SEE RELATED ARTICLE, P. 483.
[Ann Emerg Med. 2008;52:493-495.]
Accurately and completely reconcile medications across the continuum of care.
—The Joint Commission, 2008 National Patient Safety Goals, Hospital Program1
Your elderly patients are on medications you don't know what they are, their other doctors don't know what they are, they don't know what they are, and the neighbor who provided them doesn't know what they are.
—Paraphrased from a lecture by John W. Rowe, MD, geriatrician, Harvard Medical School, circa 1995
In this issue of Annals, Miller et al2 present a detailed view of medication reconciliation as carried out in a large rural medical center. They selected admitted trauma patients and evaluated medication lists compiled by the trauma team, an admitting nurse, and a clinical pharmacist. Using the clinical pharmacist-generated list as their criterion standard, the authors calculated that the lists created by the trauma team and admitting nurse were correct in all details only 15% of the time. The clinical pharmacist uncovered actual medication errors for 4% of patients, including one who experienced an adverse drug event. The authors conclude that medication histories obtained for these admitted trauma patients were “quite inaccurate.”
Inaccurate medication lists can have significant repercussions. Essential medications may be missed. Unintended medications may be taken. Doses, routes, and timing may all be mistaken and potential pharmaceutical interactions may go unrecognized until too late. At the time of significant transitions, such as hospital admission or discharge, incorrect medications may find their way onto the list, and suddenly the patient is regularly taking a proton-pump inhibitor or a daily aspirin. Accurate medication reconciliation is clearly an appropriate goal.
What is immediately striking from this study, though, is how remarkably difficult accurate and complete medication reconciliation turned out to be. This study evaluated 373 patients during 13 months. The authors excluded 299 subjects who were taking no prescription medications and who would typically be considered “reconciled,” skewing their results. Inclusion of these patients would have instantly lifted the accurate reconciliation rate to approximately 50%. Many patients presenting reconciliation challenges were also excluded, including those who could not name a pharmacy, who obtained medications by mail order, or who were taking samples from a physician's office. To put the sample size in perspective, the same hospital ED treated more than 70,000 patients from July 2005 to June 2006.3 In evaluating a small subset of less than 1% of total patients, this study provides a window on the gritty details of medication reconciliation.
The production of an initial medication list took 3 distinct steps. (I've added times and costs in a back-of-the-envelope estimate. These were not provided in the article.)
Step 1: The trauma team attempted to obtain a medication history, ideally from the patient, but alternatively from the family, chart, and primary care physician.
Step 2: An admitting nurse used the trauma team notes and other items in the medical record to develop a medication list. A similar system exists in many hospitals for admitted patients. For a patient receiving medications, it likely takes about 10 minutes per patient. At $30 per hour for a registered nurse, this is $5.00 per patient.4
Step 3: During the next few days, the clinical pharmacist interviewed family and physicians and called outpatient pharmacies to clarify the medication history. This process produced a criterion standard medication list. At a ballpark estimate of 30 minutes, this represents $25 per patient.5 The laborious and time-consuming nature of this step, along with Monday to Friday staffing, meant that 76 patients, or more than 20% of the sample who were taking medications, were discharged or died before completion of pharmacist review.
The study's results depend on the accuracy of the criterion standard pharmacist medication list. Conflicts among the 3 sets of lists serve mostly to point out the challenges in achieving accuracy. Many medication lists may exist and multiple lists inevitably conflict. There is what the primary physician considers the list, what pharmacies have on record, what the patient states is being taken, what the family states is being taken, and what the patient actually takes. When the last is not readily available, it is not at all evident which of the others should be the criterion standard for medication reconciliation, but one must be chosen or constructed as a guide for continuing therapy. Unfortunately, there is nothing particularly compelling to the idea that what the clinical pharmacist obtains through telephone calls firmly resembles what the patient actually takes.
The authors provide a sense of the barriers to the seemingly simple task of producing a medication list: inaccurate histories, use of multiple pharmacies and mail-order pharmacies, use of samples, foreign purchases, inability to spell or pronounce medications, use of herbals, illicit use, fear, confusion, anxiety, and forgetfulness. Add to this the rapid introduction of new medications and formulary changes and a fuller picture develops. Knowing that the admitting nurse would revise the medication list and that a clinical pharmacist would follow up, the trauma team may have deliberately chosen to focus on other essential tasks, reducing the applicability of the findings across sites and suggesting an unintended consequence of additional staff assigned to reconcile medications.
An estimated 40 minutes of nurse and pharmacist time, at a cost of $30.00 per patient, suggests the high costs of proper medication reconciliation. Fully implemented medication reconciliation would require a similar approach for many more patients, but it would also have to be more efficient, creating final lists before patient discharge. For a 50,000-visit emergency department, with 35% of patients requiring a medication reconciliation—the percentage taken from this study, excluding both those not receiving prescription medications and those excluded for one reason or another—and using the ballpark figures above, this would require 2,900 hours of nursing time and 8,750 hours of pharmacist time, or one-and-a-half full-time nurses and 4 full-time pharmacists for only the generation of initial medication lists. For admitted patients alone, assuming a 15% admission rate and 55% of patients receiving prescription medications, medication reconciliation would absorb 680 nursing hours and 2,060 pharmacist hours. Cutting the assumption of 30 minutes of pharmacist time per reconciliation in half reduces the hours markedly, but the totals remain prohibitive and do not include the time of the pharmacists on the other end of the telephone calls.
These estimates neglect additional opportunity costs. The trauma team members involved in medication reconciliation are not available for the next trauma. The admitting nurse focused on medication lists is not available to assist at the bedside. The pharmacist focused on pharmacy follow-up is not available to review current orders. All of these people are in short supply and wise use of their time is essential.
There are also subtle dangers in producing multiple medication lists. Many patients already carry medication lists with them. Providing them with new and separate lists may at once introduce inaccuracies and dissuade them from maintaining their own lists. Additional medication information of questionable validity also competes for clinicians' limited attention. Furthermore, the demands of medication reconciliation create an incentive to reduce prescription changes and make only recommendations, pushing the actual prescribing and reconciliation back on a potentially difficult-to-access and likely overwhelmed primary care physician.
Ideally, with complete establishment of medication reconciliation, one list would follow the patient from place to place and a criterion standard would be established. In a disjointed medical system, however, a potpourri of confused lists, with cross-outs and add-ons, follows many patients from place to place. Other patients pull sheaves of lists from pockets and purses, along with old discharge instructions and lunch receipts, and hand them over.
Medication reconciliation is a process that touches on most aspects of medical care. One challenge of the current approach, through the National Patient Safety Goals, is that it becomes a process suddenly owned by hospitals and inpatient institutions. Miller et al2 point out that reconciliation requires the cooperation of patients, physicians, hospitals, and pharmacists. The involvement of payers, in whose databases much information also resides, could make the mandate of reconciliation more powerful. Success will be evident when all participants, and most notably patients and clinicians, adopt an approach because it helps them in their daily lives.
The laudable goal of medication reconciliation is hard to argue. Miller et al2 point out both the need and the challenges. The Joint Commission has recognized some of the problems in developing standards for 2009 that distinguish between long-term and short-term medication prescription.6 Risks and limitations abound, however, and we need to avoid rosy assumptions about costs and benefits. Extensive mandates without supportive evidence and realistic potential for implementation risk wasting time, increasing costs, compounding errors, and squandering good will. At the bedside and in national policy, the challenge is to develop broad systems that track medications in a manner that recognizes the potential benefits of reconciliation without simultaneously negating them through increased cost and unanticipated risk.
FULL TEXT
