 |
 |
 |
 |
 |
 |
 |
|
|
|
|||||||||
|
|||||||||||
Clinical Pharmacist, Department of Geriatrics, Pharmacy Services, Rabin Medical Center, Beilinson Campus, Sackler School of Medicine, Tel Aviv University, Petach Tikvah, Israel
Senior Physician, Department of Geriatrics, Rabin Medical Center, Beilinson Campus, Sackler School of Medicine, Tel Aviv University, Petach Tikvah
Head, Department of Geriatrics, Rabin Medical Center, Beilinson Campus, Sackler School of Medicine, Tel Aviv University
Reprints: Dr. Beloosesky, Department of Geriatrics, Rabin Medical Center, Beilinson Campus, Petach Tikvah, Israel 49372, fax 972-3-937-6817, beloy{at}clalit.org.il
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
OBJECTIVES: To investigate the extent of in-hospital modification of medication regimens of elderly patients and its relationship to medication adherence as well as one-month postdischarge drug regimen modifications and to examine the relationship of the modifications, adherence, and polypharmacy to mortality and readmissions 3 months postdischarge.
METHODS: Clinical and demographic data, postdischarge medication modifications, and adherence were prospectively obtained in 212 elderly patients. Inhospital drug regimen modifications were retrospectively recorded.
RESULTS: The average ± SD in-hospital medication regimen modification rate was 49.8% ± 28.4. No modifications were found in 9.7% of the patients. Using demographic and clinical parameters, we performed regression analysis and found that patients who were admitted with polypharmacy, discharged home, and cognitively normal experienced fewer medication modifications (p < 0.05). At one month postdischarge, the average medication regimen modification rate was 37.5% ± 25.4. In- and posthospital modifications were directly correlated (p = 0.047). Three months postdischarge, 17 patients had died and 50 had been readmitted. The independent risk factors for mortality were in-hospital modification rate of 50% or greater (OR 6.4; 95% CI 1.3 to 29.7), impaired cognition (OR 4.2; 95% CI 1.4 to 12.3), and each chronic disease (OR 1.2; 95% CI 1 to 1.5). No relationships were found between in-hospital medication regimen modifications and readmissions or with postdischarge modifications, adherence, and polypharmacy to mortality and readmissions.
CONCLUSIONS: Hospitalization of elderly patients is characterized by extensive medication regimen modifications, which are directly correlated with postdischarge modifications and may indicate an increased risk of mortality.
Key Words: adherence, elderly, hospitalization, medication modifications, mortality
Published Online, April 29, 2008. www.theannals.com, DOI 10.1345/aph.1L070
Prescription drug expenditures are the fastest growing component of healthcare costs in the US. They are expected to increase by 6.7–10.3% per year during the coming decade.3,4 Interest in the elderly and medication management has grown, given that this population uses more than 30% of all medications in the US and other developed countries.5,6
During patients' admission to hospitals, 28–40% of medications are discontinued, and 45% of the drugs prescribed on discharge were started in the hospital.7,8 There is a decrease of 34% in the number (6.7) of daily drugs at admission versus the number (4.4) at discharge.9 After patient discharge, drug regimen maintenance may be difficult due to several reasons: lack of communication between the hospital and the community healthcare services; lack of hospital discharge papers; or insufficient detailed information regarding in-hospital drug modifications.7,10-13 According to a previous study, drug prescriptions for 55.5% of the patients were modified after admission; 38.7% were modified within the first 3 days after hospital discharge, and 37.8% were modified during the next 3 months.5 Other studies showed a 50% turnover in drug use between the general practitioner and the hospital.7 These modifications may place elderly patients at a higher risk for polypharmacy and its complications, such as poor adherence.8,11,14 Many patients continued to take discontinued medications; others showed no adherence or took the correct medications but incorrect doses.10,15 It is widely recognized that poor adherence has a substantial worsening effect on disease, admissions, death, and increased healthcare costs in the US.14,16-20 Polypharmacy among the elderly has been associated with a higher risk of adverse events, higher healthcare costs, and increased hospitalization and mortality.21-23 It has also been associated with re-hospitalization and mortality after discharge.24-27
To our knowledge, no study has evaluated the effect of medication regimen modifications during hospitalization on continuance and adherence of elderly patients to their drug therapy after hospital discharge. Also, little is known about the relationship of medication modifications during hospitalization to posthospitalization or about the relationship of postdischarge adherence to readmissions and mortality. The objectives of this study were to investigate the extent of modifications in medication regimens during the hospital stays of a widely varied group of unselected elderly patients; its relationship to medication modifications and adherence 1 month postdischarge; and whether modifications, adherence, and polypharmacy were related to readmissions and mortality 3 months postdischarge.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
DATA COLLECTION
Information regarding medication modifications during hospitalization was
recorded retrospectively from patients' hospital files. Clinical and
demographic data, postdischarge medication modifications, and information on
adherence were obtained prospectively. Each patient was interviewed by a
pharmacist. Information collected during hospitalization included
sociodemographic characteristics, cognitive and functional status, and
diseases. Subsequently, one month postdischarge (33 ± 7.5 days),
home-dwelling patients or caregivers (including family members if patients
were demented) were interviewed by telephone. In addition, general
practitioners were occasionally interviewed to clarify reasons for medication
modifications. In the case of institutionalized patients, local physicians
were interviewed. The interview consisted of unlimited, nonthreatening, and
nonjudgmental questions that concentrated on the drug regimen, deviation from
the regimen upon discharge, and reasons for modifications. Only home-dwelling
patients were evaluated for determination of adherence; this was done only
when the regimen was clearly understood by patients or their caregivers.
Mortality and readmission data were collected from the national and hospital
computer registries.
MEASURES
The Mini-Mental State Examination (MMSE) was used to assess the patient's
cognitive
level.28
Most patients were assessed during the first days of hospitalization and then
prior to discharge. The higher score was recorded. The MMSE was performed
twice to evaluate patients while mentally stable; delirious patients were
excluded. Patients were classified into 3 cognitive groups: severely impaired
(<14 points out of 30), moderately impaired (15–24 points), and
normal (
25
points).29
The Katz Index of Activities of Daily Living was used to categorize the
patients into 3 functional groups: fully dependent, partially dependent, and
independent.30
Medication regimen modification was defined as the addition of a drug, increase in the dose, decrease of a dose, cessation of a drug, or a switch within the same pharmacologic group (eg, statins including simvastatin, pravastatin). The extent of modifications between prescribed medication at admission and on discharge was measured by the percentage of modified medications in each patient. For example, if an admitted patient was taking 6 drugs and 3 were modified during hospitalization, the percentage of modified medications would be 50%. The overall mean change was calculated by averaging the percentage of modified medications of all patients. The extent of regimen change between hospital-prescribed medication on discharge and medication taken one month later was measured in the same manner.
Based on patients' self-reporting, overall adherence was measured in 2 stages: as the fraction of pills taken as prescribed for each medication, and then as the average of this proportion of all drugs taken by each patient.18 The overall mean adherence was calculated by averaging the adherence of all patients. Because the overall adherence could conceal poor adherence to one or several drugs, we also investigated a patient's nonadherence to at least one drug. Nonadherence was defined as underadherence of 70% or less or overadherence of 110% or more. We chose a low cut-off level of 70% to detect a more serious pattern, as chosen in a previous study.14 Regarding the upper cut-off, many trials have defined overadherence as 120% or more.14 We adapted a stricter criterion to detect more cases, given that the elderly are usually underrather than overadherent to their drug therapy.
Polypharmacy was defined as taking 5 or more medications per day.22-24
STATISTICAL ANALYSIS
The t-test was used to test the relationship between normally
distributed continuous variables and categorical variables. The
2 test or Fisher's Exact test was used to examine the
relationship between categorical variables. Pearson correlation was used to
measure the relationship between 2 continuous variables. Multiple linear and
logistic regression analyses were performed to determine the independent
variables affecting the modifications in the medication regimen and mortality
3 months postdischarge, respectively. A p value less than 0.05 was considered
significant.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
Baseline characteristics of the study population are summarized in Table 1. Mean age was 81.1 ± 7.25 years (median 81; range 66–103). At admission, the patients had 6.3 ± 2.6 chronic diseases. The most common causes for hospital admission were neurologic and orthopedic (18% each), cardiovascular (16%), falls or disequilibrium (15%), hemato-oncological (11%), infections (10%), gastrointestinal (4%), urologic (1%), and other diseases, including social or nonmedical causes, such as abuse, neglect, or lack of appropriate social support (7%). Mean hospital length of stay was 10.7 ± 6.3 days (median 9 days; range 2–39).
|
Patients' medication treatment characteristics at the 3 points of observation are summarized in Table 2. The number of prescriptions was significantly lower (p < 0.01) on admission and at discharge compared with one month postdischarge. Polypharmacy at admission was more frequent among older patients: 72.3% in patients 85 years of age and older and just 48.8% in those 65–74 years of age (p = 0.035). Patients taking 5 or more medications per day at admission had more diseases than did those taking fewer medications (6.7 ± 2.6 vs 5.7 ± 2.7, respectively; p < 0.01).
|
IN-HOSPITAL MODIFICATIONS
The average medication modification rate in the 207 patients was 49.8%
± 28.4 (95% CI 45.9 to 53.7). In 9.7% of patients, there was no change
in medication treatment, while 23.7% of the patients had modifications in
75–100% of their medications (Figure
1). Patients admitted with polypharmacy had fewer medication
modifications during hospitalization than did those without polypharmacy
(45.8% ± 26.6 vs 57.7% ± 30.9, respectively; p < 0.01). Using
regression analysis on demographic and clinical parameters, we found that
patients who were admitted with polypharmacy, who were discharged home, and
who were cognitively normal experienced fewer medication regimen modifications
(p < 0.01, p = 0.024, p = 0.04, respectively, R2 = 0.08).
No relationships were found between medication regimen modifications and age,
sex, number of chronic diseases, and hospital length of stay.
|
RELATIONSHIP OF IN-HOSPITAL MODIFICATIONS TO ONE-MONTH POSTHOSPITAL MODIFICATIONS
The average medication regimen modifications in-hospital and one month
postdischarge were 49.2% ± 28.0% (95% CI 45.1 to 53.3) and 37.5%
± 25.4% (95% CI 33.8 to 41.2), respectively. As seen in
Figure 1, at one month
postdischarge, 16.7% of the patients had no modifications in their medication
treatment, while 8.1% had modifications in 75–100% of their medication.
Modifications while in the hospital and posthospital were directly correlated
(p = 0.047). A subgroup analysis found this correlation in patients who were
admitted from home (p = 0.026), for reasons other than cardiovascular diseases
(p = 0.032), or with polypharmacy (p = 0.011); in patients who had a caregiver
(p = 0.027) or who visited their general practitioner at least twice within
the first month of discharge (p < 0.01); and in community discharged
patients (p = 0.047).
RELATIONSHIP OF MODIFICATIONS, ADHERENCE, AND POLYPHARMACY TO MORTALITY AND READMISSIONS
During the 3 months after discharge, 17 patients died and 50 patients were
readmitted. Patients who died had more medication modifications during
hospitalization compared with those who survived (69.1% ± 23.7 vs 48.1%
± 28.1, respectively; p < 0.01), had more diseases (7.64 ±
3.79 vs 6.12 ± 2.4, respectively; p = 0.019), were more cognitively
impaired (53% vs 19%, respectively; p < 0.01), and tended to be more
functionally dependent (64.7% vs 39.5%, respectively; p = 0.08). Logistic
regression analysis found that in-hospital medication modifications of 50% or
more, impaired cognition, and each chronic disease were independent risk
factors for mortality 3 months postdischarge (OR 6.4, 95% CI 1.3 to 29.7, p =
0.018 vs OR 4.2, 95% CI 1.4 to 12.3, p < 0.01 vs OR 1.2, 95% CI 1.0 to 1.5,
p = 0.033, respectively).
No other relationships were found between in-hospital medication regimen modifications and readmissions, nor in one-month postdischarge modifications, adherence, and polypharmacy to mortality and readmissions 3 months after discharge.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
It is well known that medication regimen modifications occur during hospitalization.7-9 In our study, we found an average in-hospital medication modification rate of 49.8%, which is comparable with the rates of 55.5% and 60% found in previous studies of elderly patients.5,7 The average modification rate one month postdischarge was 37.5%, which is also comparable with results found elsewhere.7
It is not surprising that patients who were cognitively normal and were discharged home had fewer in-hospital medication regimen modifications; they were expected to be more medically stable, enabling their community discharge. Patients who were demented, frail, discharged to nursing homes, and had more diseases31 were more complicated clinically; thus, they required more interventions and more in-hospital medication modifications than did community-discharged patients.
One would expect to find more in-hospital drug regimen modifications in patients admitted with polypharmacy; however, this was contrary to our findings. Given that polypharmacy was more common in older patients and in patients with more chronic diseases, we hypothesized that, in these complicated cases, hospital physicians avoided making numerous medication modifications, thus preventing these patients from experiencing polypharmacy-related problems such as adverse reactions and drug–drug interactions.32,33
To the best of our knowledge, the direct correlation that we found between in-hospital and one-month posthospital medication modifications had not been previously reported. Patients who are discharged home, live in the community, have caregivers, have polypharmacy, and often visit their general practitioners are expected to be in a better medical state than are other patients, but they may also be subject to more postdischarge modifications to their drug regimens. Moreover, in patients admitted with polypharmacy,34 most of the adverse drug events would occur after discharge, thus leading to consequent medication modifications. In addition, the fact that posthospital medical errors become more common as the number of prescribed drugs increases15 may explain the subsequent increase in postdischarge medication modifications among patients admitted with polypharmacy.
No relationship was found between in-hospital medication regimen modifications and the overall average adherence or adherence to at least one drug. Nevertheless, in an earlier study (unpublished), we found that patients who were nonadherent to at least one drug had significantly more postdischarge medication modifications.
Polypharmacy in elderly patients is generally a significant predictor of hospitalization and death,23,26 especially after hospital discharge.24,25,27 However, we failed to find such an effect, perhaps due to the characteristics of the patients, the study sizes, the follow-up period, and the differences related to the definition of polypharmacy. There is no consensus regarding this definition. Some studies consider only the number of drugs23; others consider the clinical indications and effects of a given drug regimen, regardless of the number of drugs used.35
A substantial fraction of the elderly admissions were found to be due to nonadherence16,36; therefore, we expected nonadherence to be correlated to readmissions. Polypharmacy and several medical visits were previously found to be associated with a higher risk of hospitalization due to nonadherence.36 Other studies have found that, in general, adherence was not a predictor of concurrent or future hospitalizations or mortality among the elderly.37
The most important finding, which, to our knowledge, has not been described, is that in-hospital medication modifications may be an independent risk factor for mortality 3 months postdischarge. Are in-hospital modifications a derivative of a poor medical condition that necessitates extensive drug modifications, thus indicating an increased mortality risk, or do extensive modifications, hypothetically, have an independent and intrinsic effect on a precarious medical state, enabling its further deterioration and contributing to the patient's death? To answer these questions, a larger, prospective, interventional study with a control group is needed.
|
Limitations |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
|
Conclusions |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Mansur, A. Weiss, and Y. Beloosesky Is There an Association Between Inappropriate Prescription Drug Use and Adherence in Discharged Elderly Patients? Ann. Pharmacother., February 1, 2009; 43(2): 177 - 184. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
 |
 |
 |
 |
