Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Beers MH et al. To develop explicit criteria Literature review of 85 published sources The experts concurred on 30 criteria A total of 30 factors defining
12 for identifying for survey development and two-round statements. Nineteen involved inappropriate use of
“Explicit criteria for inappropriate medication survey, based on Delphi methods. medications that should generally be medications in nursing home
determining use in nursing home avoided and 11 described doses, elderly was agreed upon by
inappropriate populations. Sample: 13 nationally and internationally frequencies or durations of medications experts.
medication use in recognized experts in that should generally not be exceeded.
nursing home residents” psychopharmacology, The criteria are intended for
pharmacoepidemiology, clinical geriatric Consensus could not be reached on use in typical clinical
Archives of Internal pharmacology, general clinical geriatrics, some aspects of use of eight situations in nursing home
Medicine and long-term care. medications or categories of settings. The authors noted
medications, including antipsychotics, that the guidelines might not
1991;151(9):1825-32. Instrument: Participants responded to a diphenhydramine, peptic ulcer disease apply under unusual clinical
48-statement survey, based on published treatments, diuretics, combination drugs, circumstances.
statements on inappropriate medication disopyramide, egoloid mesylates, and
use in frail, elderly persons. Criteria digoxin. The criteria are also limited
statements addressed 30 of the most by the fact that they exclude
commonly prescribed medications in clinical data in addition to
nursing homes. medication use. Thus, the
guidelines do not take into
Two types of statements: 1) individual account the entire medication
medications or classes of medications that regimen and should be
should be avoided except for special applied with caution.
circumstances; and 2) doses, frequencies,
and durations of medications that Studies of the validity of the
generally should not be exceeded. guidelines in measuring drug
Experts indicated their agreement or therapy appropriateness in
disagreement to the first type of nursing home residents are
statements and set dose, frequency and needed.
duration of therapy limits on the second
type. All responses were based on Also, studies on the impact of
medication use in elderly persons (>65 these guidelines on patient
years) who reside in nursing homes. health outcomes are
warranted.
Targeted drugs or drug categories:
sedative/hypnotics, antidepressants,
antipsychotics, antihypertensives,
nonsteroidal anti-inflammatory drugs, oral
hypoglycemics, analgesics, dementia
treatments, platelet inhibitors, histamine2
blockers, antibiotics, decongestants, iron
supplements, muscle relaxants,
gastrointestinal antispasmotics, and
antiemetics.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Gupta S et al. To examine the Retrospective, observational, cross- A total of 51.8% was not prescribed any Inappropriate medication use,
14 relationship between sectional review of Medicaid Claim History of the inappropriate medications. as measured by a modified
“Inappropriate drug inappropriate medications, and Drug files. Beers’ criteria, was related to
prescribing and related health care provider The cost of pharmaceutical services higher cost of pharmaceutical
outcomes for elderly factors, and patient Setting: Louisiana’s Intermediate Care ranged from $0 to $26,827. services but not related to
Medicaid beneficiaries demographics to cost of Facilities I. mortality.
residing in nursing pharmaceuticals and Only 4.7% of beneficiaries died in 1994.
homes” mortality. Sample: N=19,932 Medicaid beneficiaries, Findings from this study are
73.6% female, and 60.1% was 81 years or The number of inappropriate medications limited by: 1) the selective
Clinical Therapeutics older. was positively correlated with the cost of characteristics of the sample;
pharmaceuticals. 2) use of a secondary
1996;18(1):183-96. Instrument: Modified Beers’ criteria were database; and 3) lack of
applied to medications prescribed to The number of inappropriate medications consideration of clinical or
beneficiaries. was not significantly associated with other medication (e.g., dose,
mortality. duration) data.
Primary outcomes: Cost of
pharmaceutical services and mortality in Studies of the validity of the
1994. Beers’ criteria in optimizing
medication use in nursing
home residents are needed.
Also, studies on the impact of
Beers’ criteria on patient
health outcomes are
warranted.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Beers MH et al. To quantify the A prospective, cohort study. Residents were prescribed an average of Findings showed that the
15 appropriateness of 7.2 medications. quality of prescribing in NHs
“Inappropriate medication prescriptions Setting: 12 nursing homes (NHs) in Los may be poor.
medication prescribing in nursing home residents. Angeles. Based on Beers’ criteria, 40.3% (N=446)
in skilled-nursing of residents were prescribed at least one The findings from this study
facilities” Sample: N=1106 nursing home residents, inappropriate medication; 10.4% (N=115) are limited by a lack of
80% female, Mean age=84 years. Note: had > 2 inappropriate medications; and consideration for clinical data
Annals of Internal No standard deviations reported. 1.2% (N=14) had > 3 inappropriate by the Beers’ criteria.
Medicine medications.
Instrument: 30 explicit criteria defining Although the content validity
1992;117(8):684-89. inappropriate medication use in NHs, also A total of 50.8% of all inappropriate of Beers’ criteria has been
known as Beers’ Criteria. prescriptions involved drugs that addressed, studies of its
generally should not be prescribed to NH specificity, sensitivity, and
Primary outcomes: Number of residents at all, according to Beers’ reliability are needed.
inappropriate medication prescriptions. criteria.
Studies of interventions
About one-third (34.4%) involved based on Beers’ criteria and
prescriptions of drugs for durations that their impact on patient health
exceeded recommended limits and outcomes in NHs are
14.8% involved doses that exceed warranted.
recommended limits.
The two most common types of
inappropriate prescriptions included high
doses of iron supplements and orders for
long-acting benzodiazepines, persantine,
and propoxyphene.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Kelly MW et al. To determine the Retrospective chart review. Residents were prescribed an average of Potentially inappropriate
16 appropriateness of 17.9 ± 5.1 medications per resident. medication use was found in
“Appropriateness of medication use in Setting: 16 long-term care facilities in Iowa many elderly residing in long-
medication use in Iowa institutionalized Medicaid (N=13 nursing facilities and N=3 A total of 729 medications was given on term care facilities and on 15
Medicaid recipients recipients receiving 15 or psychiatric facilities). a scheduled basis. or more medications.
residing in long-term more medications.
care facilities and Sample: N=60 Medicaid recipients Based on Beers’ criteria: Findings from this study are
receiving 15 or more receiving 15 or more medications, 82% 13.4% (98 of 729) of medications were limited by the selective
medications” female, average age=73 ± 15 years. deemed potentially inappropriate and 42 characteristics of the sample.
of 60 received at least one potentially
Consultant Pharmacist Instruments: Beers’ criteria, medication inappropriate medication. The MAI was not developed
appropriateness index (MAI), and U.S. for the nursing home setting.
2000;16(12):1184-91. Health Care Financing Administration’s Propoxyphene, diphenhydramine and
(HCFA) Indicators for surveyor amitriptyline accounted for 66.1% of Studies of the validity of the
assessment of the performance of drug inappropriate prescriptions involving assessment tools in
regimen reviews were applied to drugs that generally should not be optimizing medication use in
medications taken by participants. prescribed. nursing home residents are
needed.
Primary outcomes: Number of potentially Based on MAI:
inappropriate medications based on 26.5% (193 of 729) of medications were Also, studies of the impact of
above criteria. deemed potentially inappropriate and 56 interventions, based on these
of 60 received at least one inappropriate tools, on patient health
medication. outcomes are warranted.
Drug cost, duration and indication
represented 67.5% of the
inappropriateness ratings.
The most common type of inappropriate
use was use of H2 blockers with no
indication or for conditions they have not
been proven effective.
Based on HCFA indicators:
12.8% (93 of 729) instances of
irregularities were identified.
42 of 60 recipients had at least one
potentially inappropriate medication.
Most common problems: continuous use
of hypnotics and concurrent use of ≥ 2
antipsychotics.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Williams B, Betley C To determine the Retrospective, cross-sectional. Residents were prescribed an average of Findings showed that
17 prevalence and patient 5.2 ± 3.1 standing (non-prn) medications. presumptively inappropriate
“Inappropriate use of characteristics of Setting: 252 nursing homes (NHs) of two medication use is high in
nonpsychotropic presumptively states in the U.S. A total of 11.9% of residents were NHs.
medications in nursing inappropriate prescribed at least one inappropriate
homes” nonpsychotropic Sample: N=21,884 nursing home standing medication. Targeting residents with high
medications of nursing residents, 73% female, Mean age=82.4 ± numbers of medications may
Journal of the American home residents of two 11.1 years. The most common types of inappropriate be an efficient way to identify
Geriatrics Society states. medications were dipyridamole, residents at high risk.
Instrument: Modified Beers’ criteria, using amitriptyline, and methyldopa.
1995;43(5):513-19. only 10 criteria that excluded psychotropic The findings from this study
medications (except for amitriptyline) and The most powerful predictors of are limited by a lack of
medications likely to be used as needed inappropriate medication use were consideration for clinical data
(prn) (e.g., spasmodics). number of standing medications and by the Beers’ criteria.
state.
Primary outcomes: Number of Although the content validity
presumptively inappropriate medications of Beers’ criteria has been
used to treat medical conditions. addressed, studies of its
specificity, sensitivity, and
reliability are needed.
Studies of interventions
based on Beers’ criteria and
their impact on patient health
outcomes in NHs are
warranted.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Sgadari A et al. To summarize the Descriptive research that examined inter- A large percentage (range: 70 to 96%) of Findings support the overall
18 evidence for the inter-rater rater reliability, using dual assessments the RAI/MDS items achieved adequate reliability of a majority of the
“Efforts to establish the reliability of RAI/MDS completed by licensed nursing personnel. (ĸ =0.4-0.75) to excellent (ĸ =0.75 or RAI/MDS items. Items
reliability of the resident items in the USA and in higher) reliability across all countries exhibiting poor reliabilities
assessment instrument” non-English speaking Setting: Denmark, Iceland, Italy, Japan, examined. need to be improved and
countries. Sweden, Switzerland and the USA. tested.
Age & Ageing The patterns of reliabilities moderately
Sample: Randomly selected residents differed among countries, whereby Studies of the validity of the
1997;26(S2):27-30. ranging from N=24 (Iceland) to N=129 Sweden and Japan had lower reliabilities RAI/MDS in measuring drug
(Japan). Number of facilities ranged from and Switzerland exhibited the highest appropriateness are needed.
2 (Italy) to 19 (Switzerland). level of inter-rater reliabilities.
Instrument: The RAI consists of: 1) a core
set of assessment items [the minimum
data set (MDS)], that provides a
comprehensive assessment of each
resident’s functional status; 2) 18
condition-focused resident assessment
protocols (RAPs) which provide additional
and more specialized assessments; and
3) a user’s manual that details how to use
the MDS and RAPS. The MDS covers 15
major domains: background and
customary routines,
communication/hearing, physical
functioning and structural problems, mood
and behavior patterns, disease diagnoses,
oral/nutritional status, skin condition,
special treatments and procedures,
cognitive patterns, vision patterns,
continence, activity pursuit patterns,
health conditions, oral/dental status, and
medication use. The RAPS consist of 18
areas: delirium, visual function, ADL
functional/rehabilitative potential,
psychosocial well-being, behavior
problem, falls, feeding tubes, dental care,
psychotropic drug use, cognitive
loss/dementia, communication, urinary
incontinence and indwelling catheter,
mood state, activities, nutritional status,
dehydration/fluid maintenance, pressure
ulcers, and physical restraints.
Primary outcomes: Inter-rater reliability
estimates (ĸ values).
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Morris JN et al. To assess the reliability of Descriptive research that examined inter- Almost all of the new items (e.g., Findings support the overall
19 new assessment items rater reliability, employing dual measures of pain, foot care, discharge reliability of the new and
“A commitment to and their perceived clinical assessments of residents by staff nurses status) achieved adequate (ĸ =0.4-0.75) revised items and the clinical
change: Revision of utility (as judged by using Version 2.0 of the minimum data set interrater reliability levels. usefulness of the new items
HCFA’s RAI” nurses) of Version 2 of the (MDS). of the MDS Version 2.0.
Resident Assessment Reliabilities for revised items (e.g., items However, more studies are
Journal of the American Instrument (RAI). Setting: 21 nursing homes in seven for mood, delirium, and change in status needed to improve the
Geriatrics Society states. indicators over the previous 90 days) reliability of poorly performing
were significantly (no p-value reported) items.
1997;45(8):1011-16. Sample: N=187 randomly selected higher than the original version. For
residents. Mean age=80.6 years (no example, the interrater reliability of the Studies of the validity of the
standard deviations reported). indicators of depression, anxiety, or sad RAI/MDS Version 2.0 in
mood items increased from 0.44 in the measuring drug
Instrument: MDS assessment items original MDS to 0.68 in the MDS Version appropriateness are needed.
covering the following areas: background 2.0.
information, cognitive patterns,
communication/hearing, vision, mood and Most (60.6%) nurse assessors made
behavior, psychosocial well-being, positive statements regarding the overall
physical functioning and structural changes to the MDS.
problems, continence, disease diagnoses,
health condition, oral/nutritional status,
dental status, skin condition, activity
pursuit patterns, medications, special
treatments and procedures, discharge
potential and overall status.
Primary outcomes: Interrater reliability
estimates (using kappa ĸ values) and staff
appraisal of the MDS 2.0.
Table 1: Detailed Description of Medication Use Assessment Tools Studies
# Article Objectives Study Design Main Findings Discussion
Ackermann RJ, von To develop Descriptive literature review. 12 general guidelines for effective drug Many drugs and drug
24 Bremen GBM recommendations for prescribing for nursing home patients regimens can be discontinued
reducing polypharmacy in Sample: number of reviewed articles not or simplified in nursing home
“Reducing nursing home patients. reported. 1. Start slow, go slowly. residents.
polypharmacy in the 2. Do not prescribe drugs for every
nursing home: An Inclusion criteria: Not reported. minor symptom or disease. All drugs should have clear
activist approach” 3. Try not to prescribe a drug to treat indications.
Targeted drugs or drug classes: 1) side effects.
Journal of the American antibiotics, 2) antiarrhythmic drugs, 4. Set priorities in therapy. Drugs used solely for
Board of Family diuretics and potassium, 3) digoxin, 4) 5. Select a drug that can treat more preventive purposes (e.g.,
Practice antihypertensives, 5) angina treatments, than one problem. lipid-lowering drugs or
6) histamine2-blockers, 7) theophylline, 8) 6. Review drug side effects, anticoagulants) may not have
1995;8(3):195-205. lipid-lowering drugs, 9) pentoxifylline, 10) interactions, and contraindications. a positive effect on patients
anticonvulsants, 11) anticoagulants, 12) 7. Monitor for compliance. with limited life expectancy.
nonsteroidal anti-inflammatory drugs, and 8. Periodically review medication list.
13) allopurinol. 9. Remember that drugs can cause Studies of the validity of the
illness. guidelines in reducing
10. Be wary of using new drugs on sick polypharmacy in nursing
and frail elders. home residents are needed.
11. Collaborate with facility pharmacist.
12. Write time-limited prescriptions. Also, studies on the impact of
these guidelines on patient
Individualized approaches to reducing health outcomes are
targeted drugs or drug classes were also warranted.
reported.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Midlöv P et al. To describe drug use and A cluster randomized, controlled trial. A total of 32 epilepsy (43%) and 31 The multi-specialty team
26 evaluate the effect of a Parkinson’s (37%) patients used at least intervention had no significant
“Descriptive study and multi-specialty team Setting: 48 nursing homes (NHs) in one drug deemed inappropriate using effect on HRQOL, ADL, and
pharmacotherapeutic intervention on health- Sweden (32 intervention and 16 control Beer’s criteria for geriatric NH patients. confusion states, except for
intervention in patients related quality of life homes). one small decrease in ADL
with epilepsy or (HRQOL), activities of For the epilepsy patients in the among Parkinson’s patients.
Parkinson’s disease at daily living (ADL), and Sample: N=157 residents (73 had intervention group, a total of 99 drug The clinical significance of
nursing homes in confusion state in epilepsy epilepsy, 83 had Parkinson’s disease and therapy changes were recommended this decrease is uncertain.
Sweden” and Parkinson’s patients one had both) (Control group: N=92, and 44 were carried out. The most
in nursing homes. mean age=79.8 in epilepsy and 82.2 for common recommendation was The authors offered several
European Journal of Parkinson’s; Intervention group: N=66, withdrawal/decreased dose of anti- possible reasons why the
Clinical Pharmacology mean age=78.6 in epilepsy and 80.7 for epileptic drug or of drugs that lower intervention yielded no
Parkinson’s). Note: No standard seizure threshold. improvements, including poor
2002;57(12):903-10. deviations were reported. methodology, poor
For the Parkinson’s patients, 109 measurement, or simply a
Intervention: A data form (consisting of changes were recommended and 33 lack of effect of the
patient-specific background information, were carried out. The most common intervention (note that only
medication, drug-related problems and recommendations included increased 35% of therapy
suggestions for changes in drug therapy) dose of anti-parkinsonian drugs, recommendations were
sent to responsible physicians. withdrawal of anti-parkinsonian drugs carried out).
Inappropriate medication use was and withdrawal/decreased dose of
evaluated using Beer’s criteria. Therapy psychotropics.
recommendations were made by a multi-
specialty team consisting of pharmacists For epilepsy patients, there were no
and physicians, some with expertise in significant differences between the
neurology. intervention and control groups in
changes in HRQOL, ADL, and confusion
Primary outcomes: HRQOL, ADL, and states.
confusion states at baseline and five to six
months post-intervention. For Parkinson’s patients, there were no
significant differences between the
intervention and control groups in
changes in HRQOL and confusion
states, but a significant (p=0.02)
decrease in ADL in the intervention
group. The mean decrease in ADL was
4.40 (CI= 1.32, 7.49) in the intervention
group and 0.63 (CI= -2.88, 4.14) in the
control group.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Gill SS et al. To determine the A descriptive study. A total of 69 potentially inappropriate Follow-up pharmacist letters
27 prevalence and predictors prescriptions were identified in 65 of the were fairly effective in
“Improving prescribing of potentially inappropriate Setting: A rehabilitation hospital/long-term 355 patients. reducing potentially
in the elderly: A study in prescribing and to care facility in London, Ontario. inappropriate prescriptions.
the long term care examine the effectiveness Twenty-five (37.9%) of the 66 remaining
setting” of follow-up pharmacist Sample: N=355 patients who were 65 potentially inappropriate prescriptions The findings are limited by
letters to physicians in years or older. (three patients died) had either been the lack of
Canadian Journal of improving prescribing. discontinued or changed according to comprehensiveness of the
Clinical Pharmacology Intervention: Follow-up letters were sent recommendations. IPET instrument.
to prescribers who had patients with
2001;8(2):78-83. potentially inappropriate medication use Compared to those without potentially The impact of the modified
as identified by the Improving Prescribing inappropriate prescriptions, patients with prescriptions on patient
in the Elderly Tool (IPET). The follow-up potentially inappropriate prescriptions: a) health outcomes could not be
letter contained concerns about potentially were taking significantly more (p<0.001) determined.
inappropriate medication use and medications (6.59 vs. 8.49, respectively);
suggestions for safer alternatives. b) had a history of mental illness
Potentially inappropriate use was verified (p=0.002); had higher MDS scores for
by a panel that consisted of a family depression (p=0.002); and c) were less
physician, a geriatric medicine specialist likely to have seen a geriatric medicine
and a geriatric pharmacist. specialist (p<0.001).
Primary outcomes: Number of potentially The five most common types of
inappropriate prescriptions two months potentially inappropriate prescribing
post-intervention. included: 1) anticholinergic drugs to
manage antipsychotic medication side
effects; 2) tricyclics with active
metabolites; 3) long-acting
benzodiazepines; 4) long-term NSAID
use for osteoarthritis; and 5) NSAID use
in patients with a history of peptic ulcer
disease, hypertension or renal
insufficiency.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Schmidt IK, Fastbom J To assess drug use in A randomized, controlled trial. Comparison of 1995 and 1998 Drug use is still extensive and
28 Swedish nursing homes Intervention NHs problematic in Swedish NHs.
“Quality of drug use in and changes in the quality Setting: 36 nursing homes (NHs) in Number of prescribed medications
Swedish nursing of drug use over a 3-year Sweden (18 intervention and 18 control significantly (p<0.05) increased from 8.0 The team training
homes: A follow-up period after a team homes). to 8.8 per resident. intervention, employing
study” training intervention. regular drug use review, had
Sample: N=1549 residents; Mean age=84 15 major drug classes: significant lasting effects on the quality
Clinical years. Note: no standard deviations (p<0.05) increases in anticoagulants, of drug use in NHs, as
Pharmacoepidemiology reported. minor analgesics, oxazepam (short- evidenced by the positive
acting), and antidepressants and findings three years post-
2000;20(6):433-46. Intervention: A multidisciplinary team of decreases in antipsychotics, diazepam intervention.
prescribing physicians, a pharmacist and (long-acting), and benzodiazepines.
nurses met regularly over a 1-year period The impact of the change in
to discuss individual patient drug use. Significantly (p<0.05) fewer residents drug use on patient health
Measurements of quality was based on were prescribed antipsychotics without a outcomes could not be
three areas: 1) drugs that cause confusion diagnosis (28.4% vs. 21.9%), less determined.
and memory impairments as classified by appropriate hypnotics (19.0% vs. 14.0%),
the Swedish Medical Products Agency less appropriate anxiolytics (21.4% vs.
(MPA) guidelines; 2) polypharmacy 18.0%), and ≥ 2 drugs from the same
regarding psychotropic drugs; and 3) therapeutic group (23.0% vs. 15.9%).
clinically-significant interactions based on
the pharmaceutical specialties in Sweden Drug interactions significantly (p<0.05)
(FASS) classification system. increased from 22.5% to 30.0%.
Primary outcomes: Number of prescribed Comparison of 1998 Intervention and
medication per resident, proportion of Control NHs
users of the 15 most commonly Significantly (p<0.05) fewer residents
prescribed drug classes, and quality used laxatives (65% vs. 76%), narcotics
indicators (use of drugs that cause (29% vs. 37%), minor analgesics (49%
confusion or memory impairments, use of vs. 62%), hypnotics (36% vs. 42%), and
psychotropics, and clinically-significant loop diuretics (38% vs. 47%).
drug interactions) at baseline and three
years post-intervention. Significantly (p<0.05) fewer residents
were prescribed tricyclics (3.1% vs.
6.7%), less appropriate hypnotics (10.1%
vs. 17.5%), and
≥ 3 medications that may cause
confusion (27% vs. 36%).
Drug interactions were not significantly
different.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Schmidt IK et al. To evaluate the impact of A randomized, controlled trial. At baseline, 77% (N=1480) had one or Overall, the use of
29 regular multidisciplinary more psychotropic prescriptions. psychotropic drugs was high
“The impact of regular team interventions on the Setting: 33 nursing homes (NHs) in in Swedish NHs.
multidisciplinary team quality and quantity of Sweden (15 intervention and 18 control The most frequently prescribed were
interventions on psychotropic drug homes). hypnotics (40%), anxiolytics (40%), and The multidisciplinary team
psychotropic prescribing prescribing in Swedish antipsychotics (38%). interventions effectively
in Swedish nursing nursing homes. Sample: N=1854 residents (N=626 reduced prescribing of
homes” Intervention, mean age=83, 70% female; The proportion of residents receiving antipsychotics and
N=1228 Control, mean age=84, 67% antipsychotics decreased by 19% nonrecommended hypnotics.
Journal of the American female). Note: No standard deviations (p=0.07), with a nonsignificant 7%
Geriatrics Society reported. decrease in control NHs (p=0.18). The impact of the intervention
on patient health outcomes
1998;46(1):77-82. Intervention: Monthly review of drug use Use of nonrecommended hypnotics could not be determined.
by a multidisciplinary team consisting of a declined by 37% (p<0.001), and
physician, a pharmacist and nurses over a acceptable hypnotics increased from 9 to
1-year period. Drugs were classified as 15% (p<0.001).
inappropriate based on the Swedish
Medical Products Agency (MPA) Use of acceptable antidepressants (from
guidelines. tricyclics to SSRIs) significantly (p<0.01)
increased in both groups.
Primary outcomes: Proportion of residents
with any psychotropic drug, with The intervention had no impact on
polymedicine (≥2 drug classes), with average number of psychotropics
therapeutic duplication (≥ 2 same class), prescribed, rates of polymedicine, and
and with nonrecommended and rates of therapeutic duplication.
acceptable drugs in each psychotropic
drug class measured at baseline and one However, there were significant
year post-intervention. increases in number of psychotropics
(p=0.02) and residents with therapeutic
duplication (p=0.04) in the control NHs.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Claesson CB, Schmidt To describe drug use in A randomized, controlled trial. During the intervention, the average Overall, the use of drugs
30 IK Swedish nursing homes number of prescribed drugs increased in deemed inappropriate was
and to assess the impact Setting: 33 nursing homes (NHs) (15 both the intervention (from 7.5 to 7.8) high in Swedish NHs.
“Drug use in Swedish of regular multidisciplinary intervention and 18 control homes). and control (from 7.8 to 8.2) homes.
nursing homes” team interventions on the
quantity of inappropriate Sample: N=1854 residents with an The most frequently prescribed drugs
The multidisciplinary team
Clinical Drug medications. average age of 83 years. Seventy were laxatives (approximately 70%), intervention resulted in a
Investigation percent were female and 42% had followed by minor analgesics and loop decrease in use of
dementia. diuretics. These remained unchanged inappropriate drugs, but an
1998;16(6):441-52. post-intervention. increase in total drug use.
Intervention: Monthly review of drug use The authors primarily
by a multidisciplinary team consisting of a The proportion of residents using attributed this finding to
physician, a pharmacist and nurses over a anxiolytics increased significantly replacing less appropriate
1-year period. Drugs were classified as (p<0.05) in the intervention group but not drugs with more
inappropriate based on the Swedish in the control group. This increase was
Medical Products Agency (MPA) primarily due to increased use of
appropriate ones.
guidelines that mainly focus on limiting oxaxepam, a drug deemed more
drugs that cause confusion and memory appropriate in the elderly according to The impact of the change
impairments. Pharmacists were also the MPA guidelines. in drug use on patient
educated in geriatrics, drug use in the health outcomes could not
elderly, and collaborative methods prior to The proportion of residents using be determined.
the intervention by experts in gerontology. antidepressants significantly (p<0.05)
increased in both intervention and control
Primary outcomes: Number of homes.
prescriptions and number of inappropriate
drugs at baseline and one month post- The proportion of users of neuroleptics
intervention. and hypnotics decreased in intervention
homes, but was not significant.
The proportion of users of inappropriate
medications decreased significantly
(p<0.05) in both intervention and control
homes; the decrease in control homes
was primarily due to switching from
tricyclic antidepressants to selective
serotonin uptake inhibitors (SSRIs).
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Karki SD et al. To examine the A post-test design. At baseline, the number of drugs, doses The multidisciplinary team
31 effectiveness of a team and drug costs were already lower for approach was effective in
“Impact of a team approach to reduce the Setting: A chronic disease hospital/ the intervention unit (A) compared to the reducing the average number
approach on reducing total number of drugs and nursing home in New York. control unit (B). of drugs and doses per day
polypharmacy” simplify patients’ taken by older residents,
regimens. Sample: N=113 residents (N=55 On the intervention unit (A), the average resulting in a reduction in
Consultant Pharmacist Intervention on Unit A and N=58 Control number of drugs per patient decreased drug costs as well.
on Unit B) with a mean age of 74.4 ± 11 from 4.1 to 3.3 (t=1.46, p<0.05); the
1991;6(2):133-37. years. average number of doses significantly The findings are limited by
decreased from 6.2 to 4.9 (t=3.92, the failure to account for initial
Intervention: Team approach based on p<0.05); and the cost of all drugs per differences in the intervention
five guidelines: 1) minimize does patient per day significantly decreased and control units.
frequency; 2) change dosage form from from $2.33 to $1.95 (t=2.69, p<0.05).
liquids to solids; 3) use monotherapy; 4) The impact of the reduction in
use a minimum number of as-needed No significant changes occurred on the medication use on patient
medications; and 5) review drug therapy control unit (B). health outcomes could not be
periodically. The team consisted of a determined.
clinical pharmacist, a geriatrician, and a
nurse practitioner. Joint weekly reviews of
medication profiles were conducted for six
months and appropriate changes were
made based on the above guidelines.
Primary outcomes: Number of drugs per
patient, doses per patient per day at
baseline and six months post-intervention.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Eide E, Schjøt J To evaluate the effects of A pretest, post-test design. Comparison of 1995 and 2000 The pharmacist-lead drug
32 a pharmacist on Intervention NHs information intervention
“Assessing the effects prescribing and Setting: Seven nursing homes (NHs) in improved the use and
of an intervention by a administration of Norway (five intervention and two control The NH population in 2000 was administration of hypnotics in
pharmacist on hypnotics. homes). significantly (p<0.01) older than the 1995 NHs.
prescribing and population (85.7 vs. 83.9 years,
administration of Sample: N=467 residents respectively). Given that the pharmacist
hypnotics in nursing [N=388 Intervention groups (for 1995 - regularly visited all study NHs
homes” baseline, N=201, Mean age=83.9; for The proportion of patients who used (including controls) and
2000, N=187, Mean age=85.7); and N=79 hypnotics was not significantly different. possibly contaminated the
Pharmacy World & Control group for 2000, Mean age=87.2]. However, the use of high doses was findings, the actual impact of
Science Note: No standard deviations were significantly (p<0.01) higher in 2000 the intervention could have
reported. compared those in 1995 (60.0% vs. been underestimated.
2001;23(6):227-31. 38.3%, respectively).
Intervention: Written and verbal drug The impact of the modified
information by a pharmacist regarding the The administration of hypnotics before 9 prescribing and
use of hypnotics based on six rules: 1) p.m. significantly (p<0.01) decreased administration on patient
avoid use if possible; 2) prefer short- from 40.4% in 1995 to 13.3% in 2000. health outcomes could not be
acting compounds; 3) avoid use of determined.
combination; 4) use lowest possible dose; Comparison of 2000 Intervention and
5) use intermittently; and 6) avoid Control NHs
administration early in the evening. This
information was provided to the entire A significantly higher proportion of
staff of each intervention NH in 1995. In control patients used hypnotics (24.1%
addition, the pharmacist visited all vs. 44.3%, p<0.01), used more than one
facilities on a regular basis, but only hypnotic (3.7% vs. 10.1%, p<0.05), and
performed usual pharmaceutical care to was administered hypnotics earlier in the
control institutions. evening (13.3% vs. 62.9%, p<0.01).
Primary outcomes: Use and No significant differences were found in
administration of hypnotics at baseline (in the use of individual hypnotics.
1995) and five years post-intervention (in
2000).
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Roberts MS et al. To evaluate the effects of A cluster randomized, controlled trial. When adjusting for NH clustering, no The clinical pharmacy
33 a year long clinical significant differences in drug use were program significantly reduced
“Outcomes of a pharmacy program on Setting: 52 nursing homes (NHs) in found between intervention and control drug use without adversely
randomized controlled drug use, mortality and Australia (13 intervention and 39 control NHs. affecting mortality and
trial of a clinical morbidity in nursing home homes). morbidity indices.
pharmacy intervention residents. When the clustering effect was not
in 52 nursing homes” Sample: N=3230 residents (Control group: accounted for, the intervention and The authors noted the
N=2325; Intervention group: N=905). In control residents were significantly importance of nursing homes
British Journal of both groups, approximately 90% of the (p<0.0005) different regarding drug use, as a determinant of drug use
Clinical Pharmacology residents were between the ages of 70 with intervention residents using one less outcomes, as evidenced by
and 99 years. drug per person. the reduction of significant
2001;51(3):257-65. findings when accounting for
Intervention: Clinical pharmacy service No significant differences in mortality the NH clustering effect.
program that included three phases: 1) rates, frequency of hospitalizations,
relationship building; 2) nurse education; number of residents experiencing
and 3) medication review by pharmacists. adverse events, and changes in disability
were found between intervention and
Primary outcomes: Drug use, deaths, and control NHs.
morbidity indices (hospitalization, adverse
events and disability) at baseline and
post-intervention.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Furniss L et al. To examine the effect of a A cluster randomized, controlled trial. Overall, the mean number of drugs per A pharmacist’s medication
34 medication review by a resident was 4.91 (range=0-17) at review may reduce drug use
“Effects of a pharmacist. Setting: 14 matched nursing homes (NHs) baseline. The reduction in drug use in without adversely affecting
pharmacist’s medication in the UK (7 intervention and 7 control intervention homes was higher (from 5.1 mental and physical health of
review in nursing homes). to 4.2) than in control homes (from 4.9 to nursing home residents.
homes” 4.4) (p=0.07). However, increases in
Sample: N=330 residents (Control group: behavioral disturbances may
British Journal of N=172, Mean age=78.9 ± 13.7, 67% The most common medications occur.
Psychiatry female; Intervention group: N=158, Mean prescribed included: laxatives (49%),
age=83.5 ± 9.2, 79% female). diuretics (44%), antipsychotics (30%), The authors noted that the
2000;176:563-67. hypnotics/anxiolytics (28%), anti-platelet cost data and comparisons
Intervention: Medication review by a drugs (28%), analgesics (27%), and were limited by sample size.
pharmacist. Use of neuroleptics was antidepressants (25%).
evaluated based on the U.S. Ombudsman
Reconciliation Act (OBRA) guidelines. A Use of neuroleptics was inappropriate in
4-month observation phase was followed 54% of cases, according to OBRA
by a 4-month intervention. guidelines.
Primary outcomes: Number and type of A total of 261 recommendations were
drugs, cognitive function, behavioral made and 239 (91.6%) were accepted by
disturbance, depression and mortality at the general practitioner, with 144 actual
baseline, four months (after observation changes being made.
period), and eight months post-
intervention. The two most common reasons for
recommended changes were: indication
for medication no longer present (85)
and safer or more efficacious use of drug
used (77).
Compared to the control group, the
intervention group experienced
significantly greater deterioration in
behavioral disturbance at Time 2 (mean
difference=-2.2, 95%CI=-4.1, -0.3).
No significant differences in cognitive
function, depression or mortality were
found between the intervention and
control groups.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Bellingan M, Wiseman To examine the effect of a A pretest, post-test design. Diagnosis was absent in 94.1% of The pharmacist-initiated DRR
35 IC pharmacist-initiated drug profiles. was effective in reducing the
regimen review (DRR) on Setting: An elderly care facility in South incidence of DRPs in
“Pharmacist intervention the incidence of drug- Africa. The following DRPs were identified in residents of an elderly care
in an elderly care related problems (DRPs) elderly residents: polypharmacy (≥ 6 facility.
facility” in elderly residents. Sample: N=85 residents (N=73 females drugs) in 68%; potential additive side
and N=12 males). effects in 98%; potential drug interactions The findings are limited by
International Journal of in 64%; excessive dose in 77%; and the fact that a vast majority of
Pharmacy Practice Intervention: A pharmacist-initiated drug duplicate therapy in 34%. profiles had no documented
regimen review (DRR) based on diagnosis.
1996;4:25-29. pharmacological evaluations and clinical A total of 85 recommendations were
assessments. If negative clinical effects presented for 44 patients and 66 (77.7%) The effects of the intervention
were presumably being caused by a DRP, of the recommendations were accepted may have been
recommendations were documented and by providers, but 20 were not underestimated, given patient
presented to prescribers on an implemented because of patient resistance to approximately
intervention form that provided patient resistance. 25% of recommendations.
details, outcome of the clinical
assessment, and DRPs identified in the A total of 76.5% of recommendations
pharmacological assessment. involved high risk drugs (e.g.,
Recommendations were made over psychotropics, NSAIDs, oral
period of 16 weeks. hypoglycemics, and cardiovascular
drugs).
Primary outcomes: Incidence of DRPs
(i.e., absence of diagnosis, additive side The incidence of all DRPs declined over
effect profiles of drugs, drug interactions, pre- and post-intervention periods. The
duplicate therapy, excessive dose, incidence of polypharmacy and drug
polypharmacy) at baseline and one and interactions declined significantly
two months post-intervention. (p<0.01) after the intervention, with
polypharmacy decreasing by 41% and
drug interactions by 39%.
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Avorn J et al. To examine the A randomized, controlled trial. Psychoactive drug use scores for all The education program was
36 effectiveness of an residents with potentially inappropriate effective in reducing the use
“A randomized trial of a educational program for Setting: Six matched pairs (N=12) of drug use (baseline index score ≥ 1) were of psychoactive drugs in NH
program to reduce the health care providers in nursing homes (NHs) in Massachusetts significantly reduced (mean difference in residents, without increasing
use of psychoactive reducing the excessive (six intervention and six control homes). risk reduction= -0.37, 95% CI= -0.08, - disruptive behavior among
drugs in nursing homes” use of sedating drugs in 0.67, p=0.02) in intervention NHs (from residents.
nursing homes. Sample: N=823 residents (Control group, 1.87 to 1.36, a 27% reduction) compared
New England Journal of N=431 and Intervention group, N=392). to control NHs (from 1.74 to 1.60, an 8% The reduction in drug use
Medicine reduction). was not always accompanied
Intervention: A 5-month educational- by positive clinical outcomes.
1992;327(3):168-73. outreach program focused on geriatric Discontinuations or modifications based
psychopharmacology, oriented around on program recommendations were The authors indicated that,
problems in long-term care (e.g., significantly higher for intervention NHs because of the small sample
insomnia, confusion). Two primary compared to controls for: antipsychotics size, findings regarding
program strategies: 1) educational (32% vs. 14%, respectively; mean clinical outcomes were not
material disseminated in three mailings to difference= -18%, 95% CI= -3, -33); robust across different
physicians in intervention NHs and also diphenhydramine (45% vs. 21%, analyses. Thus, larger
used in face-to-face sessions with clinical difference= -24%, 95% CI= -54, 5); and clinical studies examining
pharmacists; and 2) training sessions with benzodiazepines (20% vs. 9%, these outcomes are
nurses and nursing assistants. difference= -11%, 95%CI= -38, 15). warranted.
Target drugs: antipsychotics, In residents who received antipsychotics
diphenhydramine, benzodiazepines, and prior to the intervention and who had a
antidepressants. baseline index score ≥ 1, more
intervention NH residents had
Primary outcomes: psychoactive maintained or improved memory scores
medication use (scores on an index of compared to controls (69% vs. 46%, rate
psychoactive drug use measuring both the ratio (rr)= 0.6, 95%CI= 0.3, 1), but
magnitude and the probable reported more depression (56% vs. 27%,
inappropriateness of medication use); and rr= 2, 95%CI= 1.1, 3.9). In those who
clinical outcomes (mental status, memory, received benzodiazepines or
anxiety, depression, behavior and sleep), diphenhydramine prior to the program,
measured at baseline and 30 days post- more intervention NH residents
intervention. maintained or improved anxiety levels
compared to controls (77% vs. 48%, rr=
0.4, 95%CI= 0.2, 1), but had greater loss
of memory (62% vs. 29%, rr= 2.1,
95%CI= 1.1, 4.2).
Table 2: Detailed Description of Medication Use Intervention Studies
# Article Objectives Study Design Main Findings Discussion
Khunti K, Kinsella B To evaluate the impact of A pretest, post-test design. A total of 71 (65%) of residents had The comprehensive visit to
37 a general practitioner visit repeat prescriptions changed: 53 (51%) the NHs resulted in a
“Effect of systematic and a comprehensive Setting: Four randomly selected nursing at least one drug discontinued; 28 (26%) significant reduction in
review of medication by repeat prescribing review homes (NHs) in Leicester, UK. had at least one drug changed to a inappropriate repeat
general practitioner on on consumption of cheaper alternative; eight (7%) had a prescribing to NH residents.
drug consumption inappropriate medications Sample: N=109 residents, mean age=82.1 new drug prescribed; and one a dose
among nursing-home in nursing homes. ± 8.6 years, 71.0% women. increased. The findings are limited by
residents” the subjective evaluation of
Intervention: One comprehensive visit by The drugs that were most commonly inappropriate and by the lack
Age & Ageing a general practitioner to discuss in detail discontinued, changed, or reduced in of a control group.
all residents of each NH. Prescribing dosage were: histamine2 blockers,
2000;29(5):451-53. records of all residents were reviewed and laxatives, diuretics, antipsychotics, The impact of the intervention
discontinued if deemed inappropriate or antidepressants, non-opioid analgesics, on patient health outcomes
unnecessary. Inappropriateness was and nonsteroidal anti-inflammatory could not be determined.
based on four dimensions: indication, cost drugs.
minimization, contraindications, and
dosage. The mean number of repeat
prescriptions per resident decreased
Primary outcomes: Proportion of residents from 4.3 ± 2.6 before the intervention to
with inappropriate repeat prescriptions 3.5 ± 2.2 after the intervention (mean
after intervention. difference=0.8, 95%CI=0.5, 0.9,
p<0.0001).
Residents who had alterations were on
significantly more drugs compared to
those who did not have alterations (4.9
vs. 3.0, Mann-Whitney U=720,
p<0.0001).