- Email: [email protected]
A Medical Resident–Pharmacist Collaboration Improves the Rate of Medication Reconciliation Verification at Discharge
The Joint Commission Journal on Quality and Patient Safety Case Study in Brief
A Medical Resident–Pharmacist Collaboration Improves the Rate of Medication Reconciliation Verification at Discharge Daniel A. Caroff, MD; Therese Bittermann, MD; Charles E. Leonard, PharmD, MSCE; Gene A. Gibson, PharmD; Jennifer S. Myers, MD
M
edication reconciliation, designated as a Joint Commission National Patient Safety Goal in 2004, became a requirement for hospital accreditation in 20061 in response to concerns regarding the prevalence of preventable adverse drug events (ADEs) during patient transitions. A systematic review of 22 studies that evaluated medication history errors on hospital admission showed that up to two thirds of patients had at least one medication history error, half of which were believed to be clinically important.2 It is estimated that ADEs account for approximately 280,000 hospital admissions in the United States annually.3 ADEs after hospital discharge may be responsible for two thirds of postdischarge complications,4 and 24%– 27% of postdischarge ADEs may be preventable.4,5 Despite efforts to improve compliance, the effectiveness of inpatient medication reconciliation to prevent ADEs has been challenged by the inconsistency of the process within institutions.6 It has been demonstrated that pharmacists obtain more accurate medication histories than do physicians2,7,8 and might perform more accurate medication reconciliation.9 Pharmacist participation in discharge planning has also been shown to improve identification of medication errors with potential for harm.10–12 As a result, institutions have attempted to determine the most effective way to perform medication reconciliation in collaboration with inpatient pharmacists while best using the skill sets of clinical team members. Although some quality improvement (QI) interventions have demonstrated impressive reductions in readmission rates as a result of enhanced error recognition by pharmacists,13,14 others have not led to meaningful change,15,16 making their use of hospital staffing and financial resources difficult to justify.17 Challenges in communication, collaboration, and task allocation have also created barriers to medication reconciliation.18–20 At our institution, we encountered similar challenges in medication reconciliation at discharge. Our standard discharge process on two general medicine units involved creation of October 2015
Article-at-a-Glance Background: At the Hospital of the University of Penn-
sylvania (Philadelphia), it is standard practice to perform medication reconciliation at patient discharge. Although pharmacists historically were available to assist resident physicians in the discharge medication reconciliation process, the process was never standardized. An internal review showed a 60%–70% rate of pharmacist review of discharge medication lists, potentially enabling medication errors to go unnoticed during transitions of care. In response, a medical resident– and pharmacist–led collaboration was designed, and a pre-postintervention study was conducted to assess its effectiveness. Methods: A new work flow was established in which housestaff notified pharmacists when a preliminary discharge medication list was ready for reconciliation and provided access for pharmacists to correct medication errors in the electronic discharge document with physician approval. Length of stay, average time of day of patient discharge, and readmission data were compared in the pre- and post-intervention periods. Results: There were 981 discharges in the preintervention period and 1,207 in the postintervention period. The rate of pharmacist reconciliation increased from 64.0% to 82.4% after the intervention (p < .0001). The average number of errors identified and corrected by pharmacists decreased from 0.979 to 0.862 per discharge (p < .0001). There was no significant change in readmission rates or time of discharge after the intervention. Conclusions: Redesigning the discharge medication reconciliation process in a teaching hospital to include a review of medical resident discharge medication lists by pharmacists provided more opportunities for discharge medication error identification and correction.
Volume 41 Number 10
Copyright 2015 The Joint Commission
457
The Joint Commission Journal on Quality and Patient Safety the discharge medication list by the primary medical resident, finalization of discharge instructions and printing of paper prescriptions by the medical resident, and, lastly, review of the medication list by pharmacy and nursing staff before discharging the patient. This final step, however, was often limited by poor or inefficient communication between team members. An internal review in March 2012 showed that pharmacist medication review occurred in only 60%–70% of discharges. Thus, there was concern that discharge medication errors would go unnoticed and uncorrected. In response, a medical resident– led QI initiative was designed to improve the identification and correction of medication errors at hospital discharge. We conducted a pre-post-intervention study to assess its effect on the rate of pharmacist-verified medication reconciliation at hospital discharge.
Methods Study Design A quasi-experimental study using multiple noncontiguous data periods (preintervention period: February 2012–July 2012; post intervention period: August 2012–January 2013) was designed. The University of Pennsylvania’s Institutional Review Board approved this study.
Setting and Participants This study occurred at the Hospital of the University of Pennsylvania, a 789-bed tertiary care teaching hospital in Philadelphia. Eight inpatient teaching services on two hospital units participated, each composed of one internal medicine resident, two medical interns, one to three medical students, and either a hospitalist or medical subspecialty attending. All patients discharged on Monday through Friday from these services were included. Patients who left against medical advice or died were excluded, as were patients discharged on a weekday national holiday because of limitations in pharmacy staffing. Four unit-based clinical pharmacists (UBCPs), who were responsible for delivering comprehensive pharmaceutical and patient-centered care, participated in the study. The ratio of UBCPs to inpatient beds was 1:20 on the two inpatient units that were studied. The UBCPs performed medication reconciliation on admis sion, provided medication counseling to patients, and reviewed inpatient medication orders. At hospital discharge, the UBCPs were expected to review the discharge medication list and counsel patients and families about important changes and possible adverse effects. UBCPs were required to document all interventions in the electronic health record (EHR) system (Allscripts, Chicago). 458
October 2015
Intervention The intervention consisted of three components, as follows. 1. Policy for Medical Residents to Obtain Pharmacist Approval. A policy was introduced that required medical resi dents to obtain pharmacist approval of the discharge medica tion list before any weekday discharge. An order within the EHR allowed medical residents to notify UBCPs when a patient’s discharge medication list was ready for review. Medical resident education on this new work flow was provided during daily meetings, in automated e-mails at the start of each new rotation, and in flyers distributed throughout the two inpatient units involved. We did not have a “hard stop” function to prevent discharges without UBCP review. 2. Ability of Pharmacists to Edit the Discharge Medication List. UBCPs were enabled to directly edit the discharge medication list, accessible through the EHR. Three months before the intervention, two 30-minute training sessions on the electronic discharge software were held for all UBCPs. Sessions were led by a medical resident, and a slide set was provided by our information technology staff. When reviewing a discharge medication list, pharmacists corrected any errors after verbal confirmation from the patient’s primary resident. Documentation of all errors and corrections were logged into the EHR system. After pharmacist review, the medical resident was notified and given approval to write the discharge order. All necessary prescriptions were then printed and signed by the medical resident. 3. Nurse’s Final Review of Medication List. A final review of the medication list was performed by the patient’s nurse at the bedside immediately before discharge. This last step was already part of the standard discharge process.
Demographic and Clinical Data The following basic demographic and clinical data were collected for all patients included in the study: date of birth, date of admission, date of discharge, inpatient medical service, length of stay (LOS), risk of readmission (as determined by two or more hospital admissions during the previous 12 months 21), and pharmacist documentation of discharge medication errors. The study group designated immunosuppressants, antimicrobials, anticoagulants, insulins, and controlled substances as “high-risk” medications on the basis of the most common ADEs for which older adults seek treatment in the emergency department.22 Errors were categorized as (1) medication omitted, (2) incorrect medication, (3) incorrect dose, (4) incorrect frequency, (5) incorrect formulation, (6) medication unintentionally listed or prescribed (medication that a patient “does not take”), (7) medication duplication, and (8) prescription-writing Volume 41 Number 10
Copyright 2015 The Joint Commission
The Joint Commission Journal on Quality and Patient Safety errors (for example, medication scheduled instead of “as needed” or use of unapproved abbreviations) (Table 1, right). The intervention began on August 27, 2012. Data were gathered during five time periods: two preintervention periods (February 1, 2012–March 30, 2012, and July 1, 2012–July 30, 2012), one period immediately following the intervention (August 27, 2012–October 18, 2012), and two subsequent postintervention periods (October 22, 2012– November 16, 2012, and January 14, 2013–February 16, 2013). All data were collected by two of the authors [D.A.C., T.B.] using EHRs. Unit-specific rates of hospital readmission and average time of day of patient discharge were obtained from institutional databases.
Data Analysis Raw data were combined into preintervention and postintervention groups. Tests of proportion equality were used to calculate p values for differences in proportions. T-tests were used to calculate p values for differences in the means of continuous variables. These analyses were conducted using Stata/MP 13.1 (StataCorp, LP, College Station, Texas).
Results Demographic and Clinical Data In total, there were 981 discharges in the two preintervention periods versus 1,207 discharges in the three postintervention periods. Basic demographic and clinical data are shown in Table 2 (page 460). Overall, there were no appreciable differences in the mean age of patients in each group, in the proportion of inpatient services represented, and in the percentage of patients identified as at risk for readmission.
Outcomes After implementation of the intervention, a statistically significant increase in the number of patients discharged with UBCP verification of their medication reconciliation was noted (64.0% versus 82.4%, p < .0001; Table 3 [page 460]). Although there was no significant difference in the number of discharge documents with at least one error present, the mean number of UBCP–identified errors per discharge decreased from 0.979 to 0.862 (p < .0001). Table 3 details the types of errors found in both groups, and in most cases there was no significant change in the proportion of specific error types before and after the intervention. There was no difference in the number of highrisk medication errors identified after the intervention. A significant difference in the mean LOS was noted (7.15 days preintervention versus 6.17 days postintervention, October 2015
Table 1. Examples of Medication Error Types Error Type
Example
Omission
Patient takes morphine sulfate; however, drug missing from document
Drug name
Patient takes morphine sulfate; however, hydromorphone listed
Dose
Patient takes morphine sulfate 15 mg; however, 30 mg dose listed
Frequency
Patient takes morphine sulfate twice daily; however, listed as three times daily
Formulation
Patient takes morphine extended release; however, immediate release listed
Not on medication
Patient does not take morphine sulfate; however, it is listed in document
Prescribing
Patient takes morphine sulfate as needed; however, it is listed as scheduled
Duplication
Patient’s morphine extended release and MS Contin* are listed simultaneously
*Purdue Pharma (Stamford, Connecticut); time-released formulation of morphine sulfate.
p < .0051). There was no correlation between LOS and the likelihood of medication reconciliation (point-biserial correlations 0.0069 preintervention and 0.0008 postintervention, with p values for rejection of the critical t values .83 and .98, respectively). In addition, there was no change in the average time of discharge after implementation of the intervention. Finally, there was no significant difference in 7- and 30-day readmission rates after the intervention (Table 3).
Discussion
We demonstrated a statistically significant improvement in the standardization of the discharge medication reconciliation process through the design of a collaborative QI intervention between medical residents and clinical pharmacists. The rate of pharmacist discharge medication review increased from 62.0% to 82.4% after the intervention. Although we did not assess medication-related adverse events after discharge, we believe that our results show value added for the patients for whom their discharge medications were corrected and verified by a pharmacist. The remaining 17.6% of postintervention discharges that did not undergo pharmacist review might have been caused by imperfect compliance, miscommunication, or competing clinical obligations on the part of the house staff, pharmacists, or nursing staff. Further investigation of these issues may help improve the effect of the intervention. In addition to improvements in the rate of discharge medication reconciliation review, we also observed a significant decrease in the error rate of physician-authored discharge Volume 41 Number 10
Copyright 2015 The Joint Commission
459
The Joint Commission Journal on Quality and Patient Safety Table 2. Demographic and Clinical Data, Pre- and Postintervention Preintervention Feb–Jul 2012 N = 981
Postintervention Aug 2012–Jan 2013 N = 1,207
Mean age at discharge, years (± SD)
Inpatient service, N (%)
55 (± 18.2*)
55 (± 17.6)
General medicine
599 (61.1)
720 (59.7)
Hepatology
134 (13.7)
156 (12.9)
Pulmonary
117 (11.9)
139 (11.5)
Infectious disease
88 (9.0)
85 (7.0)
Nephrology
43 (4.4)
107 (8.9)
291 (29.7)
367 (30.4)
Number of high-readmission-risk patients
†
SD, standard deviation. * Sums to 100.1% because of rounding. † Patients are flagged as high readmission risk in the electronic health record if they have been admitted to the hospital ≥ 2 times in the 12 months previous to the admission.
Table 3. Measures of Discharge Medication Accuracy, Pre- and Postintervention Preintervention Feb–Jul 2012
Postintervention Aug 2012–Jan 2013
P Value
Mean length of stay (days ± SD)
7.15 ± 9.66
6.17 ± 6.62
< .0051
Total Discharges, N (%)
981 (100.0)
1,207 (100.0)
—
Discharges with medication reconciliation, N (%)
628 (64.0)
995 (82.4)
< .0001
7-day readmissions, N
68
77
.6254
30-day readmissions, N
296
328
.1548
Reconciliations completed, N (%)
628 (100.0)
995 (100.0)
—
Reconciliations with ≥ 1 pharmacist-identified error, N (%)
309 (49.2)
448 (45.0)
.1003
0.979 ± 0.142
0.862 ± 0.345
< .0001
615 (100.0*)
858 (100.0)
—
208 (33.8)
265 (30.9)
.2341 .0653
Discharges:
Readmissions:
Medication reconciliation:
Pharmacist-identified errors per reconciliation (mean ± SD) Errors, N (%) Total pharmacist-identified errors Omission Drug name
31 (5.0)
27 (3.1)
121 (19.7)
150 (17.5)
.2842
Frequency
54 (8.8)
87 (10.1)
.3818
Formulation
12 (2.0)
38 (4.4)
.0096
Not on medication
40 (6.5)
47 (5.5)
.4100
Prescribing
135 (22.0)
215 (25.1)
.1671
Duplication
14 (2.3)
29 (3.4)
.2147
127 (20.7)
182 (21.2)
.7940
Dose
High-risk medication †
* Sums to 100.1% because of rounding. † High-risk medications include immunosuppressants, antimicrobials, anticoagulants, insulins, and controlled substances.
460
October 2015
medication lists. There are several possibilities that could explain this finding. First, it is possible that an increased focus by both the hospital and the residency program led residents to recognize the importance of medical reconciliation and perform the task more carefully. It is also possible that medical residents changed their behavior and performed medication reconciliation more carefully, as they knew their discharge medication lists were being reviewed by a pharmacy colleague. Finally, it is possible that a pharmacist’s sensitivity for finding errors was decreased because of a larger number of documents being reviewed. Interestingly, before the intervention, there was concern that discharge medication errors made by medical residents might increase as a result of this type of intervention, given their knowledge that a pharmacist would be correcting their errors; fortunately, this did not occur. Although internal data have suggested that medication reconciliation by pharmacists reduces readmissions, our intervention did not reduce unit-specific readmission rates. Numerous other factors and concurrent institutional QI initiatives related to the discharge process could have affected the observed readmission rates and potentially confounded our intervention’s effect on this specific outcome. Importantly, the time of hospital discharge did not change before and after the intervention. We hypothesize that the time spent by pharmacists on additional discharge medication reconciliations was offset by earlier identification of patients to be discharged, and improved communication and coordination regarding discharge by the resident, pharmacist, and nurse as a result of the intervention. It is notable that this intervention did not require additional pharmacy staffing. Although it did require a small amount of pharmacist time for training, the positive impact on the discharge medication reconciliation process and likely prevention of outpatient ADEs affords this intervention at worst a cost-neutral economic impact and, very possibly, a net savings, although additional research is needed to validate this assumption. There are several limitations to this study. The quasi-experimental design is vulnerable to
Volume 41 Number 10
Copyright 2015 The Joint Commission
The Joint Commission Journal on Quality and Patient Safety confounding by a number of related QI initiatives and trends within our hospital, although this limitation was partially addressed by including multiple data periods both before and after the intervention. We did not record the total number of medications prescribed per patient, which, if significantly lower in the postintervention period, could conceivably have caused bias. We are unable to account for the pre- or postintervention rate of medication errors in the minority of patients who were discharged without pharmacist review. Findings may not be reproducible at other institutions or residency programs due to cultural or practical differences, such as limitations in pharmacy resources. Within the study hospital, however, in light of the positive results of this study, the intervention was expanded from the two original inpatient units to all inpatient services in the Department of Medicine, and the process has been sustained with positive feedback from both pharmacists and medical residents up through the present time. Being at the forefront of clinical care in teaching hospitals, medical and surgical residents are in an ideal position to identify QI and patient safety problems and design solutions to mitigate them.23,24 Resident-led initiatives are an ideal way to provide practical QI training and experience to trainees that meets the expectations of the Accreditation Council for Graduate Medical Education in systems-based practice.25 Furthermore, we suspect that the resident-led nature of this intervention minimized the frontline physician resistance to change that can be a barrier to successful QI project implementation.26
Conclusion
A collaborative medical resident–pharmacist intervention was effective in creating a sustained improvement in the discharge medication reconciliation process at the Hospital of the University of Pennsylvania. J Daniel A. Caroff, MD, formerly Internal Medicine Resident, De partment of Medicine, University of Pennsylvania, Philadelphia, is a Fellow in Infectious Disease, Brigham and Women’s Hospital and Massachusetts General Hospital, Boston. Therese Bittermann, MD, formerly Internal Medicine Resident, Department of Medicine, University of Pennsylvania, is a Gastroenterology Fellow, University of Pennsylvania. Charles E. Leonard, PharmD, MSCE, is Senior Research Investigator and Adjunct Assistant Professor of Epidemiology, Center for Pharmacoepidemiology Research and Training, Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania. Gene A. Gibson, PharmD, is Associate Director, Department of Pharmacy Services, Hospital of the University of Pennsylvania. Jennifer S. Myers, MD, is Associate Professor of Clinical Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania. Please address correspondence to Daniel A. Caroff, [email protected].
October 2015
References
1. The Joint Commission. Using medication reconciliation to prevent errors. Sentinel Event Alert. 2006 Jan 25;(35):1–4. Accessed Aug 26, 2015. www.joint commission.org/sentinel_event_alert_issue_35_using_medication_reconciliation _to_prevent_errors/. 2. Tam VC, et al. Frequency, type and clinical importance of medication history errors at admission to hospital: A systematic review. CMAJ. 2005 Aug 30;173(5): 510–515. 3. US Department of Health & Human Services, Office of Disease Prevention and Health Promotion (ODPHP). National Action Plan for Adverse Drug Event Prevention. Washington, DC: ODPHP, 2014. Accessed Aug 28, 2015. http://health.gov /hai/pdfs/ADE-Action-Plan-508c.pdf. 4. Forster AJ, et al. The incidence and severity of adverse events affecting patients after discharge from the hospital. Ann Intern Med. 2003 Feb 4;138(3):161–167. 5. Forster AJ, et al. Adverse drug events occurring following hospital discharge. J Gen Intern Med. 2005;20(4):317–323. 6. Clay BJ, et al. Results of a medication reconciliation survey from the 2006 Society of Hospital Medicine national meeting. J Hosp Med. 2008;3(6):465–472. 7. Reeder TA, Mutnick A. Pharmacist- versus physician-obtained medication histories. Am J Health Syst Pharm. 2008 May 1;65(9):857–860. 8. Carter MK, et al. Pharmacist-acquired medication histories in a university hospital emergency department. Am J Health Syst Pharm. 2006 Dec 15;63(24): 2500–2503. 9. Kushniruk AW, et al. Cognitive analysis of a medication reconciliation tool: Applying laboratory and naturalistic approaches to system evaluation. Stud Health Technol Inform. 2011;164:203–207. 10. Abdel-Qader DH, et al. Pharmacists’ interventions in prescribing errors at hospital discharge: An observational study in the context of an electronic prescribing system in a UK teaching hospital. Drug Saf. 2010 Nov 1;33(11):1027–1044. 11. Feldman LS, et al. Nurse-pharmacist collaboration on medication reconciliation prevents potential harm. J Hosp Med. 2012;7(5):396–401. 12. Allende Bandrés MÁ, et al. Pharmacist-led medication reconciliation to reduce discrepancies in transitions of care in Spain. Int J Clin Pharm. 2013;35(6):1083– 1090. 13. Eisenhower C. Impact of pharmacist-conducted medication reconciliation at discharge on readmissions of elderly patients with COPD. Ann Pharmacother. 2014;48(2):203–208. 14. Pal A, Babbott S, Wilkinson ST. Can the targeted use of a discharge pharmacist significantly decrease 30-day readmissions? Hosp Pharm. 2013;48(5):380–388. 15. Kripalani S, et al; PILL-CVD (Pharmacist Intervention for Low Literacy in Cardiovascular Disease) Study Group. Effect of a pharmacist intervention on clinically important medication errors after hospital discharge: A randomized trial. Ann Intern Med. 2012 Jul 3;157(1):1–10. 16. Kwan JL, et al. Medication reconciliation during transitions of care as a patient safety strategy: A systematic review. Ann Intern Med. 2013 Mar 5;158(5 Pt 2) :397–403. 17. Lu Y, et al. Quality improvement through implementation of discharge order reconciliation. Am J Health Syst Pharm. 2013 May 1;70(9):815–820. 18. Meguerditchian AN, et al. Medication reconciliation at admission and discharge: A time and motion study. BMC Health Serv Res. 2013 Nov 21;13:485. 19. van Sluisveld N, et al. Medication reconciliation at hospital admission and discharge: Insufficient knowledge, unclear task reallocation and lack of collaboration as major barriers to medication safety. BMC Health Serv Res. 2012 Jun 21;12:170. 20. Boockvar KS, et al. Medication reconciliation: Barriers and facilitators from the perspectives of resident physicians and pharmacists. J Hosp Med. 2011;6(6):329– 337. 21. Baillie CA, et al. The readmission risk flag: Using the electronic health record to automatically identify patients at risk for 30-day readmission. J Hosp Med. 2013;8(12):689–695. 22. Budnitz DS, et al. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. 2011 Nov 24;365(21):2002–2012. 23. Fleischut PM, et al. 2011 John M. Eisenberg Patient Safety and Quality Awards. The effect of a novel Housestaff Quality Council on quality and patient safety. Innovation in patient safety and quality at the local level. Jt Comm J Qual Patient Saf. 2012;38(7):311–317. 24. Tess AV, et al. Combining clinical microsystems and an experiential quality improvement curriculum to improve residency education in internal medicine. Acad Med. 2009;84(3):326–334. 25. Nasca TJ, et al. The next GME accreditation system—Rationale and benefits. N Engl J Med. 2012 Mar 15;366(11):1051–1056. 26. McAlearney AS, Hefner JL. Facilitating central line-associated bloodstream infection prevention: A qualitative study comparing perspectives of infection control professionals and frontline staff. Am J Infect Control. 2014;42(10 Suppl):S216–222.
Volume 41 Number 10
Copyright 2015 The Joint Commission
461
A Medical Resident–Pharmacist Collaboration Improves the Rate of Medication Reconciliation Verification at Discharge Daniel A. Caroff, MD; Therese Bittermann, MD; Charles E. Leonard, PharmD, MSCE; Gene A. Gibson, PharmD; Jennifer S. Myers, MD
M
edication reconciliation, designated as a Joint Commission National Patient Safety Goal in 2004, became a requirement for hospital accreditation in 20061 in response to concerns regarding the prevalence of preventable adverse drug events (ADEs) during patient transitions. A systematic review of 22 studies that evaluated medication history errors on hospital admission showed that up to two thirds of patients had at least one medication history error, half of which were believed to be clinically important.2 It is estimated that ADEs account for approximately 280,000 hospital admissions in the United States annually.3 ADEs after hospital discharge may be responsible for two thirds of postdischarge complications,4 and 24%– 27% of postdischarge ADEs may be preventable.4,5 Despite efforts to improve compliance, the effectiveness of inpatient medication reconciliation to prevent ADEs has been challenged by the inconsistency of the process within institutions.6 It has been demonstrated that pharmacists obtain more accurate medication histories than do physicians2,7,8 and might perform more accurate medication reconciliation.9 Pharmacist participation in discharge planning has also been shown to improve identification of medication errors with potential for harm.10–12 As a result, institutions have attempted to determine the most effective way to perform medication reconciliation in collaboration with inpatient pharmacists while best using the skill sets of clinical team members. Although some quality improvement (QI) interventions have demonstrated impressive reductions in readmission rates as a result of enhanced error recognition by pharmacists,13,14 others have not led to meaningful change,15,16 making their use of hospital staffing and financial resources difficult to justify.17 Challenges in communication, collaboration, and task allocation have also created barriers to medication reconciliation.18–20 At our institution, we encountered similar challenges in medication reconciliation at discharge. Our standard discharge process on two general medicine units involved creation of October 2015
Article-at-a-Glance Background: At the Hospital of the University of Penn-
sylvania (Philadelphia), it is standard practice to perform medication reconciliation at patient discharge. Although pharmacists historically were available to assist resident physicians in the discharge medication reconciliation process, the process was never standardized. An internal review showed a 60%–70% rate of pharmacist review of discharge medication lists, potentially enabling medication errors to go unnoticed during transitions of care. In response, a medical resident– and pharmacist–led collaboration was designed, and a pre-postintervention study was conducted to assess its effectiveness. Methods: A new work flow was established in which housestaff notified pharmacists when a preliminary discharge medication list was ready for reconciliation and provided access for pharmacists to correct medication errors in the electronic discharge document with physician approval. Length of stay, average time of day of patient discharge, and readmission data were compared in the pre- and post-intervention periods. Results: There were 981 discharges in the preintervention period and 1,207 in the postintervention period. The rate of pharmacist reconciliation increased from 64.0% to 82.4% after the intervention (p < .0001). The average number of errors identified and corrected by pharmacists decreased from 0.979 to 0.862 per discharge (p < .0001). There was no significant change in readmission rates or time of discharge after the intervention. Conclusions: Redesigning the discharge medication reconciliation process in a teaching hospital to include a review of medical resident discharge medication lists by pharmacists provided more opportunities for discharge medication error identification and correction.
Volume 41 Number 10
Copyright 2015 The Joint Commission
457
The Joint Commission Journal on Quality and Patient Safety the discharge medication list by the primary medical resident, finalization of discharge instructions and printing of paper prescriptions by the medical resident, and, lastly, review of the medication list by pharmacy and nursing staff before discharging the patient. This final step, however, was often limited by poor or inefficient communication between team members. An internal review in March 2012 showed that pharmacist medication review occurred in only 60%–70% of discharges. Thus, there was concern that discharge medication errors would go unnoticed and uncorrected. In response, a medical resident– led QI initiative was designed to improve the identification and correction of medication errors at hospital discharge. We conducted a pre-post-intervention study to assess its effect on the rate of pharmacist-verified medication reconciliation at hospital discharge.
Methods Study Design A quasi-experimental study using multiple noncontiguous data periods (preintervention period: February 2012–July 2012; post intervention period: August 2012–January 2013) was designed. The University of Pennsylvania’s Institutional Review Board approved this study.
Setting and Participants This study occurred at the Hospital of the University of Pennsylvania, a 789-bed tertiary care teaching hospital in Philadelphia. Eight inpatient teaching services on two hospital units participated, each composed of one internal medicine resident, two medical interns, one to three medical students, and either a hospitalist or medical subspecialty attending. All patients discharged on Monday through Friday from these services were included. Patients who left against medical advice or died were excluded, as were patients discharged on a weekday national holiday because of limitations in pharmacy staffing. Four unit-based clinical pharmacists (UBCPs), who were responsible for delivering comprehensive pharmaceutical and patient-centered care, participated in the study. The ratio of UBCPs to inpatient beds was 1:20 on the two inpatient units that were studied. The UBCPs performed medication reconciliation on admis sion, provided medication counseling to patients, and reviewed inpatient medication orders. At hospital discharge, the UBCPs were expected to review the discharge medication list and counsel patients and families about important changes and possible adverse effects. UBCPs were required to document all interventions in the electronic health record (EHR) system (Allscripts, Chicago). 458
October 2015
Intervention The intervention consisted of three components, as follows. 1. Policy for Medical Residents to Obtain Pharmacist Approval. A policy was introduced that required medical resi dents to obtain pharmacist approval of the discharge medica tion list before any weekday discharge. An order within the EHR allowed medical residents to notify UBCPs when a patient’s discharge medication list was ready for review. Medical resident education on this new work flow was provided during daily meetings, in automated e-mails at the start of each new rotation, and in flyers distributed throughout the two inpatient units involved. We did not have a “hard stop” function to prevent discharges without UBCP review. 2. Ability of Pharmacists to Edit the Discharge Medication List. UBCPs were enabled to directly edit the discharge medication list, accessible through the EHR. Three months before the intervention, two 30-minute training sessions on the electronic discharge software were held for all UBCPs. Sessions were led by a medical resident, and a slide set was provided by our information technology staff. When reviewing a discharge medication list, pharmacists corrected any errors after verbal confirmation from the patient’s primary resident. Documentation of all errors and corrections were logged into the EHR system. After pharmacist review, the medical resident was notified and given approval to write the discharge order. All necessary prescriptions were then printed and signed by the medical resident. 3. Nurse’s Final Review of Medication List. A final review of the medication list was performed by the patient’s nurse at the bedside immediately before discharge. This last step was already part of the standard discharge process.
Demographic and Clinical Data The following basic demographic and clinical data were collected for all patients included in the study: date of birth, date of admission, date of discharge, inpatient medical service, length of stay (LOS), risk of readmission (as determined by two or more hospital admissions during the previous 12 months 21), and pharmacist documentation of discharge medication errors. The study group designated immunosuppressants, antimicrobials, anticoagulants, insulins, and controlled substances as “high-risk” medications on the basis of the most common ADEs for which older adults seek treatment in the emergency department.22 Errors were categorized as (1) medication omitted, (2) incorrect medication, (3) incorrect dose, (4) incorrect frequency, (5) incorrect formulation, (6) medication unintentionally listed or prescribed (medication that a patient “does not take”), (7) medication duplication, and (8) prescription-writing Volume 41 Number 10
Copyright 2015 The Joint Commission
The Joint Commission Journal on Quality and Patient Safety errors (for example, medication scheduled instead of “as needed” or use of unapproved abbreviations) (Table 1, right). The intervention began on August 27, 2012. Data were gathered during five time periods: two preintervention periods (February 1, 2012–March 30, 2012, and July 1, 2012–July 30, 2012), one period immediately following the intervention (August 27, 2012–October 18, 2012), and two subsequent postintervention periods (October 22, 2012– November 16, 2012, and January 14, 2013–February 16, 2013). All data were collected by two of the authors [D.A.C., T.B.] using EHRs. Unit-specific rates of hospital readmission and average time of day of patient discharge were obtained from institutional databases.
Data Analysis Raw data were combined into preintervention and postintervention groups. Tests of proportion equality were used to calculate p values for differences in proportions. T-tests were used to calculate p values for differences in the means of continuous variables. These analyses were conducted using Stata/MP 13.1 (StataCorp, LP, College Station, Texas).
Results Demographic and Clinical Data In total, there were 981 discharges in the two preintervention periods versus 1,207 discharges in the three postintervention periods. Basic demographic and clinical data are shown in Table 2 (page 460). Overall, there were no appreciable differences in the mean age of patients in each group, in the proportion of inpatient services represented, and in the percentage of patients identified as at risk for readmission.
Outcomes After implementation of the intervention, a statistically significant increase in the number of patients discharged with UBCP verification of their medication reconciliation was noted (64.0% versus 82.4%, p < .0001; Table 3 [page 460]). Although there was no significant difference in the number of discharge documents with at least one error present, the mean number of UBCP–identified errors per discharge decreased from 0.979 to 0.862 (p < .0001). Table 3 details the types of errors found in both groups, and in most cases there was no significant change in the proportion of specific error types before and after the intervention. There was no difference in the number of highrisk medication errors identified after the intervention. A significant difference in the mean LOS was noted (7.15 days preintervention versus 6.17 days postintervention, October 2015
Table 1. Examples of Medication Error Types Error Type
Example
Omission
Patient takes morphine sulfate; however, drug missing from document
Drug name
Patient takes morphine sulfate; however, hydromorphone listed
Dose
Patient takes morphine sulfate 15 mg; however, 30 mg dose listed
Frequency
Patient takes morphine sulfate twice daily; however, listed as three times daily
Formulation
Patient takes morphine extended release; however, immediate release listed
Not on medication
Patient does not take morphine sulfate; however, it is listed in document
Prescribing
Patient takes morphine sulfate as needed; however, it is listed as scheduled
Duplication
Patient’s morphine extended release and MS Contin* are listed simultaneously
*Purdue Pharma (Stamford, Connecticut); time-released formulation of morphine sulfate.
p < .0051). There was no correlation between LOS and the likelihood of medication reconciliation (point-biserial correlations 0.0069 preintervention and 0.0008 postintervention, with p values for rejection of the critical t values .83 and .98, respectively). In addition, there was no change in the average time of discharge after implementation of the intervention. Finally, there was no significant difference in 7- and 30-day readmission rates after the intervention (Table 3).
Discussion
We demonstrated a statistically significant improvement in the standardization of the discharge medication reconciliation process through the design of a collaborative QI intervention between medical residents and clinical pharmacists. The rate of pharmacist discharge medication review increased from 62.0% to 82.4% after the intervention. Although we did not assess medication-related adverse events after discharge, we believe that our results show value added for the patients for whom their discharge medications were corrected and verified by a pharmacist. The remaining 17.6% of postintervention discharges that did not undergo pharmacist review might have been caused by imperfect compliance, miscommunication, or competing clinical obligations on the part of the house staff, pharmacists, or nursing staff. Further investigation of these issues may help improve the effect of the intervention. In addition to improvements in the rate of discharge medication reconciliation review, we also observed a significant decrease in the error rate of physician-authored discharge Volume 41 Number 10
Copyright 2015 The Joint Commission
459
The Joint Commission Journal on Quality and Patient Safety Table 2. Demographic and Clinical Data, Pre- and Postintervention Preintervention Feb–Jul 2012 N = 981
Postintervention Aug 2012–Jan 2013 N = 1,207
Mean age at discharge, years (± SD)
Inpatient service, N (%)
55 (± 18.2*)
55 (± 17.6)
General medicine
599 (61.1)
720 (59.7)
Hepatology
134 (13.7)
156 (12.9)
Pulmonary
117 (11.9)
139 (11.5)
Infectious disease
88 (9.0)
85 (7.0)
Nephrology
43 (4.4)
107 (8.9)
291 (29.7)
367 (30.4)
Number of high-readmission-risk patients
†
SD, standard deviation. * Sums to 100.1% because of rounding. † Patients are flagged as high readmission risk in the electronic health record if they have been admitted to the hospital ≥ 2 times in the 12 months previous to the admission.
Table 3. Measures of Discharge Medication Accuracy, Pre- and Postintervention Preintervention Feb–Jul 2012
Postintervention Aug 2012–Jan 2013
P Value
Mean length of stay (days ± SD)
7.15 ± 9.66
6.17 ± 6.62
< .0051
Total Discharges, N (%)
981 (100.0)
1,207 (100.0)
—
Discharges with medication reconciliation, N (%)
628 (64.0)
995 (82.4)
< .0001
7-day readmissions, N
68
77
.6254
30-day readmissions, N
296
328
.1548
Reconciliations completed, N (%)
628 (100.0)
995 (100.0)
—
Reconciliations with ≥ 1 pharmacist-identified error, N (%)
309 (49.2)
448 (45.0)
.1003
0.979 ± 0.142
0.862 ± 0.345
< .0001
615 (100.0*)
858 (100.0)
—
208 (33.8)
265 (30.9)
.2341 .0653
Discharges:
Readmissions:
Medication reconciliation:
Pharmacist-identified errors per reconciliation (mean ± SD) Errors, N (%) Total pharmacist-identified errors Omission Drug name
31 (5.0)
27 (3.1)
121 (19.7)
150 (17.5)
.2842
Frequency
54 (8.8)
87 (10.1)
.3818
Formulation
12 (2.0)
38 (4.4)
.0096
Not on medication
40 (6.5)
47 (5.5)
.4100
Prescribing
135 (22.0)
215 (25.1)
.1671
Duplication
14 (2.3)
29 (3.4)
.2147
127 (20.7)
182 (21.2)
.7940
Dose
High-risk medication †
* Sums to 100.1% because of rounding. † High-risk medications include immunosuppressants, antimicrobials, anticoagulants, insulins, and controlled substances.
460
October 2015
medication lists. There are several possibilities that could explain this finding. First, it is possible that an increased focus by both the hospital and the residency program led residents to recognize the importance of medical reconciliation and perform the task more carefully. It is also possible that medical residents changed their behavior and performed medication reconciliation more carefully, as they knew their discharge medication lists were being reviewed by a pharmacy colleague. Finally, it is possible that a pharmacist’s sensitivity for finding errors was decreased because of a larger number of documents being reviewed. Interestingly, before the intervention, there was concern that discharge medication errors made by medical residents might increase as a result of this type of intervention, given their knowledge that a pharmacist would be correcting their errors; fortunately, this did not occur. Although internal data have suggested that medication reconciliation by pharmacists reduces readmissions, our intervention did not reduce unit-specific readmission rates. Numerous other factors and concurrent institutional QI initiatives related to the discharge process could have affected the observed readmission rates and potentially confounded our intervention’s effect on this specific outcome. Importantly, the time of hospital discharge did not change before and after the intervention. We hypothesize that the time spent by pharmacists on additional discharge medication reconciliations was offset by earlier identification of patients to be discharged, and improved communication and coordination regarding discharge by the resident, pharmacist, and nurse as a result of the intervention. It is notable that this intervention did not require additional pharmacy staffing. Although it did require a small amount of pharmacist time for training, the positive impact on the discharge medication reconciliation process and likely prevention of outpatient ADEs affords this intervention at worst a cost-neutral economic impact and, very possibly, a net savings, although additional research is needed to validate this assumption. There are several limitations to this study. The quasi-experimental design is vulnerable to
Volume 41 Number 10
Copyright 2015 The Joint Commission
The Joint Commission Journal on Quality and Patient Safety confounding by a number of related QI initiatives and trends within our hospital, although this limitation was partially addressed by including multiple data periods both before and after the intervention. We did not record the total number of medications prescribed per patient, which, if significantly lower in the postintervention period, could conceivably have caused bias. We are unable to account for the pre- or postintervention rate of medication errors in the minority of patients who were discharged without pharmacist review. Findings may not be reproducible at other institutions or residency programs due to cultural or practical differences, such as limitations in pharmacy resources. Within the study hospital, however, in light of the positive results of this study, the intervention was expanded from the two original inpatient units to all inpatient services in the Department of Medicine, and the process has been sustained with positive feedback from both pharmacists and medical residents up through the present time. Being at the forefront of clinical care in teaching hospitals, medical and surgical residents are in an ideal position to identify QI and patient safety problems and design solutions to mitigate them.23,24 Resident-led initiatives are an ideal way to provide practical QI training and experience to trainees that meets the expectations of the Accreditation Council for Graduate Medical Education in systems-based practice.25 Furthermore, we suspect that the resident-led nature of this intervention minimized the frontline physician resistance to change that can be a barrier to successful QI project implementation.26
Conclusion
A collaborative medical resident–pharmacist intervention was effective in creating a sustained improvement in the discharge medication reconciliation process at the Hospital of the University of Pennsylvania. J Daniel A. Caroff, MD, formerly Internal Medicine Resident, De partment of Medicine, University of Pennsylvania, Philadelphia, is a Fellow in Infectious Disease, Brigham and Women’s Hospital and Massachusetts General Hospital, Boston. Therese Bittermann, MD, formerly Internal Medicine Resident, Department of Medicine, University of Pennsylvania, is a Gastroenterology Fellow, University of Pennsylvania. Charles E. Leonard, PharmD, MSCE, is Senior Research Investigator and Adjunct Assistant Professor of Epidemiology, Center for Pharmacoepidemiology Research and Training, Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania. Gene A. Gibson, PharmD, is Associate Director, Department of Pharmacy Services, Hospital of the University of Pennsylvania. Jennifer S. Myers, MD, is Associate Professor of Clinical Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania. Please address correspondence to Daniel A. Caroff, [email protected].
October 2015
References
1. The Joint Commission. Using medication reconciliation to prevent errors. Sentinel Event Alert. 2006 Jan 25;(35):1–4. Accessed Aug 26, 2015. www.joint commission.org/sentinel_event_alert_issue_35_using_medication_reconciliation _to_prevent_errors/. 2. Tam VC, et al. Frequency, type and clinical importance of medication history errors at admission to hospital: A systematic review. CMAJ. 2005 Aug 30;173(5): 510–515. 3. US Department of Health & Human Services, Office of Disease Prevention and Health Promotion (ODPHP). National Action Plan for Adverse Drug Event Prevention. Washington, DC: ODPHP, 2014. Accessed Aug 28, 2015. http://health.gov /hai/pdfs/ADE-Action-Plan-508c.pdf. 4. Forster AJ, et al. The incidence and severity of adverse events affecting patients after discharge from the hospital. Ann Intern Med. 2003 Feb 4;138(3):161–167. 5. Forster AJ, et al. Adverse drug events occurring following hospital discharge. J Gen Intern Med. 2005;20(4):317–323. 6. Clay BJ, et al. Results of a medication reconciliation survey from the 2006 Society of Hospital Medicine national meeting. J Hosp Med. 2008;3(6):465–472. 7. Reeder TA, Mutnick A. Pharmacist- versus physician-obtained medication histories. Am J Health Syst Pharm. 2008 May 1;65(9):857–860. 8. Carter MK, et al. Pharmacist-acquired medication histories in a university hospital emergency department. Am J Health Syst Pharm. 2006 Dec 15;63(24): 2500–2503. 9. Kushniruk AW, et al. Cognitive analysis of a medication reconciliation tool: Applying laboratory and naturalistic approaches to system evaluation. Stud Health Technol Inform. 2011;164:203–207. 10. Abdel-Qader DH, et al. Pharmacists’ interventions in prescribing errors at hospital discharge: An observational study in the context of an electronic prescribing system in a UK teaching hospital. Drug Saf. 2010 Nov 1;33(11):1027–1044. 11. Feldman LS, et al. Nurse-pharmacist collaboration on medication reconciliation prevents potential harm. J Hosp Med. 2012;7(5):396–401. 12. Allende Bandrés MÁ, et al. Pharmacist-led medication reconciliation to reduce discrepancies in transitions of care in Spain. Int J Clin Pharm. 2013;35(6):1083– 1090. 13. Eisenhower C. Impact of pharmacist-conducted medication reconciliation at discharge on readmissions of elderly patients with COPD. Ann Pharmacother. 2014;48(2):203–208. 14. Pal A, Babbott S, Wilkinson ST. Can the targeted use of a discharge pharmacist significantly decrease 30-day readmissions? Hosp Pharm. 2013;48(5):380–388. 15. Kripalani S, et al; PILL-CVD (Pharmacist Intervention for Low Literacy in Cardiovascular Disease) Study Group. Effect of a pharmacist intervention on clinically important medication errors after hospital discharge: A randomized trial. Ann Intern Med. 2012 Jul 3;157(1):1–10. 16. Kwan JL, et al. Medication reconciliation during transitions of care as a patient safety strategy: A systematic review. Ann Intern Med. 2013 Mar 5;158(5 Pt 2) :397–403. 17. Lu Y, et al. Quality improvement through implementation of discharge order reconciliation. Am J Health Syst Pharm. 2013 May 1;70(9):815–820. 18. Meguerditchian AN, et al. Medication reconciliation at admission and discharge: A time and motion study. BMC Health Serv Res. 2013 Nov 21;13:485. 19. van Sluisveld N, et al. Medication reconciliation at hospital admission and discharge: Insufficient knowledge, unclear task reallocation and lack of collaboration as major barriers to medication safety. BMC Health Serv Res. 2012 Jun 21;12:170. 20. Boockvar KS, et al. Medication reconciliation: Barriers and facilitators from the perspectives of resident physicians and pharmacists. J Hosp Med. 2011;6(6):329– 337. 21. Baillie CA, et al. The readmission risk flag: Using the electronic health record to automatically identify patients at risk for 30-day readmission. J Hosp Med. 2013;8(12):689–695. 22. Budnitz DS, et al. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. 2011 Nov 24;365(21):2002–2012. 23. Fleischut PM, et al. 2011 John M. Eisenberg Patient Safety and Quality Awards. The effect of a novel Housestaff Quality Council on quality and patient safety. Innovation in patient safety and quality at the local level. Jt Comm J Qual Patient Saf. 2012;38(7):311–317. 24. Tess AV, et al. Combining clinical microsystems and an experiential quality improvement curriculum to improve residency education in internal medicine. Acad Med. 2009;84(3):326–334. 25. Nasca TJ, et al. The next GME accreditation system—Rationale and benefits. N Engl J Med. 2012 Mar 15;366(11):1051–1056. 26. McAlearney AS, Hefner JL. Facilitating central line-associated bloodstream infection prevention: A qualitative study comparing perspectives of infection control professionals and frontline staff. Am J Infect Control. 2014;42(10 Suppl):S216–222.
Volume 41 Number 10
Copyright 2015 The Joint Commission
461