Outpatient Prescription History in the Hospital
Introduction: Historical Perspective
In February of 2001, the nations three largest pharmacy benefit managers (PBMs) Advance PCS, Express Scripts, and Medco Health Solutions jointly founded RxHub, an electronic information exchange, with the mission of promoting electronic prescribing and improving medication safety. Each of these organizations were pursuing their own initiatives for electronic prescribing but found that the most compelling factor in physician adoption was to represent a critical mass of patients in a physicians practice. By coming together and pooling their lives, this problem could be solved. This landmark collaborative effort created a single point of contact for ambulatory care e-prescribing technology vendors seeking to conduct transactions with multiple PBMs.
The major strategies for preventing medication errors and adverse drug events (ADEs) include solutions that can improve communications, make knowledge more readily accessible, assist with calculations, perform checks in real-time, and provide decision support. Nearly half of serious medication errors have been found to result from the fact that clinicians have insufficient information about the patient and the drug. Although it has been shown that clinical decision support provided by information technology tools can have great impact on quality, the success of such tools is essentially founded upon availability and access to information. Information technology solutions function merely as the delivery vehicles for information, yet it is the information itself that is the real valueresulting in enhanced clinical outcomes. The lack of access to comprehensive, up-todate information on a patients outpatient medications contributes to medication errors and ADEs in the hospital and after discharge.
Medication errors and ADEs have become among the countrys most serious health care problems, accounting for over 770,000 injuries or deaths each year in hospitals, and costing up to $5.6 million per year, per hospital, depending on size.
Having outpatient prescription claims history available at the time of patient admission to a hospital not only helps prevent errors, but improves intake efficiency and provides a unique window into patient compliance.
Medication Errors: The Role Of Outpatient Prescription History
Much has been written about medication errors since the release of the first of two major reports on medical errors by the Institute of Medicine in November of 1999. For a general overview of the problem, readers are referred to these reports,1,2 as well as to numerous individual studies measuring specific aspects of the problem,3-5 estimating associated costs,6,7 and quantifying the benefits of specific interventions.8,9 Of note, the available data on medication errors is overwhelmingly from inpatient settings, with studies from ambulatory settings only recently beginning to appear.10-14
While disagreement exists regarding the true magnitude of the medication error problem,15 few would argue that the problem is not serious or that interventions to correct it are not needed. Indeed, several states most notably California have enacted legislation requiring health care institutions to implement error prevention strategies, with an emphasis on technology-based interventions. (For a recent review of the role of information technology in preventing medical errors, see Bates.8) Similar legislation has been proposed at the federal level. Health care institutions are also feeling pressure from nonlegislative sources, including accreditation bodies such as the Joint Commission on Accreditation of Hospital Organizations, medicaquality certification organizations such as the National Committee on Quality Assurance, and employer consortiums such as Leapfrog. Lastly, reimbursement-based incentives to deploy error prevention tools are also being actively considered, particularly by the Center for Medicare and Medicaid Services.
It should thus be no surprise that hospitals have begun to invest heavily in medication error prevention technologies, including computerized physician order entry, pharmacy robots, dispensing cabinets, bar coded medication administration, and nursing documentation systems. These technology interventions each address specific components of the medication ordering/dispensing administration documentation cycle, and have variably well-characterized benefits.
While the error prevention benefits of providing outpatient prescription claims to hospital clinicians seem obvious, few studies have examined this directly. One recent study did actually look at the clinical effect of making prescription claims data available to physicians during patient encounters, though these were clinic rather than hospital encounters.16 Nonetheless, the results are quite relevant here. In this study conducted at the Henry Ford Health System (HFHS) six months of prescription claims data was printed and attached to the chart at the time of a scheduled clinic visit for 105 patients seen by staff physicians at HFHS primary care clinics (treatment group). Prescription claims were not provided for clinic visits for a similar group of 126 patients (control group). The mean age in both groups was 61 years old, and the mean number of drugs used at the time of the scheduled appointment was 5.5 in each group. The primary focus of this study was the impact of prescription claims data on physician detection of patient noncompliance, and no attempt was made to evaluate impact on medication errors. Even so, the results demonstrated that physicians were significantly more likely to make drug regimen changes when prescription claims data was available to them at the time of a patient encounter than when it was not.16 Seventy-one percent of the patients in the treatment group had at least one drug change made during their clinic encounter versus only 22 percent in the control group. There were 1.3 drug changes per encounter in the treatment group versus 0.3 per encounter in the control group (p = 0.001). Specific types of drug changes that were more common in treatment group patients included dose changes (21 percent versus 7 percent), drug additions (42 percent versus 14 percent), and drugs discontinued (15 percent versus 4 percent).16
The higher frequency of drug changes in the Henry Ford treatment group suggests that it is easier for physicians to assess whether patients are being treated adequately when a comprehensive medication list is available during a clinic appointment.16 Prescription claims data could arguably be of even greater importance to physicians evaluating patients during an ER visit or hospital admission. For instance, the physicians in the Henry Ford study were seeing patients in staff model HMO clinics, where they had access to chart-based medication lists whether or not prescription claims data was made available. In contrast, when patients are seen in a hospital, outpatient charts are much less often available. The Henry Ford study may thus actually underestimate the clinical benefit of prescription claims data for patients seen in a hospital setting.
It certainly makes sense that prescription claims data would be useful to hospital clinicians and could help them to avoid serious medication ordering mistakes. At least half of preventable medication errors are the result of a faulty physician order,9,17 and order entry is exactly the time when prior medication history is of most relevance.18,19 An inaccurate or incomplete listing of current outpatient medications may result in the inadvertent discontinuation of an important drug upon hospital admission, and increases the likelihood of unanticipated drug-drug interactions resulting from new drugs ordered after admission. Similarly, incomplete information about prior outpatient prescriptions can increase the likelihood of a repeat adverse reaction to a drug ordered in the hospital (e.g., the same drug might have been prescribed some months earlier and then discontinued as a result of poor tolerance patients frequently do not report these events on admission or, if they do, they may be unable to remember the name of the drug that they reacted to).
It is also common for patients to be confused about dosing instructions and/or dose forms. The problem of intake medication history requiring clarification is nicely illustrated by the following example, taken from Nester:20
A 67-year-old male has [admission] orders for several medications: isosorbide 20 mg bid, fluticasone inhaler prn, furosemide 80 mg once daily, bumetanide 1 mg once daily, and several other medications. The pharmacist must clarify the isosorbide as mononitrate or dinitrate and as sustained release or regular release The pharmacist is told that the inhaled corticosteroid is ordered on an as-needed basis because that is how the patient takes it at home. Also, the pharmacist must clarify whether the patient is really taking two loop diuretics
The above scenario is not simply a matter of anecdote or conventional clinical wisdom; rather, scenarios like this have been quantified by formal studies. In fact, the frequency of incomplete admission medication histories has been demonstrated to be quite large. For instance, Lau compared the number and types of medications recorded in hospital admission records with those in outpatient pharmacy records for 304 patients (mean age 71 years) admitted to a general internal medicine service.21 Discrepancies were resolved by patient interview. The admission records included 1,196 drugs. The pharmacy records revealed an additional 518 drugs that had not been recorded in the admission records, 410 of which were confirmed to be in active use. Drugs missing from the admission records included such important agents as NSAIDs, cardiovascular agents, H2 blockers, and oral antidiabetics (see Figure 1).
Sixty-one percent of patients had at least one drug missing from their admission record, and 33 percent had two or more missing drugs (see Figure 2).
Figure 1. Most commonly used medications and frequency and percentage of omission errors in 304 patients admitted to a general internal medicine service (from Lau21).
Figure 2. Distribution of number of omission errors per patient in the hospital medical records of 304 patients (from Lau21).
Improving the Accuracy and Completeness Of Medication History Intake
A significant body of literature already exists regarding the value of obtaining a more accurate outpatient medication history upon admission to the hospital by means other than claims data. Several studies from the hospital pharmacy domain have examined the clinical and financial impact of having pharmacists augment a patients admission medication history usually by conducting an intense medication history interview and by contacting community pharmacies when further clarification was needed about dispensed prescriptions.
These studies leave little doubt that clinical pharmacists given adequate time obtain more accurate and reliable medication histories from patients than do other clinicians operating under normal time constraints. A recent study by Nester and Hale at Wichita State University is illustrative.20 One hundred adult patients admitted to a 760-bed tertiary care facility were alternately assigned to pharmacist (treatment) versus nurse (control) medication history interviews. Clinical interventions occurred in 34 percent of the treatment group patients versus 16 percent of the control group (p < 0.001). There were also nearly three times as many mean interventions per patient in the treatment group (0.60 versus 0.22, p = 0.027). These clinical interventions included identification of route or dosage discrepancies (5 versus 2); identification of incomplete or missing orders (14 versus 5), and; therapeutic duplications avoided (3 versus 0).20
In spite of the remarkable clinical value delivered by having pharmacists conduct medication history interviews upon hospital admission, implementing such a service remains the exception: only 3 percent of U.S. hospitals currently do this.22 The reason, quite simply, is cost. Obtaining a good medication history takes a lot of time. Paying highly trained clinical pharmacists to routinely conduct medication interviews is expensive. Hospitals remain reluctant to launch this service, in part because it is not directly reimbursed in most states. Moreover, clinical pharmacists are in short supply. Even if a hospital were to implement this service, it would probably only include patients actually admitted to the hospital. Emergency room patients who dont get admitted, outpatient surgery patients, and possibly others for whom obtaining accurate medication histories is as important as it is elusive, would thus not benefit from the service.
Prescription Claims Data to the Rescue
Making dispensed prescription history available to hospital clinicians via claims data is a practical alternative to pharmacist-conducted medication interviews. While prescription claims data cannot fully replace the value of pharmacist-conducted interviews, the data can provide much of the same information with much less effort and expense. The Henry Ford study albeit from an outpatient setting confirms that prescription claims data alone does provide substantial clinical value.16 The increase in clinical pharmacy interventions seen in the Henry Ford study in the outpatient setting with claims-augmented history was quite similar to that seen in the inpatient setting in the Wichita State study with pharmacist-augmented history.16,20 Also, the pharmacists in the Wichita State study still had to contact patients community pharmacies to clarify home medications for 24 percent of the patients they interviewed.20 Many of these calls would probably have been unnecessary had prescription claims history been available. In other words, prescription claims data could be a useful adjunct even for those hospitals that already have a pharmacist-conducted medication interview service.
Further Support for the Importance Of Prescription History to Error Prevention
The view that providing better access to prescription history will help prevent errors in the hospital is supported by other informatics and decision support research. For instance, just-in-time models of physician decision making make it clear that increasing the availability of prior patient data including medication history is critical for reducing errors.19,23 Studies from the laboratory data domain have also left little doubt regarding the impact of increasing the availability of prior patient test results on current test ordering behavior.24-26 Lastly, several decisionmaking simulation studies have shown that even minor alterations in how prior patient data is displayed can have a major impact on physician decision-making performance.27,28
There is also indirect support from the adverse drug event literature itself. In particular, at least one ADE study has characterized the role of prior patient history in current medication errors. In this study the largest ever of the systematic causes of preventable ADEs in hospitalized patients inadequate access to patient data was the second most common of 16 types of system failures, accounting for 18 percent of events.18 Although individual instances of inadequate patient data access were not further subcategorized or reported in this study, one of the lead authors estimates that one-fifth were specifically due to incomplete outpatient medication history.
Improving the Efficiency Of Medication History Intake
When a patient is admitted to the hospital, nurses, physicians, and pharmacists all spend precious time obtaining or clarifying medication history. In the Wichita State study, the study pharmacists assigned to conducting medication interviews spent 13.4 + 6.7 minutes per interview for the treatment group patients, while nurses spent 24.3 + 19.8 minutes per control group patient.20 Interestingly, dispensing pharmacists had to spend an additional 6.0 + 6.5 minutes reconciling home medication orders for treatment group patients even after the initial pharmacist conducted the interview, and 14.4 + 9.9 additional minutes for nurse-interviewed patients. The total professional staff time not including physicians spent acquiring medication history in the control group was thus nearly 40 minutes per patient.20 Another study found that nurses spent an average of 30 minutes per patient obtaining medication history.29
There is no data available on how much time physicians spend reconciling medications on admission, nor is there information on how much time would be saved by having prescription claims history available during the medication intake process. However, in the Henry Ford study cited above: 46 percent of the physicians who received prescription claims history at the time of a patients clinic encounter reported that having the claims history saved them 1 to 3 minutes per visit, and 3 percent said that they saved 4 to 6 minutes.16 While this data comes from the ambulatory care setting and is based on physician self-report, it suggests that prescription claims data does indeed improve the efficiency of medication history intake.
Assessing Patient Compliance With Chronic Medications
Patient noncompliance with prescribed therapy is a major clinical problem, particularly when the noncompliance is undetected. Consider the example of a noncompliant patient with a poorly controlled condition such as hypertension, diabetes, or congestive heart failure whose treating physician is unaware of the noncompliance. Rather than increasing efforts to understand and resolve issues contributing to the noncompliance with existing drug therapy, the physician is likely to recommend unwarranted dosage increases or add new medications in an attempt to bring the condition under better control. While most physicians have the skills to ferret out noncompliance, they often are unable to take the necessary time to do this effectively. Prescription claims history provides a unique and powerful vantage point for quickly assessing patient compliance with chronic medications. By decreasing dependence on highly variable factors such as physician skill and available time, prescription claims data can dramatically improve the performance and reduce the variability of the assessment of patient noncompliance. The Henry Ford study provides strong support for this position: When the last six months of prescription claims history was available on the chart, physicians detected noncompliance in nearly a third of patients, whereas they detected none in the control group patients (30.5 percent versus 0 percent, p < 0.001).16 Pharmacist abstractors who had prescription claims data available to them for all of the patients in the study found noncompliance in nearly 60 percent of patients in both the treatment and control groups.
Cost Benefits Of Claims-Augmented Medication History
Available data from the ADE literature permits indirect estimation of potential cost savings that could result from making prescription claims history routinely available for patients admitted to the hospital. According to Bates, et al., preventable ADEs occur in 2 percent of admissions, resulting in average increased hospital costs of $4,685 per event, or about $2.8 million annually for a 700-bed teaching hospital.6 Roughly one-fifth of preventable ADEs are due to inadequate access to patient information.18 If, as noted above, one-fifth of these are specifically due to inadequate access to outpatient prescription history, then a fifth of a fifth or 4 percent of preventable ADEs could be avoided by improving access to outpatient prescription history. Using the $2.8 million annual cost of preventable ADEs to a 700-bed hospital, a 4 percent reduction in ADEs would produce a cost savings of $112,000. This estimate does not include other patients affected by the preventable ADE problem, such as emergency room patients who are not admitted or outpatient surgery patients. This estimate also does not consider potential ADE prevention benefits related to the improved ability to assess patient noncompliance. 16 Nor does it include benefits related to improved efficiency of medication history intake.16-20 If nurses, pharmacists, and physicians combined saved only 10 minutes per hospital admission, then a 700-bed facility with 30,000 admissions per year would save 300,000 minutes or 5,000 hours of valuable professional staff time. Even if this time savings didnt translate into full-time employee reductions, these clinicians could spend extra time delivering clinically important service rather than performing rote information retrieval. Moreover, having prescription claims data available will likely reduce the frustration typically associated with gathering outpatient medication history, and could improve job satisfaction.
The above estimate of cost savings of prescription claimsaugmented medication history intake derived indirectly from data from studies of preventable ADEs appears remarkably conservative when compared to more direct estimates of cost savings based on studies of pharmacist-augmented medication history intake. In fact, the financial and patient safety implications of these studies are nothing short of staggering: Pharmacist-conducted medication histories have been estimated to save $7 million and prevent 128 deaths per hospital per year.30 Even if prescription claims data alone provided only one-tenth of the benefit of pharmacist-conducted medication history, that still translates into $700,000 saved (and 13 deaths prevented) per hospital per year.
Endnotes
1 Committee on Quality of Health Care in America, Institute of Medicine. To Err is Human: Building a Safer Health System, ed. L.T. Kohn, J. M. Corrigan, and M. S. Donaldson, 1999. Washington, D.C.: National Academy Press.
2 Committee on Quality of Health Care in America, Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. 2001, Washington, D.C.: National Academy Press.
3 Bates, D. W., et al., Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA, 1995. 274 (1): p. 29-34.
4 Bates, D. W., L. L. Leape, and S. Petrycki. Incidence and preventability of adverse drug events in hospitalized adults. J Gen Intern Med, 1993. 8 (6): p. 289-94.
5 Bates, D.W., Frequency, consequences and prevention of adverse drug events. J Qual Clin Pract, 1999. 19 (1): p. 13-7.
6 Bates, D.W., et al., The costs of adverse drug events in hospitalized patients. Adverse Drug Events Prevention Study Group. JAMA, 1997. 277 (4): p. 307-11.
7 Bates, D.W., Costs of drug-related morbidity and mortality: enormous and growing rapidly. J Am Pharm Assoc (Wash), 2001. 41 (2): p. 156-7.
8 Bates, D.W. and A. A. Gawande, Improving safety with information technology. N Engl J Med, 2003. 348 (25): p. 2526-34.
9 Bates, D.W., et al., The impact of computerized physician order entry on medication error prevention. J Am Med Inform Assoc, 1999. 6 (4): p. 313-21.
10 Gandhi, T. K., et al., Drug complications in outpatients. J Gen Intern Med, 2000. 15 (3): p. 149-54.
11 Gandhi, T. K., et al., Adverse drug events in ambulatory care. N Engl J Med, 2003. 348 (16): p. 1556-64.
12 Gurwitz, J. H., et al., Incidence and preventability of adverse drug events among older persons in the ambulatory setting. JAMA, 2003. 289 (9): p. 1107-16.
13 Petty, B. G., Identifying and reducing complications of outpatient medications. J Gen Intern Med, 2000. 15 (3): p. 207-8.
14 Shaughnessy, A. F. and R. O. Nickel. Prescription writing patterns and errors in a family medicine residency program. J Fam Pract, 1989. 29 (3): p. 290-5.
15 McDonald, C. J., M. Weiner, and S. L. Hui. Deaths due to medical errors are exaggerated in Institute of Medicine report. JAMA, 2000. 284 (1): p. 93-5.
16 Bieszk, N., et al., Detection of medication nonadherence through review of pharmacy claims data. Am J Health Syst Pharm, 2003. 60 (4): p. 360-6.
17 Bates, D. W., et al., Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA, 1998. 280 (15): p. 1311-6.
18 Leape, L. L., et al., Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA, 1995. 274 (1): p. 35-43.
19 Elson, R. B., J. G. Faughnan, and D. P. Connelly. An industrial process view of information delivery to support clinical decision making: implications for systems design and process measures. J Am Med Inform Assoc, 1997. 4 (4): p. 266-78.
20 Nester, T. M. and L. S. Hale. Effectiveness of a pharmacist-acquired medication history in promoting patient safety. Am J Health Syst Pharm, 2002. 59 (22): p. 2221-5.
21 Lau, H. S., et al., The completeness of medication histories in hospital medical records of patients admitted to general internal medicine wards. Br J Clin Pharmacol, 2000. 49 (6): p. 597-603.
22 Bond, C. A., C. L. Raehl, and T. Franke. Clinical pharmacy services and hospital mortality rates. Pharmacotherapy, 1999. 19 (5): p. 556-64.
23 Chueh, H. and G.O. Barnett. Just-in-time clinical information. Acad Med, 1997. 72 (6): p. 512-7.
24 Connelly, D. P., B. H. Sielaff, and K. E. Willard. A clinicians workstation for improving laboratory use. Integrated display of laboratory results. Am J Clin Pathol, 1995. 104 (3): p. 243-52.
25 Tierney, W. M., et al., Computerized display of past test results. Effect on outpatient testing. Ann Intern Med, 1987. 107 (4): p. 569-74.
26 Bates, D. W., et al., A randomized trial of a computer-based intervention to reduce utilization of redundant laboratory tests. Am J Med, 1999. 106 (2): p. 144-50.
27 Willard, K. E., D. P. Connelly, and J. R. Johnson. Radical improvements in the display of clinical microbiology results: a Web-based clinical information system [published erratum appears in Am J Med 1997 Jul; 103 (1): 86]. Am J Med, 1996. 101 (5): p. 541-9.
28 Elson, R. B. and D. P. Connelly. The impact of anticipatory patient data displays on physician decision making: a pilot study. Proc AMIA Annu Fall Symp, 1997: p. 233-7.
29 Rozich, J. D. and R. K. Resa. Medication safety: one organizations approach to the challenge. J Clin Outcomes Manage, 2001, 8: p. 27-34.
30 Bond, C. A., C. L. Raehl, and T. Franke. Clinical pharmacy services, hospital pharmacy staffing, and medication errors in U.S. hospitals. Pharmacotherapy, 2002. 22 (2): p. 134-47.
