537469
research-article2014
AOPXXX10.1177/1060028014537469Annals of PharmacotherapyWimmer et al
Article
Medication Regimen Complexity and Unplanned Hospital Readmissions in Older People
Annals of Pharmacotherapy 1–9 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1060028014537469 aop.sagepub.com
Barbara C. Wimmer, MSc1,2, Elsa Dent, PhD3,4, J. Simon Bell, PhD1,2, Michael D. Wiese, PhD2, Ian Chapman, PhD3, Kristina Johnell, PhD4, 5, and Renuka Visvanathan, PhD3,6
Abstract Background: Medication-related problems and adverse drug events are leading causes of preventable hospitalizations. Few previous studies have investigated the possible association between medication regimen complexity and unplanned rehospitalization. Objective: To investigate the association between discharge medication regimen complexity and unplanned rehospitalization over a 12-month period. Methods: The prospective study comprised patients aged ≥70 years old consecutively admitted to a Geriatrics Evaluation and Management (GEM) unit between October 2010 and December 2011. Medication regimen complexity at discharge was calculated using the 65-item validated Medication Regimen Complexity Index (MRCI). Cox proportional-hazards regression was used to compute unadjusted and adjusted hazard ratios (HRs) with 95% CIs for factors associated with rehospitalization over a 12-month follow-up period. Results: Of 163 eligible patients, 99 patients had one or more unplanned hospital readmissions. When adjusting for age, sex, activities of daily living, depression, comorbidity, cognitive status, and discharge destination, MRCI (HR = 1.01; 95% CI = 0.81-1.26), number of discharge medications (HR = 1.01; 95% CI = 0.94-1.08), and polypharmacy (≥9 medications; HR = 1.12; 95% CI = 0.69-1.80) were not associated with rehospitalization. In patients discharged to nonhome settings, there was an association between rehospitalization and the number of discharge medications (HR = 1.12; 95% CI = 1.01-1.25) and polypharmacy (HR = 2.24; 95% CI = 1.02-4.94) but not between rehospitalization and MRCI (HR = 1.32; 95% CI = 0.98-1.78). Conclusion: Medication regimen complexity was not associated with unplanned hospital readmission in older people. However, in patients discharged to nonhome settings, the number of discharge medications and polypharmacy predicted rehospitalization. A patient’s discharge destination is an important factor in unplanned medication-related readmissions. Keywords medication regimen complexity, hospital readmission, elderly
Introduction Australians aged ≥65 years comprise 14% of the population but account for 39% of hospital admissions.1 These figures are comparable to those in the United Kingdom and United States where people aged ≥65 years account for 39% and 36% of hospital admissions and discharges, respectively.2,3 Medication-related problems and adverse drug events are leading causes of preventable hospital admissions.4,5 Anticholinergic and sedative medications, warfarin, insulin, and oral antiplatelet and oral hypoglycemic medication have been associated with an increased risk of hospitalization.5,6 Use of potentially inappropriate medications and polypharmacy have also been associated with hospital admissions.4,7 The risk of medication-related hospitalization may be increased in people with other risk factors for
1
Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical Sciences, Monash University, Melbourne, Australia 2 Sansom Institute, School of Pharmacy and Medical Sciences, University of South Australia, Adelaide, Australia 3 Discipline of Medicine, University of Adelaide, Adelaide, Australia 4 Discipline of Public Health, University of Adelaide, Adelaide, Australia 5 Aging Research Center, Karolinska Institutet and Stockholm University, Stockholm, Sweden 6 Aged and Extended Care Services, The Queen Elizabeth Hospital and the Adelaide Geriatrics Training and Research with Aged Care (GTRAC) Centre, University of Adelaide, Adelaide, Australia Corresponding Author: Barbara C. Wimmer, Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical Sciences, Monash University, 381 Royal Parade, Parkville, VIC, Australia. Email: [email protected]
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
2
Annals of Pharmacotherapy
hospitalization, including advanced age, ≥5 comorbidities, cognitive impairment, living in a nonhome setting, depression, and recent admission.8-10 A link between polypharmacy and medication-related hospital admission has been demonstrated in older people.4 However, the number of medications taken by a patient is only one element of regimen complexity. Other elements that contribute to regimen complexity include the number of different dose forms, dosing frequencies, and special instructions for medication use.11 These additional elements may affect whether or not an older person can manage their medication regimen at home. Having a high medication regimen complexity has also been associated with lower medication adherence.12 However, few previous studies have looked at the complexity of medication regimens and rehospitalization.13,14 The objective of this study was to investigate the association between discharge medication regimen complexity and unplanned hospital readmission over a 12-month period. We hypothesized that medication regimen complexity would be associated with unplanned rehospitalization, particularly in people with cognitive impairment and in those discharged directly to home. This was because people with cognitive impairment or dementia and those discharged to home may experience difficulties with or lack of support in managing their medications.
Methods Setting and Participants The setting of this prospective study was the Geriatrics Evaluation and Management (GEM) unit of a public hospital in Adelaide, South Australia. The study included patients aged ≥70 years admitted consecutively to the GEM unit between October 22, 2010, and December 23, 2011. Potential participants were excluded if a proxy was required but unavailable (eg, in case of dementia, unresolved delirium within 72 hours, or language barriers), their treating clinician recommended exclusion (eg, as a result of elder abuse, physical aggression, or being medically unwell), they were deemed infectious, they were missed by the researcher, they declined to participate, or they were transferred to another hospital or to palliative care service. Of the 166 included participants discharged from the GEM unit, 2 were transferred to another hospital and 1 to palliative care service. Thus, this study investigated rehospitalization among 163 patients managed and discharged from the GEM unit.
Data Collection Participants were recruited within 72 hours of their admission. All physiological assessment data were collected by the same researcher (ED). Discharge medication data were extracted directly from the hospital separation summary
prepared predominantly by medical practitioners and recorded in the Open Architecture Clinical Information System (OACIS, South Australia Health Department 2009). At the GEM unit, there is a full-time clinical pharmacist, and discharge medication lists are routinely provided to both the patients and their primary care physician to inform ongoing prescribing. All medication data were extracted from OACIS by the same pharmacist researcher (BCW).
Medication Assessment Each patient’s medication regimen was assessed at hospital discharge. Medications were categorized according to the Anatomical Therapeutic Chemical Classification System recommended by the World Health Organization. Each patient’s regimen complexity was calculated using the original 65-item validated Medication Regimen Complexity Index (MRCI).11 The MRCI considers dosage forms, dosing frequencies, and additional directions, with higher MRCI scores reflecting more complex medication regimens.11 Prescription and nonprescription medications, nutritional supplements, health products, dermatological preparations, and short-term medications (eg, antibiotics) were all considered when computing the MRCI. The inclusion of nonprescription medications was consistent with other recent studies utilizing the MRCI.14,15 The numbers of medications at hospital discharge were counted. Polypharmacy was defined as the use of 9 or more medications on a regular or as-needed basis. This definition was consistent with that recommended for use in long-term facilities.16 The more conservative commonly used definition of 5 or more medications was not practical to use because almost all participants (n = 155) took 5 or more medications.17
Main Outcome Measure The main outcome measure was unplanned rehospitalization in the 12 months following hospital discharge. The 12-month follow-up period was defined individually for each patient using the date of their first discharge as the index date. Each patient was only included once in the study. Data on unplanned rehospitalization to emergency departments over a 12-month follow-up period were extracted from the administrative records of the 3 major acute care hospitals serving the Central-Northern regions of metropolitan Adelaide by hospital case-mix staff. Dates of death were obtained for all patients who died within the 12-month follow-up period from the official register of Births, Deaths and Marriages.
Covariates The Charlson’s Comorbidity Index was calculated to assess comorbidity for each patient. This index is a weighted index that is suitable for use with medical records.18 Depressive
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
3
Wimmer et al symptoms were assessed using the Geriatric Depression Scale (GDS-15).19 Activities of daily living (ADLs) were computed using the 10-item Barthel’s ADL index. This is a performance scale with possible scores ranging from 0 to 100, with higher scores indicating greater independence.20 Discharge destination data were extracted from patients’ separation summaries. Patients were categorized as being “discharged directly to home” if they returned to their own home immediately after hospital discharge, both with or without new community services such as Meals on Wheels, domiciliary care, or home and community care services. Patients were considered “discharged to non-home settings” if they were discharged to either high- or low-level residential aged care, transition care (ie, restorative care in a residential aged care facility), or off-site inpatient rehabilitation. For the purpose of the analyses, patients were defined as having cognitive impairment if they had an inhospital Mini Mental State Examination (MMSE)21 or Rowland Universal Dementia Assessment Scale (RUDAS)22 score ≤24 or a clinical diagnosis of dementia recorded in the hospital discharge summary.
Statistical Analyses An a priori sample size calculation was performed based on a previous study, which reported that 85% of patients with an MRCI >20 were readmitted over 12 months compared with 15% of patients with an MRCI ≤20.14 Based on these findings, assuming a power of 90% and α = 0.05, a sample size of 29 patients would be required to detect a significant difference in rehospitalization. Study data were summarized using means, medians, and SDs. Shapiro-Wilk’s test (P > 0.05) and visual inspection of histograms, Q-Q plots, and box plots were used to test for normality of the data. Normally distributed characteristics were summarized using means with SDs. Nonnormally distributed variables were reported using medians and ranges. Dichotomous variables were reported as numbers and percentages. χ2 Tests were used to investigate potential differences between categorical variables, t tests were used for normally distributed continuous variables, and Mann Whitney U tests for nonnormally distributed variables. Two-sided tests were used to investigate potential differences between patients who were readmitted and those who were not readmitted during the 12-month follow-up period. Kaplan-Meier survival curves were computed to visually depict the relationship between MRCI divided in quartiles and time to rehospitalization. Univariate Cox proportionalhazards regression was then used to compute hazard ratios (HRs) with 95% CIs for factors associated with unplanned rehospitalization. Variables that were associated with unplanned readmission in the unadjusted analyses (defined as P < 0.1) or those deemed clinically relevant on the basis of previous research4,8-10 were included in multivariate models. Variables in each model were checked for
multicollinearity. Age, ADLs, depression, comorbidities, and number of medications were analyzed as continuous variables. The total MRCI score was divided by 10 and analysed as a continuous variable. Sex, the presence of cognitive impairment or dementia, and polypharmacy were analyzed as dichotomized variables. To account for deaths that occurred during the follow-up period, the analyses were censored at each patient’s date of death or the end of the 12-month follow-up period, whichever occurred first. Subgroup analyses were performed stratifying participants into (1) patients with and without dementia or cognitive impairment and (2) patients discharged directly to home or to nonhome settings. The stratifications were performed because previous research has demonstrated that both cognitive status and discharge destination can have an impact on readmission rates.8 Sensitivity analyses were performed to investigate the association between MRCI and rehospitalization when only participants with a dementia diagnosis were considered (as opposed to either a dementia diagnosis or MMSE/RUDAS ≤24). Data were analyzed using the Statistical Package for the Social Sciences (SPSS 20, SPSS Inc, Chicago, IL).
Ethical Considerations All potential participants were provided with written information about the study. Written informed consent to participate was obtained from all participants themselves or their health proxy. The study was approved by the Human Research Ethics Committees at The Queen Elizabeth Hospital and the University of South Australia.
Results Main Results For the 163 patients discharged from the GEM unit, the mean MRCI was 27.86 (SD = 11.63), and the mean number of medications was 9.7 (SD = 3.60); 43 had an MRCI of ≤20, and 40 patients had an MRCI of >35., In total, 99 (61%) patients had one or more unplanned rehospitalizations within 12 months of being discharged. Among the 99 patients readmitted, the mean MRCI was 28.01 (SD = 12.48), the mean number of medications was 9.7 (SD = 3.70), 60 had cognitive impairment or dementia, 55 had been discharged directly to home, and 23 died during the 12-month follow-up. In total, 64 patients were not readmitted to hospital. Among these patients, the mean MRCI was 27.62 (SD = 10.26), the mean number of medications was 9.7 (SD = 3.48), 23 had cognitive impairment or dementia, 32 had been discharged directly to home, and 7 died during the 12-month follow-up (Table 1). The MRCI was not significantly different in patients who were readmitted and not readmitted over 12 months (mean difference = −0.39; 95% CI = −4.09 to 3.30). People
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
4
Annals of Pharmacotherapy
Table 1. Characteristics of the Patients With and Without Unplanned Hospital Readmission Over the 12-Month Follow-up Stratified by Cognitive Function and Discharge Destination.a All Patients (n = 163) Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) DD directly to home n (%) GDS-15 (range) Cognitive impairment or dementia n (%) Cognitive Impairment or Dementia (n = 83)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) DD directly to home, n (%)
No Cognitive Impairment or Dementia (n = 80)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) DD directly to home, n (%)
DD Directly to Home (n = 87)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) Cognitive impairment or dementia, n (%)
DD to Nonhome Settings (n = 76)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) Cognitive impairment or dementia, n (%)
Readmitted (n = 99)
Not Readmitted (n = 64)
84.9 (6.20) 68 (68.7) 28.01 (12.48) 68.5 (8-100) 3.0 (0-10) 55 (55.6) 4.0 (0-15) 60 (60.6) Readmitted (n = 60)
85.6 (6.74) 50 (78.1) 27.62 (10.26) 76.0 (21-100) 2.0 (0-8) 32 (50.0) 3.0 (0-15) 23 (35.9) Not Readmitted (n = 23)
84.4 (6.33) 39 (65.0) 27.98 (11.47) 65.0 (8-100) 3.0 (0-10) 4.0 (0-15) 31 (51.7) Readmitted (n = 39)
85.04 (7.33) 18 (78.3) 25.70 (10.75) 71.0 (31-91) 2.0 (0-8) 3.0 (0-15) 8 (34.8) Not Readmitted (n = 41)
85.7 (5.98) 29 (74.4) 28.06 (14.05) 72.0 (19-94) 3.0 (0-7) 4.0 (0-14) 24 (61.5) Readmitted (n = 55)
85.9 (6.47) 32 (78.0) 28.70 (9.94) 81.0 (21-100) 2.0 (0-8) 3.5 (0-12) 24 (58.5) Not Readmitted (n = 32)
84.5 (6.75) 38 (69.1) 24.91 (9.78) 73.0 (23-94) 3.0 (0-10) 4.0 (0-13) 31 (56.4)
84.7 (7.15) 21 (65.6) 28.08 (9.42) 84.5 (21-100) 2.0 (0-7) 3 (0-15) 8 (25.0)
P Value 0.50 0.19 0.83 0.006b 0.17 0.49 0.44 0.002b P Value 0.68 0.24 0.41 0.08 0.50 0.76 0.17 P Value 0.88 0.70 0.82 0.15 0.27 0.61 0.78 P Value 0.89 0. 74 0.14 20 to 27 MRCI: >27 to 35 MRCI: >35
0.2
0.4
Cumulative probability of not being readmitted
MRCI: 1 to 20 MRCI: >20 to 27 MRCI: >27 to 35 MRCI: >35
0.6
0.8
MRCI: 1 to 20 MRCI: >20 to 27 MRCI: >27 to 35 MRCI: >35
1.0
B
0.8
1.0
A
0
60
120
180
240
300
360
Time (days) Figure 1. Kaplan-Meier survival curves on MRCI (quartiles) and unplanned hospital readmission for (A) all patients, (B) patients discharged directly to home, (C) patients discharged to nonhome settings, and number of patients (n) in each subgroup. MRCI Quartile 1st 2nd 3rd 4th Total
All Patients (n)
DD Directly to Home (n)
DD Nonhome Settings (n)
43 41 39 40 163
24 25 25 13 87
19 16 14 27 76
Abbreviations: MRCI, Medication Regimen Complexity Index; DD, discharge destination; MRCI 1st quartile, 1-20; 2nd quartile, 20.1-27; 3rd quartile, 27.1-35; 4th quartile, 35.1-73.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
7
Wimmer et al
Subgroup Analyses for Discharge Destination In the unadjusted analyses, there was no association between MRCI and rehospitalization in the subgroups of people discharged directly to home or in people discharged to nonhome settings (Table 2). However, in the multivariate model, MRCI had an inverse association with readmission for people who were discharged directly to home (HR = 0.67; 95% CI = 0.45-0.98). Specifically, for people discharged directly to home, the cumulative probability of not being readmitted to hospital was highest in the quartile of patients with the highest MRCI (Figure 1B). In people discharged to nonhome settings, those with the highest MRCI had the lowest cumulative probability of not being readmitted to hospital (Figure 1C). In the multivariate model for people discharged directly to home, there was also an inverse association between number of medications and unplanned readmission (HR = 0.88; 95% CI = 0.79-0.99) but not with polypharmacy and unplanned readmission (HR = 0.59; 95% CI = 0.30-1.16). In the multivariate model for people discharged to nonhome settings, the association between MRCI and unplanned readmission did not achieve statistical significance (HR = 1.32; 95% CI = 0.98-1.78). Conversely, there was an association between number of medications and unplanned readmissions (HR = 1.12; 95% CI = 1.01-1.25) and polypharmacy and unplanned readmissions (HR = 2.24; 95% CI = 1.02-4.94).
Subgroup Analyses for People With Cognitive Impairment or Dementia In the unadjusted analyses, there was no association between MRCI and rehospitalization in people with or without cognitive impairment (Table 2). In the multivariate model, MRCI was not associated with unplanned rehospitalization in people with cognitive impairment or dementia (HR = 1.07; 95% CI = 0.79-1.45) or in people without cognitive impairment or dementia (HR = 0.83; 95% CI = 0.57-1.20). In the sensitivity analyses, there was no association between MRCI and rehospitalization when only people with a dementia diagnosis were considered (HR = 1.51; 95% CI = 0.84-2.73).
Discussion The main finding of this study was that medication regimen complexity was not associated with unplanned rehospitalization in older people discharged from a GEM unit. However, in patients discharged directly to a nonhome setting, the number of medications prescribed at hospital discharge and the presence of polypharmacy (ie, patient was taking ≥9 medications at hospital discharge) were associated with unplanned rehospitalization.
The number of medications at discharge was a determinant of rehospitalization in people discharged to nonhome settings. This finding was consistent with a recent systematic review that identified polypharmacy as a risk factor for hospitalization.4 The number of medications prescribed has been correlated with the risk of adverse drug reactions, which in turn, are a leading cause of preventable hospitalization in older people. Interestingly, in our study, the number of medications and polypharmacy predicted rehospitalization in nonhome settings, but medication regimen complexity did not. When people were discharged to a nonhome setting, the number of medications may be an equally or more important determinant of rehospitalization than regimen complexity because medication administration in residential aged care facilities is usually supervised and supported. Medications for residents of aged care facilities are usually packed into dose administration aids, and topical or other preparations may be administered by nurses or other trained staff.23 Another potential reason as to why the number of medications and polypharmacy but not medication regimen complexity predicted rehospitalization could be related to the particular medications prescribed. Warfarin, insulin, oral antiplatelet medications, oral hypoglycemic medications, anticholinergic medications, sedative medications, and Beers Criteria medications have been associated with hospitalization in previous studies.5-7 Whereas warfarin and insulin often have complex dosing regimens, oral antiplatelet medications and sedative medications are often dosed once daily. Therefore, these medications are relatively minor contributors to regimen complexity but are associated with a high risk of unplanned rehospitalization through events such as bleeding and falls.This may contribute to why polypharmacy was a stronger predictor of rehospitalization than regimen complexity. Despite the association between polypharmacy and rehospitalization in our study, caution is needed before adopting polypharmacy as an indicator of medication inappropriateness more generally. This is because polypharmacy has been associated with underprescribing of clinically indicated medications,17 and it may not predict hospitalization in primary care patients with a large number of comorbidities.24 The inverse association between regimen complexity and rehospitalization in patients discharged directly to home was unexpected. This may have been a result of the special nature of our study participants. Patients discharged directly to home with a complex regimen may have had a system in place to support medication taking (eg, vigilant family members and dose administration aids). The inverse association between regimen complexity and rehospitalization suggests that older people with relatively simple medication regimens may still require support to manage their medications.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
8
Annals of Pharmacotherapy
Another possible explanation is that patients’ medication regimens may be simplified via the discontinuation or “deprescribing” of medications.25 This may have occurred because clinicians believed that the likely benefits of specific medications no longer outweighed the higher risk of adverse drug events resulting from factors such as agerelated renal function decline, increasing frailty, or multimorbidity. People in the final stages of life may be discharged to home with relatively simple medication regimens but may still be prone to unplanned rehospitalization. However, it is unclear to what extent this might have explained the inverse association because the GEM unit where this study was conducted selected patients for admission with a view to discharging them home. We had hypothesized that MRCI would be associated with rehospitalization in people with cognitive impairment, who often need support in medication taking.26,27 However, no association between MRCI and rehospitalization was identified in this subgroup. Further research with larger sample sizes should be conducted on this topic. It is possible that older people discharged directly to home with relatively simple medication regimens may still require help managing their medications. Should further studies support the value of the MRCI, it may be suitable for inclusion in prescribing and dispensing software as a prompt for clinicians.
Strengths and Limitations The strengths of this study include the medication data being extracted by an experienced pharmacist researcher in order to minimize the likelihood of error.28 Validated scales were used to assess comorbidities,18 ADLs,20 depression,19 cognition,21,22 and medication regimen complexity.11 Patients admitted to the GEM unit were older people (mean age 85.15 ± 6.4 years), and this age group is underrepresented in other research in this area. Data on unplanned rehospitalization was comprehensive because it was obtained from the 3 major hospitals in Central-Northern and South-Western Adelaide. Similarly, the dates of death were accurate because they were obtained for all patients who died within the 12-month follow-up period from the official register of Births, Deaths and Marriages. A readmission period of 12 months compensated for potentially seasonal fluctuations in hospital admissions. Medication regimen complexity may reflect patients’ overall health condition because people who are more ill or otherwise in need of support may be prescribed more complex regimens. We accounted for this by adjusting our regression models for age, sex, ADLs, comorbidities, cognition, and depression. However, as in all observational studies, there remains the possibility of confounding. The study may have been underpowered to detect an association between regimen complexity and unplanned
rehospitalization. However, the sample size far exceeded the minimum sample size of 29, which was calculated based on earlier research.14 In addition, our study identified an association between a number of other medication and non– medication-related parameters and rehospitalization. Social support or the use of dose administration aids can have an impact on unplanned rehospitalization.29 However, we could not account for these factors. In the absence of validated cutoffs for MRCI, we analyzed MRCI as a continuous variable . An additional potential limitation was the reliance on patients’ hospital medical records because these are often incomplete.30 Yet data utilized in the study were the same data forwarded to the patients’ primary care physicians for the purpose of medical decision making and repeat prescribing postdischarge. Generalizability may be limited because study participants were discharged from a single GEM unit. The GEM unit selects patients and, therefore, our participants may not be representative of all hospitalized patients. However, patients’ characteristics (MRCI scores, sex distribution, and comorbidity) were comparable with those reported in earlier research.15
Conclusion Medication regimen complexity was not associated with unplanned rehospitalization in older people discharged from a GEM unit. However, in patients discharged directly to a nonhome setting, the number of medications prescribed at hospital discharge and polypharmacy were associated with unplanned rehospitalization. A patient’s discharge destination can impact the association between medicationrelated factors and rehospitalization. These factors should be considered in future analyses on the risk of rehospitalization performed in larger studies of people discharged from hospital. Acknowledgments The authors would like to thank Ms Anne-Marie Young, Casemix manager, The Queen Elizabeth Hospital, for her support in the acquisition of readmission data and Ms Cassie Hewton for her support at the GEM unit.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
Authors’ Note Barbara C. Wimmer received a Monash Graduate Scholarship and an International President’s Scholarship from the University of
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
9
Wimmer et al South Australia to undertake the research reported in this manuscript.
References 1. Australian Institute of Health and Welfare. Australia’s Welfare 2013. Australia’s Welfare No. 11. Cat. No. AUS 174. Australian Hospital Statistics 2011–12. Health Services Series No. 50. Cat. No. HSE 134. Canberra, Australia: AIHW and AIWH; 2013. 2. NHS Health and Social Care Information Centre Hospital Episode Statistics: Headline Figures 2011-12. http://www. hesonline.nhs.uk Accessed 18 Dec 2013 3. US Department of Health and Human Services, Agency for Healthcare Research and Quality. National and Regional Estimates on Hospital Use for All Patients From the HCUP Nationwide Inpatient Sample (NIS), 2011: National Statistics. http://hcupnet.ahrq.gov/ Accessed 18 Dec 2013 4. Al Hamid A, Ghaleb M, Aljadhey H, Aslanpour Z. A systematic review of hospitalisation resulting from medicine related problems in adult patients [published online November 28, 2013]. Br J Clin Pharmacol. doi:10.1111/ bcp.12293. 5. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. 2011;365:2002-2012. 6. Gnjidic D, Hilmer SN, Hartikainen S, et al. Impact of high risk drug use on hospitalization and mortality in older people with and without Alzheimer’s disease: A National Population Cohort Study. PLoS One. 2014;9:e83224. doi:10.1371/journal.pone.0083224. 7. Price SD, Holman CAJ, Sanfilippo FM, Emery JD. Association between potentially inappropriate medications from the Beers criteria and the risk of unplanned hospitalization in elderly patients. Ann Pharmacother. 2014;84:6-16. 8. Nazir A, LaMantia M, Chodosh J, et al. Interaction between cognitive impairment and discharge destination and its effect on rehospitalization. J Am Geriatr Soc. 2013;61:19581963. 9. Dobrzanska L, Newell R. Readmissions: a primary care examination of reasons for readmission of older people and possible readmission risk factors. J Clin Nurs. 2006;15:599-606. 10. Marcantonio ER, McKean S, Goldfinger M, Kleefield S, Yurkofsky M, Brennan TA. Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan. Am J Med. 1999;107:13-17. 11. George J, Phun YT, Bailey MJ, Kong DC, Stewart K. Development and validation of the medication regimen complexity index. Ann Pharmacother. 2004;38:1369-1376. 12. Elliott RA, Ross-Degnan D, Adams AS, Safran DG, Soumerai SB. Strategies for coping in a complex world: adherence behavior among older adults with chronic illness. J Gen Intern Med. 2007;22:805-810. 13. Willson MN, Greer CL, Weeks DL. Medication regimen complexity and hospital readmission for an adverse drug event. Ann Pharmacother. 2014;48:26-32.
14. Dierich MT, Mueller C, Westra BL. Medication regimens in older home care patients. J Gerontol Nurs. 2011;37(12):45-55. 15. Libby AM, Fish DN, Hosokawa PW, et al. Patient-level medication regimen complexity across populations with chronic disease. Clin Ther. 2013;35:385-398.e1. 16. Hanlon JT, Schmader KE, Ruby CM, Weinberger M. Suboptimal prescribing in older inpatients and outpatients. J Am Geriatr Soc. 2001;49:200-209. 17. Kuijpers MA, Van Marum RJ, Egberts AC, Jansen PA. Relationship between polypharmacy and underprescribing. Br J Clin Pharmacol. 2007;65:130-133. 18. Beloosesky Y, Weiss A, Mansur N. Validity of the medication-based Disease Burden Index compared with the Charlson Comorbidity Index and the Cumulative Illness Rating Scale for geriatrics: a cohort study. Drugs Aging. 2011;28:1007-1014. 19. Lesher EL, Berryhill JS. Validation of the Geriatric Depression Scale–Short Form among inpatients. J Clin Psychol. 1994;50:256-260. 20. Hartigan I. A comparative review of the Katz ADL and the Barthel Index in assessing the activities of daily living of older people. Int J Older People Nurs. 2007;2:204-212. 21. Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: a comprehensive review. J Am Geriatr Soc. 1992;40:922-935. 22. Rowland JT, Basic D, Storey JE, Conforti DA. The Rowland Universal Dementia Assessment Scale (RUDAS) and the Folstein MMSE in a multicultural cohort of elderly persons. Int Psychogeriatr. 2006;18:111-120. 23. Bell JS, Johnell K, Wimmer BC, Wiese MD. Multidose drug dispensing and optimising drug use in older people. Age Ageing. 2013;42:556-558. 24. Payne RA, Abel G, Avery A, Mercer S, Roland M. Is polypharmacy always hazardous? A retrospective cohort analysis using linked electronic health records from primary and secondary care [published online January 15, 2014]. Br J Clin Pharmacol. doi:10.1111/bcp.12292. 25. Hardy JE, Hilmer SN. Deprescribing in the last year of life. J Pharm Pract Res. 2011;41:146-151. 26. Cotrell V, Wild K, Bader T. Medication management and adherence among cognitively impaired older adults. J Gerontol Soc Work. 2006;47(3-4):31-46. 27. Adams R, May H, Swift L, Bhattacharya D. Do older patients find multi-compartment medication devices easy to use and which are the easiest? Age Ageing. 2013;42:715-720. 28. Meyer C, Stern M, Woolley W, Jeanmonod R, Jeanmonod D. How reliable are patient-completed medication reconciliation forms compared with pharmacy lists? Am J Emerg Med. 2011;30:1048-1054. 29. Rodríguez-Artalejo F, Guallar-Castillón P, Herrera MC, et al. Social network as a predictor of hospital readmission and mortality among older patients with heart failure. J Card Fail. 2006;12:621-627. 30. Lau HS, Florax C, Porsius AJ, De Boer A. The completeness of medication histories in hospital medical records of patients admitted to general internal medicine wards. Br J Clin Pharmacol. 2000;49:597-603.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
research-article2014
AOPXXX10.1177/1060028014537469Annals of PharmacotherapyWimmer et al
Article
Medication Regimen Complexity and Unplanned Hospital Readmissions in Older People
Annals of Pharmacotherapy 1–9 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1060028014537469 aop.sagepub.com
Barbara C. Wimmer, MSc1,2, Elsa Dent, PhD3,4, J. Simon Bell, PhD1,2, Michael D. Wiese, PhD2, Ian Chapman, PhD3, Kristina Johnell, PhD4, 5, and Renuka Visvanathan, PhD3,6
Abstract Background: Medication-related problems and adverse drug events are leading causes of preventable hospitalizations. Few previous studies have investigated the possible association between medication regimen complexity and unplanned rehospitalization. Objective: To investigate the association between discharge medication regimen complexity and unplanned rehospitalization over a 12-month period. Methods: The prospective study comprised patients aged ≥70 years old consecutively admitted to a Geriatrics Evaluation and Management (GEM) unit between October 2010 and December 2011. Medication regimen complexity at discharge was calculated using the 65-item validated Medication Regimen Complexity Index (MRCI). Cox proportional-hazards regression was used to compute unadjusted and adjusted hazard ratios (HRs) with 95% CIs for factors associated with rehospitalization over a 12-month follow-up period. Results: Of 163 eligible patients, 99 patients had one or more unplanned hospital readmissions. When adjusting for age, sex, activities of daily living, depression, comorbidity, cognitive status, and discharge destination, MRCI (HR = 1.01; 95% CI = 0.81-1.26), number of discharge medications (HR = 1.01; 95% CI = 0.94-1.08), and polypharmacy (≥9 medications; HR = 1.12; 95% CI = 0.69-1.80) were not associated with rehospitalization. In patients discharged to nonhome settings, there was an association between rehospitalization and the number of discharge medications (HR = 1.12; 95% CI = 1.01-1.25) and polypharmacy (HR = 2.24; 95% CI = 1.02-4.94) but not between rehospitalization and MRCI (HR = 1.32; 95% CI = 0.98-1.78). Conclusion: Medication regimen complexity was not associated with unplanned hospital readmission in older people. However, in patients discharged to nonhome settings, the number of discharge medications and polypharmacy predicted rehospitalization. A patient’s discharge destination is an important factor in unplanned medication-related readmissions. Keywords medication regimen complexity, hospital readmission, elderly
Introduction Australians aged ≥65 years comprise 14% of the population but account for 39% of hospital admissions.1 These figures are comparable to those in the United Kingdom and United States where people aged ≥65 years account for 39% and 36% of hospital admissions and discharges, respectively.2,3 Medication-related problems and adverse drug events are leading causes of preventable hospital admissions.4,5 Anticholinergic and sedative medications, warfarin, insulin, and oral antiplatelet and oral hypoglycemic medication have been associated with an increased risk of hospitalization.5,6 Use of potentially inappropriate medications and polypharmacy have also been associated with hospital admissions.4,7 The risk of medication-related hospitalization may be increased in people with other risk factors for
1
Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical Sciences, Monash University, Melbourne, Australia 2 Sansom Institute, School of Pharmacy and Medical Sciences, University of South Australia, Adelaide, Australia 3 Discipline of Medicine, University of Adelaide, Adelaide, Australia 4 Discipline of Public Health, University of Adelaide, Adelaide, Australia 5 Aging Research Center, Karolinska Institutet and Stockholm University, Stockholm, Sweden 6 Aged and Extended Care Services, The Queen Elizabeth Hospital and the Adelaide Geriatrics Training and Research with Aged Care (GTRAC) Centre, University of Adelaide, Adelaide, Australia Corresponding Author: Barbara C. Wimmer, Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical Sciences, Monash University, 381 Royal Parade, Parkville, VIC, Australia. Email: [email protected]
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
2
Annals of Pharmacotherapy
hospitalization, including advanced age, ≥5 comorbidities, cognitive impairment, living in a nonhome setting, depression, and recent admission.8-10 A link between polypharmacy and medication-related hospital admission has been demonstrated in older people.4 However, the number of medications taken by a patient is only one element of regimen complexity. Other elements that contribute to regimen complexity include the number of different dose forms, dosing frequencies, and special instructions for medication use.11 These additional elements may affect whether or not an older person can manage their medication regimen at home. Having a high medication regimen complexity has also been associated with lower medication adherence.12 However, few previous studies have looked at the complexity of medication regimens and rehospitalization.13,14 The objective of this study was to investigate the association between discharge medication regimen complexity and unplanned hospital readmission over a 12-month period. We hypothesized that medication regimen complexity would be associated with unplanned rehospitalization, particularly in people with cognitive impairment and in those discharged directly to home. This was because people with cognitive impairment or dementia and those discharged to home may experience difficulties with or lack of support in managing their medications.
Methods Setting and Participants The setting of this prospective study was the Geriatrics Evaluation and Management (GEM) unit of a public hospital in Adelaide, South Australia. The study included patients aged ≥70 years admitted consecutively to the GEM unit between October 22, 2010, and December 23, 2011. Potential participants were excluded if a proxy was required but unavailable (eg, in case of dementia, unresolved delirium within 72 hours, or language barriers), their treating clinician recommended exclusion (eg, as a result of elder abuse, physical aggression, or being medically unwell), they were deemed infectious, they were missed by the researcher, they declined to participate, or they were transferred to another hospital or to palliative care service. Of the 166 included participants discharged from the GEM unit, 2 were transferred to another hospital and 1 to palliative care service. Thus, this study investigated rehospitalization among 163 patients managed and discharged from the GEM unit.
Data Collection Participants were recruited within 72 hours of their admission. All physiological assessment data were collected by the same researcher (ED). Discharge medication data were extracted directly from the hospital separation summary
prepared predominantly by medical practitioners and recorded in the Open Architecture Clinical Information System (OACIS, South Australia Health Department 2009). At the GEM unit, there is a full-time clinical pharmacist, and discharge medication lists are routinely provided to both the patients and their primary care physician to inform ongoing prescribing. All medication data were extracted from OACIS by the same pharmacist researcher (BCW).
Medication Assessment Each patient’s medication regimen was assessed at hospital discharge. Medications were categorized according to the Anatomical Therapeutic Chemical Classification System recommended by the World Health Organization. Each patient’s regimen complexity was calculated using the original 65-item validated Medication Regimen Complexity Index (MRCI).11 The MRCI considers dosage forms, dosing frequencies, and additional directions, with higher MRCI scores reflecting more complex medication regimens.11 Prescription and nonprescription medications, nutritional supplements, health products, dermatological preparations, and short-term medications (eg, antibiotics) were all considered when computing the MRCI. The inclusion of nonprescription medications was consistent with other recent studies utilizing the MRCI.14,15 The numbers of medications at hospital discharge were counted. Polypharmacy was defined as the use of 9 or more medications on a regular or as-needed basis. This definition was consistent with that recommended for use in long-term facilities.16 The more conservative commonly used definition of 5 or more medications was not practical to use because almost all participants (n = 155) took 5 or more medications.17
Main Outcome Measure The main outcome measure was unplanned rehospitalization in the 12 months following hospital discharge. The 12-month follow-up period was defined individually for each patient using the date of their first discharge as the index date. Each patient was only included once in the study. Data on unplanned rehospitalization to emergency departments over a 12-month follow-up period were extracted from the administrative records of the 3 major acute care hospitals serving the Central-Northern regions of metropolitan Adelaide by hospital case-mix staff. Dates of death were obtained for all patients who died within the 12-month follow-up period from the official register of Births, Deaths and Marriages.
Covariates The Charlson’s Comorbidity Index was calculated to assess comorbidity for each patient. This index is a weighted index that is suitable for use with medical records.18 Depressive
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
3
Wimmer et al symptoms were assessed using the Geriatric Depression Scale (GDS-15).19 Activities of daily living (ADLs) were computed using the 10-item Barthel’s ADL index. This is a performance scale with possible scores ranging from 0 to 100, with higher scores indicating greater independence.20 Discharge destination data were extracted from patients’ separation summaries. Patients were categorized as being “discharged directly to home” if they returned to their own home immediately after hospital discharge, both with or without new community services such as Meals on Wheels, domiciliary care, or home and community care services. Patients were considered “discharged to non-home settings” if they were discharged to either high- or low-level residential aged care, transition care (ie, restorative care in a residential aged care facility), or off-site inpatient rehabilitation. For the purpose of the analyses, patients were defined as having cognitive impairment if they had an inhospital Mini Mental State Examination (MMSE)21 or Rowland Universal Dementia Assessment Scale (RUDAS)22 score ≤24 or a clinical diagnosis of dementia recorded in the hospital discharge summary.
Statistical Analyses An a priori sample size calculation was performed based on a previous study, which reported that 85% of patients with an MRCI >20 were readmitted over 12 months compared with 15% of patients with an MRCI ≤20.14 Based on these findings, assuming a power of 90% and α = 0.05, a sample size of 29 patients would be required to detect a significant difference in rehospitalization. Study data were summarized using means, medians, and SDs. Shapiro-Wilk’s test (P > 0.05) and visual inspection of histograms, Q-Q plots, and box plots were used to test for normality of the data. Normally distributed characteristics were summarized using means with SDs. Nonnormally distributed variables were reported using medians and ranges. Dichotomous variables were reported as numbers and percentages. χ2 Tests were used to investigate potential differences between categorical variables, t tests were used for normally distributed continuous variables, and Mann Whitney U tests for nonnormally distributed variables. Two-sided tests were used to investigate potential differences between patients who were readmitted and those who were not readmitted during the 12-month follow-up period. Kaplan-Meier survival curves were computed to visually depict the relationship between MRCI divided in quartiles and time to rehospitalization. Univariate Cox proportionalhazards regression was then used to compute hazard ratios (HRs) with 95% CIs for factors associated with unplanned rehospitalization. Variables that were associated with unplanned readmission in the unadjusted analyses (defined as P < 0.1) or those deemed clinically relevant on the basis of previous research4,8-10 were included in multivariate models. Variables in each model were checked for
multicollinearity. Age, ADLs, depression, comorbidities, and number of medications were analyzed as continuous variables. The total MRCI score was divided by 10 and analysed as a continuous variable. Sex, the presence of cognitive impairment or dementia, and polypharmacy were analyzed as dichotomized variables. To account for deaths that occurred during the follow-up period, the analyses were censored at each patient’s date of death or the end of the 12-month follow-up period, whichever occurred first. Subgroup analyses were performed stratifying participants into (1) patients with and without dementia or cognitive impairment and (2) patients discharged directly to home or to nonhome settings. The stratifications were performed because previous research has demonstrated that both cognitive status and discharge destination can have an impact on readmission rates.8 Sensitivity analyses were performed to investigate the association between MRCI and rehospitalization when only participants with a dementia diagnosis were considered (as opposed to either a dementia diagnosis or MMSE/RUDAS ≤24). Data were analyzed using the Statistical Package for the Social Sciences (SPSS 20, SPSS Inc, Chicago, IL).
Ethical Considerations All potential participants were provided with written information about the study. Written informed consent to participate was obtained from all participants themselves or their health proxy. The study was approved by the Human Research Ethics Committees at The Queen Elizabeth Hospital and the University of South Australia.
Results Main Results For the 163 patients discharged from the GEM unit, the mean MRCI was 27.86 (SD = 11.63), and the mean number of medications was 9.7 (SD = 3.60); 43 had an MRCI of ≤20, and 40 patients had an MRCI of >35., In total, 99 (61%) patients had one or more unplanned rehospitalizations within 12 months of being discharged. Among the 99 patients readmitted, the mean MRCI was 28.01 (SD = 12.48), the mean number of medications was 9.7 (SD = 3.70), 60 had cognitive impairment or dementia, 55 had been discharged directly to home, and 23 died during the 12-month follow-up. In total, 64 patients were not readmitted to hospital. Among these patients, the mean MRCI was 27.62 (SD = 10.26), the mean number of medications was 9.7 (SD = 3.48), 23 had cognitive impairment or dementia, 32 had been discharged directly to home, and 7 died during the 12-month follow-up (Table 1). The MRCI was not significantly different in patients who were readmitted and not readmitted over 12 months (mean difference = −0.39; 95% CI = −4.09 to 3.30). People
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
4
Annals of Pharmacotherapy
Table 1. Characteristics of the Patients With and Without Unplanned Hospital Readmission Over the 12-Month Follow-up Stratified by Cognitive Function and Discharge Destination.a All Patients (n = 163) Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) DD directly to home n (%) GDS-15 (range) Cognitive impairment or dementia n (%) Cognitive Impairment or Dementia (n = 83)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) DD directly to home, n (%)
No Cognitive Impairment or Dementia (n = 80)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) DD directly to home, n (%)
DD Directly to Home (n = 87)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) Cognitive impairment or dementia, n (%)
DD to Nonhome Settings (n = 76)
Age (SD) Female sex, n (%) MRCI (SD) Barthel’s ADL index (range) CCI (range) GDS-15 (range) Cognitive impairment or dementia, n (%)
Readmitted (n = 99)
Not Readmitted (n = 64)
84.9 (6.20) 68 (68.7) 28.01 (12.48) 68.5 (8-100) 3.0 (0-10) 55 (55.6) 4.0 (0-15) 60 (60.6) Readmitted (n = 60)
85.6 (6.74) 50 (78.1) 27.62 (10.26) 76.0 (21-100) 2.0 (0-8) 32 (50.0) 3.0 (0-15) 23 (35.9) Not Readmitted (n = 23)
84.4 (6.33) 39 (65.0) 27.98 (11.47) 65.0 (8-100) 3.0 (0-10) 4.0 (0-15) 31 (51.7) Readmitted (n = 39)
85.04 (7.33) 18 (78.3) 25.70 (10.75) 71.0 (31-91) 2.0 (0-8) 3.0 (0-15) 8 (34.8) Not Readmitted (n = 41)
85.7 (5.98) 29 (74.4) 28.06 (14.05) 72.0 (19-94) 3.0 (0-7) 4.0 (0-14) 24 (61.5) Readmitted (n = 55)
85.9 (6.47) 32 (78.0) 28.70 (9.94) 81.0 (21-100) 2.0 (0-8) 3.5 (0-12) 24 (58.5) Not Readmitted (n = 32)
84.5 (6.75) 38 (69.1) 24.91 (9.78) 73.0 (23-94) 3.0 (0-10) 4.0 (0-13) 31 (56.4)
84.7 (7.15) 21 (65.6) 28.08 (9.42) 84.5 (21-100) 2.0 (0-7) 3 (0-15) 8 (25.0)
P Value 0.50 0.19 0.83 0.006b 0.17 0.49 0.44 0.002b P Value 0.68 0.24 0.41 0.08 0.50 0.76 0.17 P Value 0.88 0.70 0.82 0.15 0.27 0.61 0.78 P Value 0.89 0. 74 0.14 20 to 27 MRCI: >27 to 35 MRCI: >35
0.2
0.4
Cumulative probability of not being readmitted
MRCI: 1 to 20 MRCI: >20 to 27 MRCI: >27 to 35 MRCI: >35
0.6
0.8
MRCI: 1 to 20 MRCI: >20 to 27 MRCI: >27 to 35 MRCI: >35
1.0
B
0.8
1.0
A
0
60
120
180
240
300
360
Time (days) Figure 1. Kaplan-Meier survival curves on MRCI (quartiles) and unplanned hospital readmission for (A) all patients, (B) patients discharged directly to home, (C) patients discharged to nonhome settings, and number of patients (n) in each subgroup. MRCI Quartile 1st 2nd 3rd 4th Total
All Patients (n)
DD Directly to Home (n)
DD Nonhome Settings (n)
43 41 39 40 163
24 25 25 13 87
19 16 14 27 76
Abbreviations: MRCI, Medication Regimen Complexity Index; DD, discharge destination; MRCI 1st quartile, 1-20; 2nd quartile, 20.1-27; 3rd quartile, 27.1-35; 4th quartile, 35.1-73.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
7
Wimmer et al
Subgroup Analyses for Discharge Destination In the unadjusted analyses, there was no association between MRCI and rehospitalization in the subgroups of people discharged directly to home or in people discharged to nonhome settings (Table 2). However, in the multivariate model, MRCI had an inverse association with readmission for people who were discharged directly to home (HR = 0.67; 95% CI = 0.45-0.98). Specifically, for people discharged directly to home, the cumulative probability of not being readmitted to hospital was highest in the quartile of patients with the highest MRCI (Figure 1B). In people discharged to nonhome settings, those with the highest MRCI had the lowest cumulative probability of not being readmitted to hospital (Figure 1C). In the multivariate model for people discharged directly to home, there was also an inverse association between number of medications and unplanned readmission (HR = 0.88; 95% CI = 0.79-0.99) but not with polypharmacy and unplanned readmission (HR = 0.59; 95% CI = 0.30-1.16). In the multivariate model for people discharged to nonhome settings, the association between MRCI and unplanned readmission did not achieve statistical significance (HR = 1.32; 95% CI = 0.98-1.78). Conversely, there was an association between number of medications and unplanned readmissions (HR = 1.12; 95% CI = 1.01-1.25) and polypharmacy and unplanned readmissions (HR = 2.24; 95% CI = 1.02-4.94).
Subgroup Analyses for People With Cognitive Impairment or Dementia In the unadjusted analyses, there was no association between MRCI and rehospitalization in people with or without cognitive impairment (Table 2). In the multivariate model, MRCI was not associated with unplanned rehospitalization in people with cognitive impairment or dementia (HR = 1.07; 95% CI = 0.79-1.45) or in people without cognitive impairment or dementia (HR = 0.83; 95% CI = 0.57-1.20). In the sensitivity analyses, there was no association between MRCI and rehospitalization when only people with a dementia diagnosis were considered (HR = 1.51; 95% CI = 0.84-2.73).
Discussion The main finding of this study was that medication regimen complexity was not associated with unplanned rehospitalization in older people discharged from a GEM unit. However, in patients discharged directly to a nonhome setting, the number of medications prescribed at hospital discharge and the presence of polypharmacy (ie, patient was taking ≥9 medications at hospital discharge) were associated with unplanned rehospitalization.
The number of medications at discharge was a determinant of rehospitalization in people discharged to nonhome settings. This finding was consistent with a recent systematic review that identified polypharmacy as a risk factor for hospitalization.4 The number of medications prescribed has been correlated with the risk of adverse drug reactions, which in turn, are a leading cause of preventable hospitalization in older people. Interestingly, in our study, the number of medications and polypharmacy predicted rehospitalization in nonhome settings, but medication regimen complexity did not. When people were discharged to a nonhome setting, the number of medications may be an equally or more important determinant of rehospitalization than regimen complexity because medication administration in residential aged care facilities is usually supervised and supported. Medications for residents of aged care facilities are usually packed into dose administration aids, and topical or other preparations may be administered by nurses or other trained staff.23 Another potential reason as to why the number of medications and polypharmacy but not medication regimen complexity predicted rehospitalization could be related to the particular medications prescribed. Warfarin, insulin, oral antiplatelet medications, oral hypoglycemic medications, anticholinergic medications, sedative medications, and Beers Criteria medications have been associated with hospitalization in previous studies.5-7 Whereas warfarin and insulin often have complex dosing regimens, oral antiplatelet medications and sedative medications are often dosed once daily. Therefore, these medications are relatively minor contributors to regimen complexity but are associated with a high risk of unplanned rehospitalization through events such as bleeding and falls.This may contribute to why polypharmacy was a stronger predictor of rehospitalization than regimen complexity. Despite the association between polypharmacy and rehospitalization in our study, caution is needed before adopting polypharmacy as an indicator of medication inappropriateness more generally. This is because polypharmacy has been associated with underprescribing of clinically indicated medications,17 and it may not predict hospitalization in primary care patients with a large number of comorbidities.24 The inverse association between regimen complexity and rehospitalization in patients discharged directly to home was unexpected. This may have been a result of the special nature of our study participants. Patients discharged directly to home with a complex regimen may have had a system in place to support medication taking (eg, vigilant family members and dose administration aids). The inverse association between regimen complexity and rehospitalization suggests that older people with relatively simple medication regimens may still require support to manage their medications.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
8
Annals of Pharmacotherapy
Another possible explanation is that patients’ medication regimens may be simplified via the discontinuation or “deprescribing” of medications.25 This may have occurred because clinicians believed that the likely benefits of specific medications no longer outweighed the higher risk of adverse drug events resulting from factors such as agerelated renal function decline, increasing frailty, or multimorbidity. People in the final stages of life may be discharged to home with relatively simple medication regimens but may still be prone to unplanned rehospitalization. However, it is unclear to what extent this might have explained the inverse association because the GEM unit where this study was conducted selected patients for admission with a view to discharging them home. We had hypothesized that MRCI would be associated with rehospitalization in people with cognitive impairment, who often need support in medication taking.26,27 However, no association between MRCI and rehospitalization was identified in this subgroup. Further research with larger sample sizes should be conducted on this topic. It is possible that older people discharged directly to home with relatively simple medication regimens may still require help managing their medications. Should further studies support the value of the MRCI, it may be suitable for inclusion in prescribing and dispensing software as a prompt for clinicians.
Strengths and Limitations The strengths of this study include the medication data being extracted by an experienced pharmacist researcher in order to minimize the likelihood of error.28 Validated scales were used to assess comorbidities,18 ADLs,20 depression,19 cognition,21,22 and medication regimen complexity.11 Patients admitted to the GEM unit were older people (mean age 85.15 ± 6.4 years), and this age group is underrepresented in other research in this area. Data on unplanned rehospitalization was comprehensive because it was obtained from the 3 major hospitals in Central-Northern and South-Western Adelaide. Similarly, the dates of death were accurate because they were obtained for all patients who died within the 12-month follow-up period from the official register of Births, Deaths and Marriages. A readmission period of 12 months compensated for potentially seasonal fluctuations in hospital admissions. Medication regimen complexity may reflect patients’ overall health condition because people who are more ill or otherwise in need of support may be prescribed more complex regimens. We accounted for this by adjusting our regression models for age, sex, ADLs, comorbidities, cognition, and depression. However, as in all observational studies, there remains the possibility of confounding. The study may have been underpowered to detect an association between regimen complexity and unplanned
rehospitalization. However, the sample size far exceeded the minimum sample size of 29, which was calculated based on earlier research.14 In addition, our study identified an association between a number of other medication and non– medication-related parameters and rehospitalization. Social support or the use of dose administration aids can have an impact on unplanned rehospitalization.29 However, we could not account for these factors. In the absence of validated cutoffs for MRCI, we analyzed MRCI as a continuous variable . An additional potential limitation was the reliance on patients’ hospital medical records because these are often incomplete.30 Yet data utilized in the study were the same data forwarded to the patients’ primary care physicians for the purpose of medical decision making and repeat prescribing postdischarge. Generalizability may be limited because study participants were discharged from a single GEM unit. The GEM unit selects patients and, therefore, our participants may not be representative of all hospitalized patients. However, patients’ characteristics (MRCI scores, sex distribution, and comorbidity) were comparable with those reported in earlier research.15
Conclusion Medication regimen complexity was not associated with unplanned rehospitalization in older people discharged from a GEM unit. However, in patients discharged directly to a nonhome setting, the number of medications prescribed at hospital discharge and polypharmacy were associated with unplanned rehospitalization. A patient’s discharge destination can impact the association between medicationrelated factors and rehospitalization. These factors should be considered in future analyses on the risk of rehospitalization performed in larger studies of people discharged from hospital. Acknowledgments The authors would like to thank Ms Anne-Marie Young, Casemix manager, The Queen Elizabeth Hospital, for her support in the acquisition of readmission data and Ms Cassie Hewton for her support at the GEM unit.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
Authors’ Note Barbara C. Wimmer received a Monash Graduate Scholarship and an International President’s Scholarship from the University of
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
9
Wimmer et al South Australia to undertake the research reported in this manuscript.
References 1. Australian Institute of Health and Welfare. Australia’s Welfare 2013. Australia’s Welfare No. 11. Cat. No. AUS 174. Australian Hospital Statistics 2011–12. Health Services Series No. 50. Cat. No. HSE 134. Canberra, Australia: AIHW and AIWH; 2013. 2. NHS Health and Social Care Information Centre Hospital Episode Statistics: Headline Figures 2011-12. http://www. hesonline.nhs.uk Accessed 18 Dec 2013 3. US Department of Health and Human Services, Agency for Healthcare Research and Quality. National and Regional Estimates on Hospital Use for All Patients From the HCUP Nationwide Inpatient Sample (NIS), 2011: National Statistics. http://hcupnet.ahrq.gov/ Accessed 18 Dec 2013 4. Al Hamid A, Ghaleb M, Aljadhey H, Aslanpour Z. A systematic review of hospitalisation resulting from medicine related problems in adult patients [published online November 28, 2013]. Br J Clin Pharmacol. doi:10.1111/ bcp.12293. 5. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. 2011;365:2002-2012. 6. Gnjidic D, Hilmer SN, Hartikainen S, et al. Impact of high risk drug use on hospitalization and mortality in older people with and without Alzheimer’s disease: A National Population Cohort Study. PLoS One. 2014;9:e83224. doi:10.1371/journal.pone.0083224. 7. Price SD, Holman CAJ, Sanfilippo FM, Emery JD. Association between potentially inappropriate medications from the Beers criteria and the risk of unplanned hospitalization in elderly patients. Ann Pharmacother. 2014;84:6-16. 8. Nazir A, LaMantia M, Chodosh J, et al. Interaction between cognitive impairment and discharge destination and its effect on rehospitalization. J Am Geriatr Soc. 2013;61:19581963. 9. Dobrzanska L, Newell R. Readmissions: a primary care examination of reasons for readmission of older people and possible readmission risk factors. J Clin Nurs. 2006;15:599-606. 10. Marcantonio ER, McKean S, Goldfinger M, Kleefield S, Yurkofsky M, Brennan TA. Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan. Am J Med. 1999;107:13-17. 11. George J, Phun YT, Bailey MJ, Kong DC, Stewart K. Development and validation of the medication regimen complexity index. Ann Pharmacother. 2004;38:1369-1376. 12. Elliott RA, Ross-Degnan D, Adams AS, Safran DG, Soumerai SB. Strategies for coping in a complex world: adherence behavior among older adults with chronic illness. J Gen Intern Med. 2007;22:805-810. 13. Willson MN, Greer CL, Weeks DL. Medication regimen complexity and hospital readmission for an adverse drug event. Ann Pharmacother. 2014;48:26-32.
14. Dierich MT, Mueller C, Westra BL. Medication regimens in older home care patients. J Gerontol Nurs. 2011;37(12):45-55. 15. Libby AM, Fish DN, Hosokawa PW, et al. Patient-level medication regimen complexity across populations with chronic disease. Clin Ther. 2013;35:385-398.e1. 16. Hanlon JT, Schmader KE, Ruby CM, Weinberger M. Suboptimal prescribing in older inpatients and outpatients. J Am Geriatr Soc. 2001;49:200-209. 17. Kuijpers MA, Van Marum RJ, Egberts AC, Jansen PA. Relationship between polypharmacy and underprescribing. Br J Clin Pharmacol. 2007;65:130-133. 18. Beloosesky Y, Weiss A, Mansur N. Validity of the medication-based Disease Burden Index compared with the Charlson Comorbidity Index and the Cumulative Illness Rating Scale for geriatrics: a cohort study. Drugs Aging. 2011;28:1007-1014. 19. Lesher EL, Berryhill JS. Validation of the Geriatric Depression Scale–Short Form among inpatients. J Clin Psychol. 1994;50:256-260. 20. Hartigan I. A comparative review of the Katz ADL and the Barthel Index in assessing the activities of daily living of older people. Int J Older People Nurs. 2007;2:204-212. 21. Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: a comprehensive review. J Am Geriatr Soc. 1992;40:922-935. 22. Rowland JT, Basic D, Storey JE, Conforti DA. The Rowland Universal Dementia Assessment Scale (RUDAS) and the Folstein MMSE in a multicultural cohort of elderly persons. Int Psychogeriatr. 2006;18:111-120. 23. Bell JS, Johnell K, Wimmer BC, Wiese MD. Multidose drug dispensing and optimising drug use in older people. Age Ageing. 2013;42:556-558. 24. Payne RA, Abel G, Avery A, Mercer S, Roland M. Is polypharmacy always hazardous? A retrospective cohort analysis using linked electronic health records from primary and secondary care [published online January 15, 2014]. Br J Clin Pharmacol. doi:10.1111/bcp.12292. 25. Hardy JE, Hilmer SN. Deprescribing in the last year of life. J Pharm Pract Res. 2011;41:146-151. 26. Cotrell V, Wild K, Bader T. Medication management and adherence among cognitively impaired older adults. J Gerontol Soc Work. 2006;47(3-4):31-46. 27. Adams R, May H, Swift L, Bhattacharya D. Do older patients find multi-compartment medication devices easy to use and which are the easiest? Age Ageing. 2013;42:715-720. 28. Meyer C, Stern M, Woolley W, Jeanmonod R, Jeanmonod D. How reliable are patient-completed medication reconciliation forms compared with pharmacy lists? Am J Emerg Med. 2011;30:1048-1054. 29. Rodríguez-Artalejo F, Guallar-Castillón P, Herrera MC, et al. Social network as a predictor of hospital readmission and mortality among older patients with heart failure. J Card Fail. 2006;12:621-627. 30. Lau HS, Florax C, Porsius AJ, De Boer A. The completeness of medication histories in hospital medical records of patients admitted to general internal medicine wards. Br J Clin Pharmacol. 2000;49:597-603.
Downloaded from aop.sagepub.com at University of Texas Libraries on June 5, 2014
