International Journal of Clinical Pharmacy

, Volume 35, Issue 2, pp 217–224

Impact of an intervention to reduce medication regimen complexity for older hospital inpatients

Authors

    • Pharmacy DepartmentAustin Health
    • Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical SciencesMonash University
  • Christopher O’Callaghan
    • Department of General MedicineAustin Health
  • Eldho Paul
    • Department of Epidemiology & Preventive Medicine, School of Public Health and Preventive Medicine, Faculty of Medicine, Nursing & Health SciencesMonash University
  • Johnson George
    • Centre for Medicine Use and Safety, Faculty of Pharmacy and Pharmaceutical SciencesMonash University
Research Article

DOI: 10.1007/s11096-012-9730-3

Cite this article as:
Elliott, R.A., O’Callaghan, C., Paul, E. et al. Int J Clin Pharm (2013) 35: 217. doi:10.1007/s11096-012-9730-3

Abstract

Background Hospitalisation often leads to increased medication regimen complexity for older patients; increased complexity is associated with medication non-adherence. There has been little research into strategies for reducing the impact of hospitalisation on medication regimen complexity. Objective To investigate the impact of pharmacist medication review, together with an educational intervention targeting clinical pharmacists and junior medical officers, on the increase in medication regimen complexity that occurs during hospitalisation. Setting Two acute general medicine wards and two subacute aged care (geriatric assessment and rehabilitation) wards at a major metropolitan public hospital in Melbourne, Australia. Methods A before-after study involving patients aged 60 years and over was undertaken over two 5-week periods. During the pre-intervention period patients received usual care. During the intervention period, clinical pharmacists were encouraged to review patients’ medication regimen complexity prior to discharge, and make recommendations to hospital medical officers to simplify regimens. Prior to the intervention period, pharmacists attended an interactive case-based education session about medication regimen simplification, and completed an assessment task. A similar, but briefer, education session was delivered to junior medical officers. Main outcome measure The primary endpoint was change in medication regimen complexity index (MRCI) score (a validated measure of regimen complexity) between admission and discharge for regularly scheduled long-term medications, adjusted for age, length of hospital stay, number of medications and regimen complexity prior to admission. Results Three hundred ninety-one patients were included (mean age 80.6 years, mean 7.4 regularly scheduled long-term medications on admission). The mean increase in MRCI score between admission and discharge was significantly smaller in the 205 intervention patients than in the 186 usual care patients (2.5 vs. 4.0, p = 0.02; adjusted difference 1.6, 95 %CI 0.3, 2.9). The intervention had greatest impact in patients discharged from subacute wards (mean adjusted difference: 2.7), not using a dose administration aid after discharge (mean adjusted difference: 2.6), and not discharged to a residential care facility (mean adjusted difference: 1.9). Mean differences in MRCI scores were equivalent to ceasing one to two medications. Conclusion An educational intervention and clinical pharmacist medication review reduced the impact of hospitalisation on the complexity of older patients’ medication regimens.

Keywords

AustraliaEducationGeriatricsHospital inpatientsMedication reviewPharmacistsPrescribingRegimen complexity

Impact of findings on practice

  • Regimen simplification should be a high priority in the management of older patients during their hospital stay.

  • Physicians and pharmacists should consider regimen complexity when prescribing medications and reviewing medication regimens for older hospital inpatients.

  • Educational interventions targeting medical officers and pharmacists focusing on regimen complexity may be valuable.

Introduction

Complexity of medication regimens is a major concern for patients with chronic illness and their caregivers [13], and increased regimen complexity is associated with medication non-adherence and medication errors [48].

Prescription of multiple medications is common in older patients, and may be unavoidable in those with multiple morbidities [9, 10]. Admission to hospital usually leads to changes to older patients’ medications including addition of medications [1113], often resulting in increased regimen complexity at discharge [11].

Although regimen complexity is related to the number of medications that a patient takes, there are a range of other factors that contribute to regimen complexity [5, 14]. This was illustrated by a recent study that assessed medication regimen complexity in almost 90,000 elderly home care patients in the United States of America [15], using the Medication Regimen Complexity Index (MRCI). The MRCI is a validated instrument that generates a score based on 65 items that take into account the dosage forms, dosing frequencies, and additional instructions relevant to drug administration (e.g. take 30 min before food) [14]. In that study, the median number of medications used by home care patients was seven, but within the group of patients taking seven medications the MRCI score ranged from 6.5 to 42 (mean 15) [15]. Higher MRCI scores were associated with increased risk of emergency department presentation and hospitalisation, even after controlling for number of medications and case-mix (demographics, co-morbidities, clinical and functional status) [15].

Medication regimens can often be simplified without altering the therapeutic intent of the regimen, for example by switching to longer-acting medications or formulations that require fewer dose-times per day, consolidating dose-times for multiple medications, using fixed-dose combination products, or switching to medications or dose-forms with less complex administration requirements [11]. Simplifying medication regimens can improve medication adherence and treatment outcomes [1619].

There has been little research into methods for reducing medication regimen complexity in older hospital inpatients [11, 20], who are at high risk of medication non-adherence, medication errors and therefore adverse medication events after discharge [21, 22].

One approach to simplifying medication regimens is to have a pharmacist review the regimen and make recommendations to the prescribing medical officer [11, 19]. This method may be more effective than expecting medical officers to simplify regimens independently because, when prescribing, they tend to focus primarily on therapeutics and may not consider regimen complexity and/or the patient’s ability to manage the regimen. Also, based on the authors’ experience, medical officers tend not to be as familiar as pharmacists with the range of products and formulations available that may help to simplify a patient’s regimen (e.g. fixed-dose combination products and longer-acting or simpler dose-forms). Clinical pharmacists in Australian hospitals usually review patients’ medication management and adherence and provide medication counseling as part of routine patient care [23], and therefore may be well placed to review regimen complexity.

Aim of the study

The aim of this study was to investigate the impact of pharmacist medication review, together with an educational intervention targeting inpatient clinical pharmacists and junior (intern and resident) medical officers (JMOs), on the increase in medication regimen complexity that occurs during hospitalisation. The hypothesis was that if pharmacists focused on reviewing regimen complexity when conducting routine in-hospital medication reviews, the increase in regimen complexity that typically results from hospitalisation would be reduced. Because it is often necessary to discharge patients on more medications than they were taking prior to admission (as a result of new diagnoses, often requiring multiple medications), the goal was to limit the increase in regimen complexity rather than reduce complexity.

Methods

Study design & setting

A before-after study was undertaken in two acute general medicine wards and two subacute aged care (geriatric assessment and rehabilitation) wards at a major metropolitan public hospital over two 5 week periods. The same weeks in two consecutive years were selected for the pre-intervention (usual care) and intervention study periods in order to minimise the impact of the JMOs’ and pharmacists’ level of experience and seasonal variations in case-mix and prescribing.

The participating wards received a clinical pharmacy service (Monday–Friday). Typically this included a pharmacist medication regimen review at admission (as part of admission medication reconciliation), during the inpatient stay (as part of regular medication chart reviews), and at discharge (as part of discharge medication reconciliation). Prior to the intervention, review of regimen complexity was not a major focus of the medication regimen review.

Subjects

The study involved patients aged 60 years and over who were discharged from the participating wards during the study periods. Patients were excluded if they were prescribed no regularly scheduled medications both prior to admission and at discharge, or if they were transferred to another hospital or ‘hospital-in-the-home’ (a program that provides care to hospital-admitted patients in their place of residence as a substitute for hospital accommodation).

Intervention

During the intervention period, as part of routine medication regimen reviews for patients under their care, pharmacists were encouraged to review regimen complexity and make recommendations to hospital medical officers to simplify medication regimens when it was clinically appropriate to do so (i.e. when the pharmacist felt that the patient would benefit from a simpler regimen). In the month prior to the intervention period, pharmacists attended an education session delivered by an experienced clinical pharmacist (RAE). The session included a 15 min presentation that addressed the impact of regimen complexity on medication adherence, the impact of hospitalisation on regimen complexity, and a case study illustrating how regimen complexity can be minimized without altering the therapeutic intent of the regimen. This was followed by a 60 min interactive tutorial in which participants, in small groups, worked through a series of cases to review the patients’ medications and minimise regimen complexity. The session was followed by an assessment task, in which pharmacists had to independently simplify a number of regimens; the pharmacist who ran the education session reviewed their simplified regimens and provided individualised feedback.

A similar, but briefer, education session (without an assessment component) was delivered to JMOs, to raise their awareness of the need to minimise regimen complexity and prime them for pharmacist recommendations related to regimen complexity. Senior medical staff (registrars and consultant physicians) were informed about the intervention but received no training.

Data collection

Pre-admission medication regimens and use of dose administration aids (DAAs) were obtained from the Medication History on Admission form within the patients’ hospital medical record. This form was routinely completed by clinical pharmacists, usually within 24–48 h of admission, as part of medication reconciliation [24]. It contained a comprehensive medication history obtained from the patient or their caregiver, and other sources such as the patient’s own medications, their community pharmacy and/or their general practitioner. Discharge regimens were obtained from the discharge prescription that was written by the treating medical unit and reviewed and reconciled by the unit’s clinical pharmacist (or weekend covering pharmacist) to ensure accuracy and inclusion of all medications the patient was required to take after discharge. Patients’ use of DAAs following discharge was obtained from the pharmacists’ discharge plan. All clinical pharmacists had received training and performance review for routine clinical pharmacy tasks including medication reconciliation on admission and discharge, and recording of clinical pharmacist activities and interventions.

Regimen complexity was assessed on admission and discharge using the MRCI [14]. Two MRCI scores were calculated for each time-point, one for regularly scheduled long-term medications only, and one for all medications (defined as regularly scheduled long-term medications plus regularly scheduled short-term [defined duration] medications and when required [prn] medications). MRCI scores were determined by a pharmacy student, who reviewed the medication regimens and entered the regimen characteristics into a purpose-designed spreadsheet which calculated the scores. The pharmacy student received training from the principal investigator (RAE), and patients’ MRCI scores were double-checked by the principal investigator until consistent agreement was achieved. A different pharmacy student was used to collect the data in the intervention period, but the same training and double-checking process was used to ensure consistency. The students were not blinded to study group allocation.

During the intervention period, pharmacists recorded whether they had reviewed regimen complexity and whether they had initiated any changes to simplify the regimen for each discharged patient. This data was not collected in the pre-intervention period. All other data were collected by the researchers.

Endpoints

The primary endpoint was the change in MRCI score between admission and discharge for regularly scheduled long-term medications.

Secondary endpoints were the change in MRCI score for all medications, the proportion of patients whose medication regimen complexity was reviewed by a pharmacist prior to discharge, and the proportion of patients who had one or more pharmacist-initiated regimen simplifications implemented.

Sample size calculation

We calculated that 193 patients would be needed in the pre-intervention and intervention groups to detect a 2.0 point reduction in the hospitalisation-induced increase in MRCI score, from 4.0 (SD 7.0) to 2.0 (SD 7.0), with 80 % power, and an alpha value of 0.05 (two-sided). A difference of 2.0 points was selected for two reasons: (1) in terms of complexity a 2.0 point reduction would be equivalent to ceasing at least one medication (e.g. a twice-daily medication, or a once-daily medication with special administration instructions), which we considered would represent a clinically relevant average reduction in regimen complexity, and (2) a pilot study suggested that an average reduction in MRCI score of 2.96 (range 0–10.5) would be the maximum achievable if all potential regimen simplifications were implemented [11].

Statistical analysis

Statistical analysis was performed using SPSS Version 19.0 (IBM SPSS Statistics, USA). The primary outcome measure (change in MRCI score) was assessed for normality and found to be well approximated by a normal distribution. Comparisons between the pre-intervention and intervention groups were performed using the independent samples Student T test for parametric continuous variables, Mann–Whitney U test for nonparametric continuous variables and Chi square test for categorical variables. Multiple linear regression analysis was conducted to assess the effect of the intervention on change in MRCI scores adjusting for age, length of hospital stay, number of medications and regimen complexity prior to admission, with results reported as parameter estimates (95 % confidence intervals). Statistical significance was set at a two-sided p value of 0.05.

Subgroup analysis was performed for four specific subgroups that were identified a priori: ward type (acute vs. subacute), discharge destination (home vs residential care), medication management after discharge (independent vs assisted) and use of a DAA after discharge (yes vs. no).

Continuous data are presented as mean (standard deviation) or median (inter-quartile range) depending on the underlying distribution of the data. Categorical data are reported as number (proportion).

The study was approved by the Monash University and Austin Health Human Research Ethics Committees.

Results

Three hundred and ninety-one patients were included in the study (Fig. 1); 186 in the pre-intervention (usual care) group and 205 in the intervention group (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11096-012-9730-3/MediaObjects/11096_2012_9730_Fig1_HTML.gif
Fig. 1

Patient flow chart. a Provision of care to hospital-admitted patients in their place of residence as a substitute for hospital accommodation

Table 1

Characteristics of patients included in the study

 

Pre-intervention patients (n = 186)

Intervention patients (n = 205)

p value

Age in years [mean (SD)]

79.7 (8.2)

81.3 (8.0)

0.05

Female [n (%)]

107 (57.5)

119 (58.0)

0.92

Length of stay in hospital, days [Median (IQR)]

9.5 (4.0–27.0)

11.0 (4.0–27.5)

0.81

Warda [n (%)]

 Acute

115 (61.8)

128 (62.4)

0.90

 Subacute

71 (38.2)

77 (37.6)

 

No. of medications prior to admission [Mean (SD)]

 Regularly scheduled long-term medications

7.1 (4.0)

7.7 (3.6)

0.11

 All medications

8.1 (4.4)

8.8 (4.1)

0.08

MRCI score prior to admission [Mean (SD)]

 Regularly scheduled long-term medications

18.2 (11.2)

19.1 (10.3)

0.45

 All medications

20.7 (12.5)

21.7 (11.6)

0.42

Living at a residential care facility prior to admission [n (%)]

37 (19.9)

40 (19.5)

0.93

Managing medications independently prior to admission [n (%)]

112 (60.2)

119 (58.0)

0.66

Using a DAA prior to admission [n (%)]

70 (37.6)

79 (38.5)

0.85

aWard from which the patient was discharged (Note: most subacute patients spent time on an acute ward before being transferred to the subacute aged care ward)

DAA Dose administration aid (also known as ‘multi-compartment adherence aid’ or ‘monitored dosage system’), MRCI Medication Regimen Complexity Index, SD Standard deviation, IQR Inter-quartile range

Hospitalisation resulted in an increase in medication regimen complexity in both groups, however the mean increase in regimen complexity between admission and discharge was significantly smaller in the intervention group compared to the pre-intervention group: 2.5 versus 4.0, adjusted difference 1.6 (95 % CI 0.3, 2.9) for regularly scheduled long-term medications; 4.7 versus 6.7, adjusted difference 2.0 (95 % CI 0.3, 3.6) for all medications (Table 2).
Table 2

Impact of the intervention on change in MRCI score between admission and discharge

 

Change in MRCI score between admission & dischargea [Mean (SD)]

Unadjusted Difference [Mean (95 %CI)]

p value (unadjusted difference)

Adjusted differenceb [Mean (95 %CI)]

p value (adjusted difference)

Pre-intervention patients (n = 186)

Intervention patients (n = 205)

All patients

 Regularly scheduled long-term medications

+4.0 (7.5)

+2.5 (6.5)

−1.5 (−2.9, −0.1)

0.03

−1.6 (−2.9, −0.3)

0.02

 All medications

+6.7 (9.5)

+4.7 (8.3)

−2.0 (−3.8, −0.2)

0.03

−2.0 (−3.6, −0.3)

0.02

Subgroup analysis

 Data reported for regularly scheduled long-term medications onlyc

 Ward type

  Acute (n = 243)

+1.8 (6.2)

+1.4 (5.7)

−0.4 (−1.9, +1.1)

0.6

−0.6 (−2.1, +0.9)

0.45

  Subacute (n = 148)

+7.6 (8.1)

+4.3 (7.2)

−3.3 (−5.8, −0.8)

0.01

−2.7 (−5.1, −0.4)

0.02

 Discharge destination

  Residential care (n = 106)

+2.8 (6.3)

+2.9 (7.3)

+0.1 (−2.6, +2.7)

0.95

−0.7 (−3.2, +1.8)

0.32

  Home (n = 285)

+4.4 (7.9)

+2.3 (6.1)

−2.1 (−3.8, −0.5)

0.01

−1.9 (−3.4, −0.4)

0.02

 Medication management after discharge

  Assisted (n = 189)

+4.1 (6.9)

+2.8 (6.3)

−1.3 (−3.2, +0.6)

0.17

−1.4 (−3.2, +0.3)

0.11

  Independent (n = 202)

+3.9 (8.0)

+2.2 (6.6)

−1.7 (−3.8, +0.3)

0.09

−1.6 (−3.5, +0.2)

0.09

 Use of a DAA after discharge

  DAA (n = 170)

+2.3 (7.4)

+2.0 (6.9)

−0.3 (−2.4, +1.9)

0.84

−0.3 (−2.4, +1.7)

0.72

  No DAA (n = 221)

+5.3 (7.4)

+2.8 (6.1)

−2.5 (−4.3, −0.7)

0.007

−2.6 (−4.2, −0.9)

0.002

DAA Dose administration aid (also known as ‘multi-compartment adherence aid’ or ‘monitored dosage system’), MRCI Medication Regimen Complexity Index, SD Standard deviation, CI Confidence interval

a+ and − symbols signify the direction of change/difference in MRCI score

bAdjusted for age, length of hospital stay, number of medications and regimen complexity prior to admission

cNb. Analysis with ‘all medications’ produced similar results (slightly larger reductions in MRCI-change, with similar p values)

Subgroup analyses indicated that the intervention had a greater impact in patients who were discharged from subacute wards, not using a DAA after discharge, and not discharged to a residential care facility (Table 2).

During the intervention period pharmacists reported reviewing regimen complexity for 173/205 (84.4 %) patients, and implementing 94 simplification-related regimen changes in 54/205 (26.3 %) patients. The mean increase in regimen complexity between admission and discharge for regularly scheduled long-term medications was similar for patients who did (n = 54) and did not (n = 151) have pharmacist-initiated changes implemented (3.0 vs. 2.3, p = 0.46).

Discussion

Intensity and complexity of drug therapy have increased over recent years as a result of increasing evidence for multi-drug regimens for the management of many of the chronic medical problems that affect older people, such as airways disease, hypertension, heart failure, diabetes and osteoporosis [9, 25, 26]. Addition of medications to patients’ regimens, to optimise their medical management, during hospitalisation is often unavoidable.

Our study suggests that delivery of an educational intervention to hospital pharmacists and JMOs, and encouraging pharmacists to focus on regimen complexity when conducting routine in-hospital medication reviews, can help to minimise the impact of hospitalisation on the complexity of discharge medication regimens.

In the subgroups of patients that benefited most from the intervention, the average reductions in the hospital-related increase in MRCI score (1.9–2.7) were equivalent to prescribing one to two fewer medications, which is likely to be a clinically relevant reduction. For some individual patients, the reduction is likely to have been much larger [11], as the average change in MRCI score is influenced by the fact that not all patients had regimen simplification changes made. In an earlier pilot study exploring regimen simplification, mean MRCI score was reduced by 2.96, but 20 % of patients had reductions in MRCI scores of 6 or more [11].

The intervention had no significant effect on medication regimen complexity for patients who were discharged to residential care or using a DAA after discharge. This may be because the increase in regimen complexity was not as great in these patients (Table 2), providing less opportunity for simplification. It is also possible that pharmacists and/or medical officers considered regimen simplification to be a lower priority in patients going to residential care (since these patients would usually not be managing their own medications after discharge) or using a DAA (that may be packed by a third party such as a community pharmacist).

The intervention was most effective in patients discharged from subacute aged care wards. This may be because these patients experience more medication changes during their hospital stay and are in hospital for a longer period [11], providing more opportunities to simplify the regimen. It could also be because the care provided on these wards focuses on improving frail, elderly patients’ capacity to function independently after discharge, and medication regimen simplification is one strategy for assisting patients’ (and/or their caregivers’) to manage their medications. Barriers to the simplification of medication regimens were reported by pharmacists, most commonly ‘lack of time’ and non-acceptance of recommendations by patients or prescribers; these are described in more detail in another paper [27].

The intervention had a similar impact on the change in regimen complexity for patients who did and did not have pharmacist-initiated medication changes implemented, suggesting that medical officers might have increased their focus on regimen complexity when prescribing, even when a pharmacist did not make a specific recommendation. This might have been a result of the JMOs’ education session, however it is likely that having pharmacists make simplification-related recommendations for more than one in four of their patients also acted as an ongoing reminder to consider regimen complexity. This is consistent with an earlier study that reported reduced medication regimen complexity when JMOs were regularly prompted to review their patients’ regimens [20].

Our study has both strengths and limitations. Regimen complexity was measured using an objective, reliable and validated method. The two study groups were well matched in terms of ward, gender, place of residence, medication management and use of DAAs; non-significant, but potentially clinically relevant, differences in length of stay and pre-admission medication regimens were adjusted for in the analysis. A limitation of the non-randomised before-after design is potential for unmeasured confounding factors. We did not collect data on reasons for admission and co-morbidities, but timing the two study periods to occur over the same months of two consecutive years is likely to have avoided seasonal variations in case-mix. This feature of the study design also meant that prescribers in the two study periods had a similar level of experience (at least 10 months for JMOs). We are not aware of any other interventions or changes in practice that occurred between the two study periods that could have influenced the complexity of medication regimens. The individuals who calculated the MRCI scores were not blinded to group allocation. However, MRCI is an objective measure and one of the authors trained both scorers and independently checked their scores to ensure consistency. The scorers did not know each others’ MRCI scores and risk of observational bias influencing the outcome was minimal. The intervention comprised multiple elements (education of pharmacists and doctors and pharmacist medication review) so it is not possible to ascertain the effectiveness of individual components. We only measured the impact of the intervention over 5 weeks; since the impact of educational interventions may decline over time, it is possible that the education may need to be repeated at regular intervals, or integrated into routine training programs for pharmacists and JMOs, to have a sustained effect. Data collection (recording of interventions) by the clinical pharmacists during the intervention period may have enhanced the impact of the educational intervention, by acting as a reminder for them to review regimen complexity (this data was not collected in the pre-intervention period). Data on the number of patients reviewed by pharmacists and the number of pharmacist-initiated changes were reliant on self-reporting and may be an under- or over-estimate.

This is only the second study, to our knowledge, to test an intervention designed to address medication regimen complexity in hospital inpatients. Muir et al. [20] provided JMOs with a ‘medication grid’ displaying all of the patient’s medications and times of administration and asked them to review the regimen with their team (including a clinical pharmacist, if available). Patients whose JMO received the grid were discharged on fewer medications and fewer doses per day [20].

Neither our study nor the study by Muir et al. [20] assessed the impact of regimen simplification on medication adherence, clinical outcomes or patient/caregiver burden post-discharge. Although there is evidence that simpler medication regimens lead to better adherence to treatment and better treatment outcomes, [1618] longitudinal studies are needed to determine whether strategies to minimise regimen complexity in hospitalised patients lead to improved adherence and outcomes after discharge.

Conclusion

An educational intervention and clinical pharmacist medication review reduced the impact of hospitalisation on the complexity of older patients’ medication regimens. Given the large body of evidence that regimen complexity is an important predictor of treatment burden, medication adherence and outcomes [18, 1618], more attention should be devoted to this aspect of prescribing.

Acknowledgments

The assistance of Patricia Ooi and Victoria Tran (pharmacy students, Monash University) and Dhineli Perera (clinical pharmacist, Austin Health) with the conduct of this study is acknowledged.

Funding

The study received no external funding or sponsorship.

Conflicts of interest

The authors have no competing interests to declare.

Copyright information

© Springer Science+Business Media Dordrecht 2012