review and enter the medication orders. Pharmacy operations are decentralized during the day and centralized during the evening. During the night when the pharmacy is closed, an off-site tele-pharmacy service processes the orders. For example, if an order is written at 2:00 a.m. when our pharmacy is closed, the order is faxed to a facility 150 miles away, where it is reviewed by a pharmacist and entered into our computer system via a secure VPN internet connection. New orders written during hours that the pharmacy is closed populate our systems in an identical fashion, just as if they had been processed on site. Nurses can also telephone these pharmacists with questions. It is virtually transparent that all of this actually takes place in another city. I consider entering orders to be part of the medication therapy management system and not the drug distribution system, because our pharmacists are making decisions on how and when each medication is to be administered while at the same time considering all of the other medications the patient is receiving. As we have literally moved closer to our patients, we have come to recognize that some medication orders written just do not make sense for the patient. Pharmacists have to get to know their patients to recognize those orders. For example, we had an order written for “three amps of bicarb in one liter of i.v. fluid” for a patient in our intensive care unit (ICU). To a pharmacist in a centralized pharmacy, this would not be an uncommon order for an ICU patient. But our floor pharmacist was more familiar with the patient and did not recall seeing anywhere that the patient was acidotic. The pharmacist therefore inquired why we were giving bicarb to this patient. The physician then realized he had intended to write “calcium” but inadvertently had written “bicarb.” In another very recent example, “Mucinex 600mg b.i.d.” was ordered for a patient. The pharmacist reviewing the order was unaware of any congestion problems with this patient. He also saw that the patient was scheduled for an imaging study and realized the order was intended to be written for “Mucomyst” 600mg, not “Mucinex.” In both cases, if the pharmacists had not been clinically familiar with the patients, these errors never would have been detected. I am not implying that we are performing any exceptional clinical interventions, but that our use of automation and technicians has enabled us to initiate these activities and achieve this level of medication safety with our existing staff, as opposed to having to hire additional staff. By offloading most of the drug distribution tasks from the pharmacists and delegating these functions to our technicians and automation, we have freed up the pharmacists’ time to concentrate more on medication therapy management in the patient care areas. Summary Our experience shows that by making extensive use of automation and enhancing the role of pharmacy technicians, we have been able to expand our activities significantly in medication therapy management with the same number of pharmacists and the same amount of pharmacists’ time. We have been able to give our pharmacists more time and opportunity to really know their patients and work directly with physicians and nurses, and as a result improve both medication safety and the quality of care. Footnotes Based on the proceedings of a symposium held December 3, 2006, during the ASHP Midyear Clinical Meeting in Anaheim, CA, and supported by an educational grant from Cardinal Health. Dr. Witkowski received an honorarium for participating in the symposium and writing this article. Dr. Witkowski reports he serves on the speakers’ bureau for Ortho-McNeil. Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.

Bates, David, W.

doi: 10.2146/ajhp070190pmid: 17617512

Abstract Purpose. To summarize key recommendations and supporting evidence from the most recent Institute of Medicine (IOM) report, Preventing Medication Errors. Summary. Starting in 2000, IOM reports brought the problem of medical safety into public awareness and made four major points: errors are common and costly, systems cause errors, errors can be prevented and safety can be improved, and medication-related adverse events are the single leading cause of injury. The most recent report is an attempt to think about what needs to be done to reach the next level of medication safety. Some have had difficulty implementing these recommendations, but these challenges can be overcome by learning from these experiences. Evidence supporting the recommendations made in this report includes research on computerized prescriber order entry (renal insufficency geriatric patients, meta-analysis, unintended consequences, pediatric transfer patients); intravenous infusion safety systems; and dispensing errors and bar-coding. Conclusion.Preventing Medication Errors lays out a blueprint for change in medication safety. The report makes clear that providers have m any opportunities to improve. Technologies, such as computerized order entry, bar-coding and smart pumps and computerized ADE monitoring, will undoubtedly play a key role, and institutions should be thinking seriously about implementing a number of these. The report also emphasizes how essential a culture change, combined with well-designed technologies, will be necessary to achieve the next level of safety called for in the IOM report. Computers, Economics, Errors, medication, Hospitals, Institute of Medicine, Patient information, Pharmacy, institutional, hospital, Quality assurance, Risk management, Technology, Toxicity Starting in 2000, the Institute of Medicine (IOM) began publishing a remarkably successful series of reports on quality in healthcare. To Err is Human brought the problem of medical safety into public awareness.1,Crossing the Quality Chasm made four major points: errors are common and costly, systems cause errors, errors can be prevented and safety can be improved, and medication-related adverse events are the single leading cause of injury.2 Even more importantly, the report suggested that between the healthcare that exists today and the healthcare that is possible lies not just a gap but a chasm, and it attempted to provide a blueprint for how that chasm might be crossed.2 That has not yet been accomplished, yet progress is being made. The most recent report, Preventing Medication Errors, is an attempt to think about what needs to be done to reach the next level of medication safety.3 The work underlying Preventing Medication Errors began in 2004 when Congress mandated the Center for Medicare and Medicaid Services (CMS) to sponsor the IOM to conduct a comprehensive study of drug safety and quality issues in order to provide a blueprint for system-wide change. First, an epidemiological review showed that, estimated very conservatively, medications harm at least 1.5 million people per year. In hospitals, there are at least 400,000 preventable adverse drug events (ADEs) per year, or approximately one medication error per patient per day. That finding probably had the greatest impact on the general public.3 IOM recommendations The IOM report’s recommendations are summarized in Table 11.3 The emphasis on the patient is evident throughout the report, especially in Recommendation 1. Recommendation 2 reflects the feeling that the resource base for consumer-oriented drug information today is not adequate. Providers today do not necessarily have access to comprehensive reference information when they are delivering care, yet this is vitally important. Table 1. Preventing Medication Errors recommendations3 Recommendation 1. To improve the quality and safety of the medication use process, specific measures should be instituted to strengthen patients’ capacities for sound medication self-management. Patients’ rights should be formalized. Patients should maintain an active list to which all providers should have access. Providers should educate patients about their medications. Consultation on medications should be available to patients at key points in the medication use process, e.g., when a patient is admitted and discharged and receives medication at a pharmacy. Recommendation 2. Government agencies (Agency for Healthcare Research and Quality [AHRQ], CMS, the Food and Drug Administration [FDA]) should enhance the resource base for consumer-oriented drug information and medication of self-management support. Pharmacy leaflets should be standardized. The National Library of Medicine (NLM) should be chief internet resource for consumers. FDA, CMS, and NLM should evaluate approaches for building a national network of drug information help-lines. Same group should confirm minimum dataset for public health records (PHRs). National plan for widespread distribution of medication safety information should be developed. Recommendation 3. All healthcare organizations should make available to providers patient information and decision-support tools to enable providers to: Access comprehensive reference information, Communicate patient-specific medication information interoperably, Assess safety through active monitoring, By 2008, prescribers should have plans in place to e-prescribe and by 2010 write all and have pharmacies receive all electronically, Subject prescriptions to decision support, Have the appropriate competencies for the medication use process, Make effective use of technologies, which will vary by setting. Recommendation 4. Better labeling is needed, as are better methods for communicating medication information to consumers. FDA should develop guidance documents to industry for labeling and packaging. Studies about design of labeling and information sheets should be done. FDA should work with industry to develop a strategy for expanding unit-of-use packaging. AHRQ should fund studies evaluating impact of samples on safety, prescribing behavior, and consumer choice. Recommendation 5. Industry and government should collaborate to establish standards affecting drug-related healthcare information technology (HIT): NLM and drug nomenclature AHRQ and safety alert mechanisms by severity, frequency, and clinical importance, including -Intelligent prompting -Human factors -Specifications for alerts. Recommendation 6. Congress should fund AHRQ to work with other agencies to develop a broad research agenda on safe and appropriate medication use, especially testing of error prevention strategies: Annual level of investment $100 million [Current levels of investment are approaching zero]. Recommendation 7. Oversight and regulatory organizations and payers should use [tactics] to motivate the adoption of practices that can reduce medication errors and ensure that providers have needed competencies. Payers and purchasers should provide explicit financial incentives. CMS should evaluate a variety of strategies for medication therapy management. Regulators should set minimum functionality standards for error prevention technologies. States should enable e-prescribing. State boards of pharmacy should undertake quality improvement (QI) initiatives related to community pharmacy practice. Medication error reporting should be promoted by all. Accreditors of professional education should require more training in medication-related areas. Recommendation 1. To improve the quality and safety of the medication use process, specific measures should be instituted to strengthen patients’ capacities for sound medication self-management. Patients’ rights should be formalized. Patients should maintain an active list to which all providers should have access. Providers should educate patients about their medications. Consultation on medications should be available to patients at key points in the medication use process, e.g., when a patient is admitted and discharged and receives medication at a pharmacy. Recommendation 2. Government agencies (Agency for Healthcare Research and Quality [AHRQ], CMS, the Food and Drug Administration [FDA]) should enhance the resource base for consumer-oriented drug information and medication of self-management support. Pharmacy leaflets should be standardized. The National Library of Medicine (NLM) should be chief internet resource for consumers. FDA, CMS, and NLM should evaluate approaches for building a national network of drug information help-lines. Same group should confirm minimum dataset for public health records (PHRs). National plan for widespread distribution of medication safety information should be developed. Recommendation 3. All healthcare organizations should make available to providers patient information and decision-support tools to enable providers to: Access comprehensive reference information, Communicate patient-specific medication information interoperably, Assess safety through active monitoring, By 2008, prescribers should have plans in place to e-prescribe and by 2010 write all and have pharmacies receive all electronically, Subject prescriptions to decision support, Have the appropriate competencies for the medication use process, Make effective use of technologies, which will vary by setting. Recommendation 4. Better labeling is needed, as are better methods for communicating medication information to consumers. FDA should develop guidance documents to industry for labeling and packaging. Studies about design of labeling and information sheets should be done. FDA should work with industry to develop a strategy for expanding unit-of-use packaging. AHRQ should fund studies evaluating impact of samples on safety, prescribing behavior, and consumer choice. Recommendation 5. Industry and government should collaborate to establish standards affecting drug-related healthcare information technology (HIT): NLM and drug nomenclature AHRQ and safety alert mechanisms by severity, frequency, and clinical importance, including -Intelligent prompting -Human factors -Specifications for alerts. Recommendation 6. Congress should fund AHRQ to work with other agencies to develop a broad research agenda on safe and appropriate medication use, especially testing of error prevention strategies: Annual level of investment $100 million [Current levels of investment are approaching zero]. Recommendation 7. Oversight and regulatory organizations and payers should use [tactics] to motivate the adoption of practices that can reduce medication errors and ensure that providers have needed competencies. Payers and purchasers should provide explicit financial incentives. CMS should evaluate a variety of strategies for medication therapy management. Regulators should set minimum functionality standards for error prevention technologies. States should enable e-prescribing. State boards of pharmacy should undertake quality improvement (QI) initiatives related to community pharmacy practice. Medication error reporting should be promoted by all. Accreditors of professional education should require more training in medication-related areas. Open in new tab Table 1. Preventing Medication Errors recommendations3 Recommendation 1. To improve the quality and safety of the medication use process, specific measures should be instituted to strengthen patients’ capacities for sound medication self-management. Patients’ rights should be formalized. Patients should maintain an active list to which all providers should have access. Providers should educate patients about their medications. Consultation on medications should be available to patients at key points in the medication use process, e.g., when a patient is admitted and discharged and receives medication at a pharmacy. Recommendation 2. Government agencies (Agency for Healthcare Research and Quality [AHRQ], CMS, the Food and Drug Administration [FDA]) should enhance the resource base for consumer-oriented drug information and medication of self-management support. Pharmacy leaflets should be standardized. The National Library of Medicine (NLM) should be chief internet resource for consumers. FDA, CMS, and NLM should evaluate approaches for building a national network of drug information help-lines. Same group should confirm minimum dataset for public health records (PHRs). National plan for widespread distribution of medication safety information should be developed. Recommendation 3. All healthcare organizations should make available to providers patient information and decision-support tools to enable providers to: Access comprehensive reference information, Communicate patient-specific medication information interoperably, Assess safety through active monitoring, By 2008, prescribers should have plans in place to e-prescribe and by 2010 write all and have pharmacies receive all electronically, Subject prescriptions to decision support, Have the appropriate competencies for the medication use process, Make effective use of technologies, which will vary by setting. Recommendation 4. Better labeling is needed, as are better methods for communicating medication information to consumers. FDA should develop guidance documents to industry for labeling and packaging. Studies about design of labeling and information sheets should be done. FDA should work with industry to develop a strategy for expanding unit-of-use packaging. AHRQ should fund studies evaluating impact of samples on safety, prescribing behavior, and consumer choice. Recommendation 5. Industry and government should collaborate to establish standards affecting drug-related healthcare information technology (HIT): NLM and drug nomenclature AHRQ and safety alert mechanisms by severity, frequency, and clinical importance, including -Intelligent prompting -Human factors -Specifications for alerts. Recommendation 6. Congress should fund AHRQ to work with other agencies to develop a broad research agenda on safe and appropriate medication use, especially testing of error prevention strategies: Annual level of investment $100 million [Current levels of investment are approaching zero]. Recommendation 7. Oversight and regulatory organizations and payers should use [tactics] to motivate the adoption of practices that can reduce medication errors and ensure that providers have needed competencies. Payers and purchasers should provide explicit financial incentives. CMS should evaluate a variety of strategies for medication therapy management. Regulators should set minimum functionality standards for error prevention technologies. States should enable e-prescribing. State boards of pharmacy should undertake quality improvement (QI) initiatives related to community pharmacy practice. Medication error reporting should be promoted by all. Accreditors of professional education should require more training in medication-related areas. Recommendation 1. To improve the quality and safety of the medication use process, specific measures should be instituted to strengthen patients’ capacities for sound medication self-management. Patients’ rights should be formalized. Patients should maintain an active list to which all providers should have access. Providers should educate patients about their medications. Consultation on medications should be available to patients at key points in the medication use process, e.g., when a patient is admitted and discharged and receives medication at a pharmacy. Recommendation 2. Government agencies (Agency for Healthcare Research and Quality [AHRQ], CMS, the Food and Drug Administration [FDA]) should enhance the resource base for consumer-oriented drug information and medication of self-management support. Pharmacy leaflets should be standardized. The National Library of Medicine (NLM) should be chief internet resource for consumers. FDA, CMS, and NLM should evaluate approaches for building a national network of drug information help-lines. Same group should confirm minimum dataset for public health records (PHRs). National plan for widespread distribution of medication safety information should be developed. Recommendation 3. All healthcare organizations should make available to providers patient information and decision-support tools to enable providers to: Access comprehensive reference information, Communicate patient-specific medication information interoperably, Assess safety through active monitoring, By 2008, prescribers should have plans in place to e-prescribe and by 2010 write all and have pharmacies receive all electronically, Subject prescriptions to decision support, Have the appropriate competencies for the medication use process, Make effective use of technologies, which will vary by setting. Recommendation 4. Better labeling is needed, as are better methods for communicating medication information to consumers. FDA should develop guidance documents to industry for labeling and packaging. Studies about design of labeling and information sheets should be done. FDA should work with industry to develop a strategy for expanding unit-of-use packaging. AHRQ should fund studies evaluating impact of samples on safety, prescribing behavior, and consumer choice. Recommendation 5. Industry and government should collaborate to establish standards affecting drug-related healthcare information technology (HIT): NLM and drug nomenclature AHRQ and safety alert mechanisms by severity, frequency, and clinical importance, including -Intelligent prompting -Human factors -Specifications for alerts. Recommendation 6. Congress should fund AHRQ to work with other agencies to develop a broad research agenda on safe and appropriate medication use, especially testing of error prevention strategies: Annual level of investment $100 million [Current levels of investment are approaching zero]. Recommendation 7. Oversight and regulatory organizations and payers should use [tactics] to motivate the adoption of practices that can reduce medication errors and ensure that providers have needed competencies. Payers and purchasers should provide explicit financial incentives. CMS should evaluate a variety of strategies for medication therapy management. Regulators should set minimum functionality standards for error prevention technologies. States should enable e-prescribing. State boards of pharmacy should undertake quality improvement (QI) initiatives related to community pharmacy practice. Medication error reporting should be promoted by all. Accreditors of professional education should require more training in medication-related areas. Open in new tab Recommendation 3 emphasizes the need to communicate patient-specific medication information “interoperatively,” i.e., so that it can be moved from one clinical setting to another.3 It should be possible, for example, for a patient to go from New York to Louisiana and still have the important medication information be accessible. Safety needs to be assessed through active monitoring. Today there is far too much reliance on spontaneous monitoring, which does not provide a solid sense of the magnitude of the problem in terms of medication safety. By 2008 prescribers should have plans in place to “e-prescribe” and by 2010 to write all, and have pharmacies receive all, prescriptions electronically. This is one of the most important recommendations in the report and is eminently achievable if the healthcare community focuses strongly on this goal and addresses the key policy issues surrounding it. Another critical point is that prescriptions need to be subjected to decision support that is more robust and standardized than it is today. In addition, all providers in the medication-use process need to have the appropriate competencies. For example, prescribers of oncology medications and pharmacists performing medication order reviews on oncology drugs need to have appropriate competencies to perform these tasks. In addition, providers need to make effective use of technologies. Recommendation 4 states that AHRQ should fund studies evaluating the impact of samples on safety, prescribing behavior, and consumer choice.3 Some of the committee members wanted to recommend that samples should just be eliminated. Recommendation 5 was to establish standards for drug-related healthcare information technology (HIT), which is essential to achieve the next level of safety.3 The National Library of Medicine (NLM) is already working on drug nomenclature but needs more support to finish this work. A standard way of representing drugs that can be used by everyone is still not available. This is remarkable, since standards are fairly robust in many other areas, such as for laboratory tests and imaging. AHRQ should evaluate safety alert mechanisms and include issues, such as intelligent prompting, human factors, and specifications for specific types of alerts. Finally, much of the benefit to be gained in medication safety will come from having good decision support, which is not currently standardized in the systems in use. Recommendation 6 was for an annual investment of approximately $100 million to fund AHRQ to work with other agencies to develop a broad research agenda on safe and appropriate medication use.3 Current levels of investment on research on patient safety broadly are approaching zero, except for issues relating to safety and health information technology. The money that AHRQ has to spend on patient safety in 2007 will all be spent on safety outside the hospital. More balance is needed. Recommendation 7, the policy recommendation calls for payers and purchasers to provide explicit financial incentives for doing things that improve medication safety.3 Financial incentives are being implemented, yet this is still in the very early stages. Many states have proposed legislation on e-prescribing that is starting to be passed, yet approximately 20 states have problems with this (e.g., electronic signatures are not permitted). Such problems need to be eliminated. Recommendations to pharmacy (Table 22) included the need to monitor the frequencies of errors and near-misses, so that corrections can be made.3 Today many pharmacies rely on spontaneous reporting. If interventions are made, it is important to track those interventions and use prescribing errors corrected by pharmacists as opportunities to identify areas for improvement. Table 2. Recommendations to Pharmacy3 Pharmacies shall: Regularly monitor the medication literature for drug error information and take action for prevention. Verify accurate entry of new prescription data. Monitor error frequencies and near misses (so corrections can be made). Report errors to external reporting programs. Verify patient identity (use bar codes). Educate consumers about error prevention. Use electronic prescribing by providers and prescriptions sent directly to pharmacies. Suppress trivial warnings from decision-support systems. Develop methods to monitor prescription-filling machines to avoid errors. Follow patients for side effects, especially if high-risk. Pharmacists shall: Review the patient’s medication list routinely (especially during transitions in care). Review different treatment options. Review the name and purpose of the selected medication. Discuss with the patient when and how to take medications. Pharmacies shall: Regularly monitor the medication literature for drug error information and take action for prevention. Verify accurate entry of new prescription data. Monitor error frequencies and near misses (so corrections can be made). Report errors to external reporting programs. Verify patient identity (use bar codes). Educate consumers about error prevention. Use electronic prescribing by providers and prescriptions sent directly to pharmacies. Suppress trivial warnings from decision-support systems. Develop methods to monitor prescription-filling machines to avoid errors. Follow patients for side effects, especially if high-risk. Pharmacists shall: Review the patient’s medication list routinely (especially during transitions in care). Review different treatment options. Review the name and purpose of the selected medication. Discuss with the patient when and how to take medications. Open in new tab Table 2. Recommendations to Pharmacy3 Pharmacies shall: Regularly monitor the medication literature for drug error information and take action for prevention. Verify accurate entry of new prescription data. Monitor error frequencies and near misses (so corrections can be made). Report errors to external reporting programs. Verify patient identity (use bar codes). Educate consumers about error prevention. Use electronic prescribing by providers and prescriptions sent directly to pharmacies. Suppress trivial warnings from decision-support systems. Develop methods to monitor prescription-filling machines to avoid errors. Follow patients for side effects, especially if high-risk. Pharmacists shall: Review the patient’s medication list routinely (especially during transitions in care). Review different treatment options. Review the name and purpose of the selected medication. Discuss with the patient when and how to take medications. Pharmacies shall: Regularly monitor the medication literature for drug error information and take action for prevention. Verify accurate entry of new prescription data. Monitor error frequencies and near misses (so corrections can be made). Report errors to external reporting programs. Verify patient identity (use bar codes). Educate consumers about error prevention. Use electronic prescribing by providers and prescriptions sent directly to pharmacies. Suppress trivial warnings from decision-support systems. Develop methods to monitor prescription-filling machines to avoid errors. Follow patients for side effects, especially if high-risk. Pharmacists shall: Review the patient’s medication list routinely (especially during transitions in care). Review different treatment options. Review the name and purpose of the selected medication. Discuss with the patient when and how to take medications. Open in new tab Pharmacists should report errors to external reporting programs, so that organizations such as the Institute for Safe Medication Practices (ISMP) can identify patterns and widespread problems. It is also recommended that pharmacists verify patient identity, use bar codes, and educate consumers about error prevention. Consumer education is an important issue. All too often, when patients receive their medications at a drug store, they merely take the packaged medications and move on without learning how to take them properly. The next set of recommendations reinforces the need to ensure that prescribing is electronic and that the prescriptions are going directly to the pharmacies. Also, trivial warnings from decision-support systems need to be suppressed. The final set of recommendations for individual pharmacists emphasized the need for repeated reviews and discussions at key points in the medication-use process, particularly upon admission and at discharge. Computerized prescriber order entry A 1998 study identified a 55% reduction in the serious medication error rate following the implementation of computerized prescribing—even a system with very limited physician support.4 There was an 83% reduction in the overall medication error rate.5 The cost of each preventable ADE was about $6000. Renal insufficiency. More recent studies have identified the results that can be obtained if decision support is added to computerized prescriber order entry (CPOE). One of the most important types of decision support is renal dosing decision support. The 2001 “Nephros” study evaluated the impact of delivering real-time decision support for patients with renal insufficiency and found significant benefit.6 The findings showed that a large number of patients (42% of 17,828) had some degree of renal insufficiency. There are several hundred drugs whose doses require adjustment for renal insufficiency. Remembering all these doses correctly exceeds the scope of any individual’s capabilities. The Nephros results showed that in the control group the prescribed dose was correct only 54% of the time, and the frequency only 35% of the time. In the intervention group, correct dose and frequency increased to 67% and 59%, respectively, and patients’ length of stay decreased by about a half-day.6 Financial analyses of the impact of decision support implementation also showed this to be one of the more important interventions. Geriatric patients. Another study looked at the impact of decision support on prescribing for geriatric inpatients.7 Elderly patients frequently get dosages that are too high, especially initial dosages. A randomized controlled trial with decision support for dosing of psychoactive drugs found that when prescribing suggestions were made, patients more often got the recommended dose (29% versus 19%). Most importantly, the rate at which patients fell decreased to 2.8 from 6.4 falls/1000 patient days. Nonetheless, the frequency with which people accepted the suggestions was still only 29% after the implementation of decision support. Thus, the results showed that it is clearly beneficial to suggest starting with a lower dosage, yet there is room for more improvement.7 Meta-analysis. A recent systematic review of the impact of CPOE on medication safety showed that of the five trials published at that time, two showed a marked decrease in the serious medication error rate, one showed improvement in the corollary order rate, one showed improvement in five prescribing behaviors, and one showed improvement in nephrotoxic drug dosing and frequency.8 Several additional studies are currently underway. Unintended consequences. Several recent studies discussed in the IOM report have been quite controversial. A study by Koppel et al.9 evaluated the commercial CPOE application at the University of Pennsylvania and asked users their impressions about the system. They found many situations in which “a leading CPOE system” facilitated medication error risks. They also found that it often took many screens to do things that people needed to do routinely, and that some needed screen views were not available.9 It is important to note that others have also reported on this issue.10 The Koppel study had a few issues.9,11 First, errors or adverse events were not actually counted. Second, the investigators said that other studies focused only on the advantages of CPOE, which is not accurate. In fact, previous studies actually evaluated how many errors were made before the implementation of order entry, how many occurred afterwards, and the difference. Our studies included the errors that were created by the system. Anytime you introduce a new system, it does cause some new errors. Third, the CPOE application that they studied was an old one.11 Nonetheless, valuable debate was stimulated and some key points identified from this published review. The first is that after a CPOE system is implemented, iterative changes have to be made. A system should not be installed and then left alone. At one time or another our institution experienced essentially all the problems reported on by Koppel et al.,9 yet following implementation serial changes were made to address those problems. The second point is that software alone is insufficient; people systems as well as software systems must be addressed.11 Pediatric transfer patients. A perhaps more troublesome study from the University of Pittsburgh looked at children who were transported in for special care.12 In that study the mortality rate increased from 2.8% to 6.3% after the introduction of a commercial CPOE application.12 Some caveats regarding these findings: first, the study design was before/after, and a number of other changes were made at the same time CPOE was implemented, making it difficult to assess the impact of CPOE alone; overall mortality was not reported; and CPOE was introduced very rapidly—over only six days. There are various ways to implement CPOE, and rapid implementation has some advantages, since parallel systems do not exist. Nonetheless, CPOE implementation is a substantial systems change. At Brigham and Women’s, CPOE was introduced over approximately three years. Moreover, after CPOE implementation, order entry was not allowed until the patient had actually entered the hospital and been logged into the system. The pediatric patients evaluated were those who were being transported in for special care because they were very sick; many were coming in by helicopter or ambulances. The previous approach was to have the house officer write the orders while the child was en route, so that everything would be ready and the pharmacy would have the drugs prepared when the child came through the door.12,13 Another issue was that after CPOE implementation, all the drugs, including the vasoactive drugs that many of these children needed, were moved to the central pharmacy, which was located away from the acute unit. A rule was implemented that said pharmacy was not allowed to process medication orders until after the orders were activated, which could not happen until the child had physically entered the institution. In addition, the decision was made to move ahead with implementation without putting in order sets. This has had a significant impact because groups of orders can be written much more quickly. Writing single orders on the computer is always slower. The net result was substantial delays in care delivery, which may have resulted in the differences seen before and after the implementation of CPOE.13 This study was quite weak methodologically. Nonetheless, the increase in mortality rate was very large and of obvious concern. Clearly, the organization broke many of the rules for implementation, and it is absolutely essential for other organizations that are implementing CPOE to handle the socio-technical aspects of implementation better.13 Intravenous infusion safety systems Another very important technology is intravenous (i.v.) infusion safety systems. These so-called “smart pumps” are important because few administration errors get caught. If a doctor makes an error in prescribing, a pharmacist is going to review it, a nurse is going to review it, and there is a good chance the error will be intercepted before it gets to a patient. But with an administration error, there is no one between the nurse and the patient to intercept the error. Intravenous errors can be especially dangerous. A significant error with a drug such as heparin or insulin can be catastrophic. For example, a physician ordered a heparin bolus dose of 4,000 units to be followed by an infusion of 890 units per hour.14 In this case, the 4,000 unit bolus dose was given appropriately, but the nurse misinterpreted the order and programmed the infusion device to deliver 4,000 units per hour and not 890 units per hour for continuous infusion. In this instance, the smart pump alerted the nurse.14 This is the kind of error that anyone could make at any time. Most of the time clinician errors have little consequence, but this kind of error could be fatal. A study published in 2005 evaluated the impact of an early-generation device that was studied in the cardiac surgical intensive care units (ICU) and step-down units.15 In the control periods the alerts were turned off, and in the intervention periods the alerts were turned on. The evaluation was a time-series study, and the incidence of errors was found by chart abstraction and log reports. As shown in Figure 11, the results showed basically no difference in the preventable ADE rate between the control and the intervention groups or in the nonintercepted potential ADE rate, i.e., the “near miss” rate, which is where a benefit had been expected to be found. However, the ideal world projections, in which nurses actually would use the drug library and not override warnings, showed that i.v. infusion safety systems could have prevented a large number of preventable and potential ADEs.15 Figure 1. Open in new tabDownload slide Smart Pump Results and Secondary Analysis.15 Seventy of intercepted PADEs in “Ideal World” use libary and no overrides. Figure 1. Open in new tabDownload slide Smart Pump Results and Secondary Analysis.15 Seventy of intercepted PADEs in “Ideal World” use libary and no overrides. This study had several issues. First, the pumps were initially set up so that nurses could either use the drug library or not use the drug library, and many times they ended up not using it, especially in conditions when risky drugs were being used. Second, when nurses did get important warnings, they often ignored the warnings and overrode the alerts. Such overrides were more extensive than investigators had anticipated.15 The study led to several factors being identified, both from the study data and from information from Taxis.16 One issue is that unstable patients at times need drugs quickly, and time demands result in increased workloads that can lead to shortcuts and override violations. In addition, there was clearly an issue of lack of perceived risk. Nurses did not believe that what they were doing was associated with any real risk. There also may have been issues with role-modeling: no feedback was provided to the nurses during the study, and there were no forcing functions built into pump use at the time when the study was done. The study clearly showed that serious i.v. infusion errors were frequent and could be detected using smart pumps.14 Prior to this study the magnitude of the problem was not well known, and at our institution we had little idea how common these errors were. Poor compliance was probably the main reason why the study did not show that smart pump use had any impact on the serious medication error rate or the preventable ADE rate. This suggests that behavioral and technological factors have to be addressed if smart pumps are to achieve their potential for improving medication safety.14 Of note, this was a first-generation smart pump, and many improvements have subsequently been made, including a number based on these results. Another study looked at the variation in infusion therapy practice in 100 hospitals.17 The results revealed unexpected, extensive variation in practice, much of which is quite clearly unnecessary and represents a safety risk. An analysis of smart pump drug libraries showed an average of 84 drugs per hospital, with an average of 8.5 names per drug across hospitals. The use of 8.5 different names for the same drug clearly has the potential to cause confusion during pump programming.17 The continuous dosing units used for the same drug also showed significant variability. These findings suggested that there is very substantial room for improvement, if hospitals will work together to standardize i.v. infusion practices across units and even across hospitals. A “toolkit” made available by the San Diego Patient Safety Consortium reports how this can be accomplished.18 Reaching consensus on infusion practices can be challenging, but standardization holds significant promise to improve infusion safety and quality of care. Dispensing errors and bar coding More recent studies have focused on dispensing errors, which are relatively common in hospital pharmacies because of the high volume of medications dispensed. More than 44,000 errors occur annually in a 735-bed hospital that dispenses about 6 million doses per year. Of these, investigators estimated that more than 9,500 errors annually had the potential to harm patients. Only about one-third of these serious errors are intercepted before medication administration.19 Dispensing and stocking errors can have serious consequences. In a recent tragic report from Indiana, three babies died in the neonatal ICU. Two strengths of heparin were available in this hospital, and a pharmacy technician filled the automated dispensing cabinet on the neonatal unit with the adult strength of heparin. When the doses of heparin were obtained, the adult-strength medication was given to five babies, three of whom died. The vials looked very much alike. It is very easy to see how that could happen, and it is hard to blame the nurses on the units who actually administered those drugs. A very recent study evaluated the impact of bar-coding drugs in the pharmacy and then checking them before they are sent to the patient care units. This study was done as a precursor to another study of the impact of bar coding in medication administration. Results showed that the dispensing error rate fell 31% after bar-code implementation, and the potential ADE rate (the “near miss” rate) fell 63% (Figure 22). This was a substantial reduction in the rate of potentially harmful errors. Moreover, projections indicated that the implementation of bar coding avoided about 13,500 medication dispensing errors and more than 6,000 ADEs. Bar-coding technology had a differential impact on different types of errors: a 58% reduction in wrong medication errors, a 53% reduction in wrong dose/strength, and complete elimination of wrong dosage form errors.20 Figure 2. Open in new tabDownload slide Dispensing Errors and Potential Adverse Drug Events: Before and After Bar-code Technology Implementation.20p<0.0001 (Chi-squared test). Figure 2. Open in new tabDownload slide Dispensing Errors and Potential Adverse Drug Events: Before and After Bar-code Technology Implementation.20p<0.0001 (Chi-squared test). Conclusion Preventing Medication Errors3 lays out a blueprint for change in medication safety. The report makes clear that providers have many opportunities to improve. Technologies, such as computerized order entry, bar coding and smart pumps, and computerized ADE monitoring, will undoubtedly play a key role, and institutions should be thinking seriously about implementing a number of these. The report also emphasizes how essential a culture change, combined with well-designed technologies, will be to achieve the next level of safety called for in the IOM report. Footnotes Based on the proceedings of a symposium held December 3, 2006, during the ASHP Midyear Clinical Meeting in Anaheim, CA, and supported by an educational grant from Cardinal Health. Dr. Bates received an honorarium for participating in the symposium and writing this article. Dr. Bates serves as a consultant to Cardinal Health and is on the clinical advisory board for IntelliDot. References 1 Institute of Medicine. To err is human: building a safer health system. L.T. Kohn, J. M. Corrigan, & M. S. Donaldson (Eds.). Washington, DC: National Academy Press, 2000 . 2 Institute of Medicine. Crossing the quality chasm: a new health system for the 21st century. Washington, DC: National Academy Press, 2001 . 3 Institute of Medicine. Preventing medication errors: quality chasm series. Washington, DC: National Academy Press, 2006 . 4 Bates DW, Leape LL, Cullen DJ et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA . 1998 ; 280 : 1311 –6. Crossref Search ADS PubMed 5 Bates DW, Teich JM, Lee J et al. The impact of computerized physician order entry on medication error prevention. J Am Med Inform Assoc . 1999 Jul–Aug; 6 (4): 313 –21. Crossref Search ADS PubMed 6 Chertow GM, Kuperman GJ, Burdick E et al. Guided medication dosing for inpatients with renal insufficiency. JAMA . 2001 ; 286 (22): 2839 –44. Crossref Search ADS PubMed 7 Peterson JF, Kuperman GJ, Shek C et al. Guided prescription of psychotropic medications for geriatric inpatients. Arch Intern Med . 2005 ; 165 (7): 802 –7. Crossref Search ADS PubMed 8 Kaushal R, Shojania KG, Bates DW. Effects of computerized physician order entry and clinical decision support systems on medication safety: a systematic review. Arch Intern Med . 2003 ; 163 (12): 1409 –16. Crossref Search ADS PubMed 9 Koppel R, Metlay JP, Cohen A et al. Role of computerized physician order entry systems in facilitating medication errors. JAMA . 2005 ; 293 (10): 1197 –203. Crossref Search ADS PubMed 10 Ash JS, Fournier L, Stavri PZ et al. Principles for a successful computerized physician order entry implementation. AMIA Annu Symp Proc . 2003 ; 36 –40. 11 Bates DW. Computerized physician order entry and medication errors: finding a balance. J Biomed Inform . 2005 ; 38 (4): 259 –61. Crossref Search ADS PubMed 12 Han YY, Carcillo JA, Venkataraman ST et al. Unexpected increased mortality after implementation of a commercially sold computerized physician order entry system. Pediatrics . 2005 ; 116 (6): 1506 –12. Erratum in: Pediatrics. 2006; 117(2):594. Crossref Search ADS PubMed 13 Phibbs CS, Milstein A, Delbanco SD et al. No Proven Link Between CPOE and Mortality. Pediatrics . 2005 ; 116 (6): 1506 –12. Crossref Search ADS PubMed 14 ISMP. “Smart” infusion pumps join CPOE and bar coding as important ways to prevent medication errors. ISMP Medication Safety Alert! 2002 Feb 6. 15 Rothschild JM, Keohane CA, Cook EF et al. A controlled trial of smart infusion pumps to improve medication safety in critically ill patients. Crit Care Med . 2005 ; 33 (3): 533 –40,679. Crossref Search ADS PubMed 16 Taxis K, Barber N. Causes of intravenous medication errors: an ethnographic study. Qual Saf Health Care . 2003 ; 12 (5): 343 –7. Crossref Search ADS PubMed 17 Bates DW, Vanderveen T, Seger DL et al. Variability in intravenous medication practices: implications for medication safety. Jt Comm J Pt Safety Qual . 2005 ; 31 (4): 203 –10. 18 San Diego Patient Safety Consortium. Getting Started Kit: Safe Administration of High-Risk IV Medications, Intra- and Inter-Hospital Standardization: Drug Concentrations and Dosing Units. How-to Guide, 2006. www.cardinalhealth.com/clinicalcenter/ (accessed 2007 Jan 24). 19 Cina JL, Gandhi TK, Churchill W et al. How many hospital pharmacy medication dispensing errors go undetected? Jt Comm J Qual Patient Saf . 2006 ; 32 (2): 73 –80. Crossref Search ADS PubMed 20 Poon EG, Cina JL, Churchill W et al. Medication dispensing errors and potential adverse drug events before and after implementing bar code technology in the pharmacy. Ann Intern Med . 2006 ; 145 (6): 426 –34. Crossref Search ADS PubMed Copyright © 2007. American Society of Health-System Pharmacists, Inc. All rights reserved.