Table 1.

Study characteristics of the included studies

Study ID	Author	Year	Country	Design	Setting	Sample size	Age	Intervention	Follow-up	Outcomes	Summary of results
1	Avorn et al. [14]	1992	USA	Cluster RCT	Nursing homes	NHs = 12 NHR = 823	65 and older	- physicians who is there prescribing of psychoactive drugs was above threshold at the baseline evaluation were invited by pharmacists for separate sessions (3 interactive visits). - All physicians of NHR received 6 literature summaries (insomnia, behavioural problems) in 3 mailings. - 4 training sessions were delivered to nurses/ nursing assistants on geriatric pharma psychology, alternatives to psychoactive drugs	5 months	-psychoactive drug use scores -proportion of residents using antipsychotics.	-psychoactive drug use mean in intervention 27% compared to 80% in control group (P = 0.02). -antipsychotics ceased in intervention 32% versus 14% in control. - no of days/patient/ month greatly reduced in intervention than control. -no of non-recommended hypnotics ceased and substituted with alternative drugs/discontinued completely were 45% in intervention versus 21% control.
2	Rovner et al. [15]	1996	USA	RCT	Nursing homes	1 NH (250 bed community NH) NHR = 89	65 and older	-As ap art of a dementia care program: implementation of new prescribing guidelines based on protocol for psychotropic drug management -Educational rounds weekly for 1 h to discuss patient’s behavioural disorders, and medical status	6 months	- antipsychotic drug - behaviour disorders -restraint use, - and cognitive/ functional status.	-Statistically significant 71% reduction in agitation with intervention versus 49% with control
3	Meador et al. [16]	1997	USA	RCT	Nursing homes	NHs = 12 NHR = 1311	65 and older	-geropsychiatrist delivered educational visit to physicians (45-60 min) -NH staff received 5–6 1-h programmes over 1 week delivered by nurse educator. -after 1 month, follow up sessions - when requested, evening meetings for families.	6 months	-Proportion of APs drug use in days/ 100 /days of stay. -severity and presence of behavioural symptoms.	-APs use per 100 days at baseline in intervention gp decreased from 25.3 days to 19.7 per 100. -Aps reduction is 23% in intervention gp to control gp. −33% NHR in intervention gp had their antipsychotics ceased.
4	Schmidt et al. 1&2 [17, 18]	1998	Sweden	Cluster RCT	Nursing homes	NH = 33 NHR = 1854	65 and older	-Monthly multidisciplinary meetings led by pharmacist for 12 months	12 months	-Proportion of residents with psychotropics -non recommended hypnotics, antidepressants.	−19% of residents in the intervention gp ceased APs (p = 0.007). −37% of residents ceased non- recommended hypnotics in intervention gp (p < 0.001).
	Schmidt et al. [19]	2000	Sweden	Cluster RCT	Nursing homes	NH = 36 NHR = 1549	65 and older	Nursing homes participated in 1995 were followed up.	three-year follow-up	Medication appropriateness	-proportion of residents prescribed non-recommended hypnotics were lower (14.0%) compared to previous study 1995 (19.0%). - in1998 5% of residents were prescribed non recommended hypnotics compared to control gp (10.1%).
	Claesson et al. [20]	1998	Sweden	Cluster-RCT	Nursing homes	NH = 33 NHR = 1854	65 and older	- regular multidisciplinary meetings (physician, pharmacist, NH nurses/assistant) reviewed resident’s drug use on a monthly basis over 12 months. -education for selected pharmacists (5 occasions = 65.5 h), topics were drug use in elderly, geriatrics.	14 months	Medication-related problems	-NH residents were prescribed on average 7.7o (range: 6–11) medications. - laxatives (70%) -psychotropic drugs (77%, range: 50–95%).
5	Furniss et al. [21]	2000	England.	Cluster RCT	Nursing homes	NH = 14 NHR = 330 residents: (172 ctrl, 158 Int)	65 and older	-Medication review led by pharmacist. - pharmacist review the medications at NH, GPs surgery, or over phone. -Pharmacist collected details of current medication, medical history and current problem identified by nursing home staff. - 3 weeks post-medicine review, NH were revisited to identify any problems and to ascertain on whether changes had been implemented.	8 months	-no of prescribed medications -Types of medications, reason for using neuroleptic medications. -hospital admission (in-patient days) -MMSE -GDS -BASDEC -CRBRS -Falls and death	- 239 of recommendations accepted by GP (91.6%). -change of medications =144 -In total MMSE were declined. - Mean CRBRS scores increased in Int compared to ctrl - deaths in ctrl were higher than Int NHs.
6	Stein et al. [22]	2001	USA	Cluster RCT	Nursing homes	NH = 20 NHR = 147	65 and older	-Staff training sessions (30 min) -Study physician visited/telephone to all primary care physicians -physicians received messages about NSAIDs risks and benefits , algorithm for stopping NSAIDs, or aternatives such as paracetamol or topical agents and non-pharmacological management for pain.	3 months	NSAIDs and paracetamol Use in the past week	-Mean number of days of NSAIDs use deceased in Int gp from 7.0–1.9 days compared to ctrl gp (7.0–6.2 days), P = 0.0001 - paracetamol use in Int gp increased (3.1 days) compared to ctrl (0.31 days), P = 0.0001.
7	Roberts et al. [23]	2001	Australia	Cluster RCT	Nursing homes	NH = 52 NHR = 3230	65 and older	-nurse education (6–9 problem-based education sessions) including geriatric medications and common problems in long care such as depression & pain. -supported by bulletins, wall charts and clinical pharmacist visits. - clinical pharmacist average contact 26 h/NH -clinical pharmacist reviewed drug regimen for 500 residents selected by home staff.	12 months	-Mortality rate -hospital admission -Drug use -ADEs -Medication-related problems	-mean no of psycholeptics administered /resident in Int gp decreased (− 0.14,95% CI − 0.28-0.0, p = 0.044) - in the intervention group mean number of benzodiazepines Administered/ resident reduced (− 0.06, 95% CI − 0.06 to 0.04, p = 0.29).
8	Crotty et al. (a) [24]	2004	Australia	Cluster RCT	Aged care facility	NH = 10 NHR = 154	65 and older	−2 multidisciplinary case conferences were conducted 6–12 weeks. -pharmacists, geriatrician, residential care staff, GP, and a representative of the Alzheimer’s Association of South Australia. -medication review prepared by the resident’s GP before case conference.	7 months	-MAI score	-Mean MAI score in Int gp 4.1 (2.1–6.1) versus 0.4 (0.4–1.2) in ctrl gp. - benzodiazepines: mean MAI score in int.gp 0.73 (0.16–1.30) versus − 0.38 (−1.02 to 0.27) in ctrl gp.
9	Crotty et al. (b) [25]	2004	Australia	RCT	long-term care facility/hospital discharge	NH = 85 NHR = 110 Discharged from 3 hospitals	65 and older	-pharmacist transition coordinator transfers the medication-related information to the family physician and community pharmacist. -case conference at facility within month of transfer include pharmacist, nurse, family physician, community pharmacist,	8 weeks	-MAI score -Hospital admission -Medication related problems -ADEs, -falls	-No change in MAI score in Int gp 2.5, 95% CI1.4–3.7) -In ctrl gp MAI score had worsened 6.5, 95%CI 3.9–9.1)
10	Crotty et al. (c) [26]	2004	Australia	RCT	Residential care facilities	NH = 20 NHR = 715	65 and older	-Educational intervention: two (30 min) outreach visits of pharmacists to doctors. - presenting detailed audit information on psychotropic use, stroke risk reduction, and fall rates. −4 (2 h training sessions) for link nurse in each facility.	7 months	-MAI score -Hospital admissions - MRP	-No significant difference in psychotropic drug use before &after intervention (0.89,95%CI 0.69–1.15). -PRN of antipsychotics drug use increased in Int gp compared to ctrl gp (4.95,95%CI 1.69–14.50). - No significant difference in BZD drug use before & after intervention (0.89,95%CI 0.69–1.15). - No significant difference in falls (1.17, 95%CI 0.86–1.58).
11	Fossey et al. [27]	2006	UK	Cluster RCT	Nursing homes	NH = 12 NHR = 349	65 and older	-Training and support to care staff on non-pharmacological interventions, alternatives to neuroleptic use. -Medication review by Led by old age psychiatrist, senior nurse every 3 months -contact between psychiatrist and prescribers to provide and wrote prescribing recommendations	12 months	-Proportion of residents receiving neuroleptics. -CMAI - QoL	- reduction in neuroleptic use/resident (19.1, 95% CI 0.5–37.7%, P = 0.045) --Neuroleptic use decrease 24% in exp. (47 to 23%) but increased in ctrl 7.6% (49.7 to 42.1%). -No significant changes in CMAI
12	Zermansky et al. [28]	2006	UK	RCT	Nursing homes and residential homes	NH = 65 NHR = 661	65 and older	- Pharmacist medication review by using the resident’s medical record. - consultation with the resident’s and carer. -pharmacist forward written recommendations to GP.	6 months	-no. of changes in medication/patient -Hospital admissions -Medication-related problems -Medicine costs -Number of medicines per participant - Mortality - Falls - SMMSE -Barthel index -GP consultations	- Increase in mean number of drug changes/patient ctrl: 2.4 versus 3.1 in Int (P < 0.01) -no of falls reduced significantly - pharmacist recommendations accepted (75.6%), and 76.6% of these recommendations were implemented.
13	Gurwitz et al. [29]	2008	USA and Canada	Cluster RCT	Two large long-term care facilities.	Facility = 2 Residents = 1118	65 and older	-Computer program (order entry with clinical decision support system). - more than 600 potentially serious drug-drug interactions alerts were reviewed. -no of ADEs were identified (preventable events including errors and drug-drug interactions were determined). -alerts included in the CDSSs were assessed to determine if any of them could have prevented the prescribing of these drugs.	1 year in one facility and 6 months in the other	-Number of preventable ADEs - ADEs severity - ADEs preventability	-None ADEs = (1.06,95% CI 0.92–1.23) Preventable ADEs = (1.02,95% CI 0.81–1.30)
14	Field et al. [30]	2009	Canada	Cluster RCT	long-term care facility	-One long-term care Facility - 22 long-stay units Residents= 833	65 and older	The 22 long-stay units were randomly assigned - for Intervention units’ prescriber: Alerts related to medication prescribing for residents with renal insufficiency were displayed. -Control units: Alerts hidden and tracked - The types alerts were: maximum recommended daily dose/frequency of administration, medication to be avoided, and missing information.	12 months	-Proportion of final drug orders alert that were appropriate	-Appropriate final drug orders proportion were high in Int (1.2, 95% CI 1.0–1.4) for frequency. -for drugs that should be avoided (2.6, 95% CI 1.4–5.0). for missing information (1.8, 95% CI 1.1 to 3.4). -Appropriate final drug orders Significant in Int (1.2 95% CI 1.0–1.4).
15	Patterson et al. [31]	2010	Ireland	Cluster RCT	Nursing homes	NH = 11 NHR = 334	65 and older	-intervention homes were visited monthly by trained pharmacists for 1 year. Resident’s information was collected from records, GP and community pharmacist. Interviews were conducted with the residents and next of kin to assess the need for medicines. - applied an algorithm to assess appropriateness of psychoactive medication and worked with GPs to improve the prescribing of these medications.	Monthly for 12 months	-Proportion of residents prescribed inappropriate psychoactive medications. -no of falls	- At 12 months, residents taking inappropriate psychoactive medications in Int gp (19.5%) decreased compared to ctrl gp (50%) intervention homes (0.26, 95% CI 0.14–0.49) -No change the falls rate
16	Testad et al. [32]	2010	Norway	Cluster RCT	Nursing homes	NH = 4 NHR = 211	65 and older	-Education and training program (2 days seminar and monthly group guidance for six months).	12 months	-% of residents using antipsychotic drugs - Restraint use	-No statically significant difference in antipsychotic use. - Significant reduction in Aggression in Int gp at 6 & 12 month follow-up. -Significant reduction in proportion of residents restrained at 6 months but not at 12 months.
17	Lapane et al. [33]	2011	United States	Cluster RCT	Nursing homes	NH = 25 NHR = 3321	65 and older	- GRAM is automatically generated to assist consultant pharmacists identify residents at risk for delirium/ falls -Detailed instruction of consultant pharmacists providing targeted medication review for all residents at high-risk. - Reports within 24 h of admission and used during monthly review.	12 months	− Mortality − Hospital admission potentially due to ADEs.	-Mortality rate /1000 resident-months, HR: 0.90 (adjusted HR 0.89, 95% CI 0.73–1.08) -Hospital admission/1000 resident-months, HR: 1.13 (adjusted HR 1.11, 95% CI 0.94–1.31).
18	Pope et al. [34]	2011	UK, Ireland	RCT	Nurse-managed continuing-care	NHR = 10 nurse-managed continuing-care Residents = 225	65 and older	-medical assessment by a geriatrician, and using Beer’s criteria for multidisciplinary panel medication review. - recommendations forwarded to the GP. - after 6 months, reassessment occurred	6 months	-no of drugs prescribed -mortality -medication cost	−92.7% of patients received medication recommendations and 80.1% accepted. - total number of medications/ patient/d reduced in Int gp (11.64–11.09 compared to ctrl 11.07–11.5).
19	Kersten et al. [35]	2013	Norway	RCT	Nursing homes	NH = 22 NHR = 87	65 and older	-A paper-based review with a view to reduce ADS scores were conducted by clinical pharmacist. -clinical pharmacist discuss discontinue or replace an anticholinergic drug with the physician before changes were implemented.	8 weeks	- Cognitive function - anti-cholinergic side-effects	- cognitive function not improved - anti-cholinergic side-effects not improved
20	Milos et al. [36]	2013	Switzerland	RCT	Nursing homes or community	NHR = 279	75 years or older	Pharmacists-led medication review that included assessment of relevant parts of (EMRs) and collection of patient’s blood sample data. - clinical pharmacist-initiated medication reviews based on the background information to identify DRPs.	2 months	- no of PIMs. - DRPs	−6% decreased in PIM in Int gp -Total no of DRPs in the intervention group was 431 [mean 2.5 (1.5) / patient (range 0–9) - No significant difference between the no of DRPs in nursing home patients [mean 2.53 (1.33)] and community-dwelling patients [mean 2.55 (1.29)] Significant in changes in the actions taken by the physician were for lowered dosage.
21	Frankenthal et al. [37]	2014	Israel	RCT	chronic care geriatric facility	NH = 1 NHR = 359	65 and older	-medication review conducted by pharmacist -to identify PIMs and PPOs medications screened with STOPP/ START criteria then followed up with recommendations to the chief physician. - chief physician decided to accept or not.	12 Months	-medication appropriateness -mortality -hospital admission -QoL -MRP -medication cost	-significant decreased in the average number of drugs prescribed in Int gp (P < .001). - significant decreased in the average number of falls in Int gp (P = .006). -decrease in the average drug costs in Int gp by US$29. - hospitalization, FIM scores, and QoL were same in both groups.
22	García-Gollarte, et al. [38]	2014	Spain	Cluster-RCT	Nursing homes	NH = 36 NHR = 716	65 and older	−30 doctors received educational intervention. - The educational intervention included general drug use in elderly, STOPP START workshop, and adverse drug reactions in older people. -participants also received educational material and references - on-demand support (via phone) for 6 months provided by the educator.	6 Months	- Medication appropriateness (STOPP-START) -Hospital admissions Medication appropriateness (STOPP-START) -Falls	- The mean number of inappropriate drugs was higher in ctrl gp (1.29–1.56) compared to Int gp (0.81–1.13). -no of falls increased in the ctrl gp from 19.3–28% and not significantly change in the intervention group from 25.3–23.9%.
23	Pitkala et al. [39]	2014	Finland	Cluster-RCT	Assisted living facilities	Facility = 20 Residents = 227	65 and older	-two 4-h interactive training sessions for nursing staff aimed to enable nurses to recognize potentially harmful medications and corresponding adverse drug events. -the second 4-h sessions: case-study-based. - nurses in this intervention were asked to identify potential MDR and highlight these to the consulting doctor.	12 months	-Medication appropriateness -Hospital admissions -Mortality -QoL -MMSE	-mean number of potentially harmful drugs lowered in int gp (−0.43, 95% CI-0.71 to −0.15) and not changed in ctrl gp (+ 0.11, 95% CI − 0.09 to + 0.31) (P = .004). -HR QoL decreased in Int gp (− 0.038, 95% CI − 0.054 to − 0.022) compared to ctrl gp (− 0.072,95% CI − 0.089 to − 0.055) (P = .005). -hospital admission decreased significantly in int gp (1.4 days/person/year, 95% CI 1.2 to −1.6) compared to ctrl gp (2.3 days/person/year; 95% CI 2.1to −2.7), RR = 0.60, 95% CI 0.49 to − 0.75, P < .001).
24	Connolly et al. [40]	2015	New Zealand	Cluster-RCT	RACFs	NH = 36 NHR = 1998	65 and older	- Gerontology nurse specialist delivered staff education and clinical coaching. - benchmarking of resident indicators including restraint use, falls, etc.). - multidisciplinary team meeting (1 h) monthly for the first 3 months.	14 months	- Hospital admissions (ambulatory sensitive hospitalisations, total acute admissions). -Mortality	-no differences between Int and ctrl gp in rates of ambulatory sensitive hospitalisations admission (1.07; 95% CI 0.85–1.36; P = 0.59). -no difference in mortality (1.11; 95% CI 0.76–1.61; P = 0.62).
25	Potter et al. [41]	2016	Australia	RCT	RACFS	Facility = 4 Residents = 95	65 and older	-medication review followed by discontinuing non-beneficial medications conducted by a GP and a geriatrician/clinical Pharmacologist - During deprescribing, the GP reviewed participants weekly.	12 months	-no of falls -mortality -no of fallers -cognitive function -QoL	-mortality 26% in int gp and 40% in ctrl gp (HR 0.60, 95%CI 0.30 to 1.22). -QoL Changes in Int gp (− 1.0 ± 4.3) compared to ctrl gp (−  1.0 ± 4.7). -Falls -Patients with one or more falls in int gp (0.56, 95% CI 0.42–0.69) compared to ctrl gp(0.65, 95% CI 0.50–0.77), (p = 0.40)

Abbreviations: RCT Randomised Controlled Trials, NHR Nursing Home Residents, NHs Nursing Homes, CDSSs Computerised Clinical Decision Support Systems, CI Confidence Interval, GP General Practitioner, no number, min minutes, hr. hour, APs Antipsychotics, gp group, ctrl control, Int intervention, MMSE Mini-Mental State Exam, GDS Geriatric Depression Scale, BASDEC Brief Assessment Schedule Depression Cards, CRBRS Crichton-Royal Behaviour Rating Scale, P, p value, NSAIDs Non-Steroidal Anti-Inflammatory Drugs, ADEs Adverse Drug Events, MAI Medication Appropriateness Index, MRP Medication-related Problem, PRN pro re nata (when necessary), BZD Benzodiazepine, CMAI Cohen-Mansfield Agitation Inventory, QoL Quality of Life, exp experiment group, % percentage, GRAM Geriatric Risk Assessment Med Guide, HR Hazard Ratio, ADS Anticholinergic Drug Scale, EMRs electronic medical records, PIMs Potential Inappropriate medications, PPOs potential prescription omissions, STOPP/START Screening Tool of Older Person’s potentially inappropriate Prescriptions and Screening Tool to Alert doctors to Right Treatment, FIM Functional Independence Measure