Interventions to Reduce Inappropriate Prescribing of Antipsychotic Medications in People With Dementia Resident in Care Homes: A Systematic Review
Journal of the American Medical Directors Association; Volume 15, Issue 10, October 2014



Antipsychotic medications are generally used to manage the behavioral and psychological signs of dementia. Several large studies have demonstrated an association between treatment with antipsychotics and increased morbidity and mortality in individuals with dementia.


To appraise the effectiveness of interventions employed to reduce inappropriate prescribing of antipsychotics on the elderly with dementia in residential care.


Systematic searches were conducted in 12 electronic databases. Reference lists coming from all included studies and forward citation searching using Web of Science were also conducted. All quantitative studies which has a comparative research design and studies in which recognized methods of qualitative data collection were chosen were included. Articles were screened for inclusion independently by 2 reviewers. Data extraction and quality appraisal were performed by 1 reviewer and checked with a second with discrepancies resolved by discussion with a third if necessary.


Twenty-two quantitative studies (reported in 23 articles) were included evaluating the strength of educational programs (n = 11), in-reach services (n = 2), medication review (n = 4), and multicomponent interventions (n = 5). No qualitative studies meeting our inclusion criteria were identified. Eleven studies were randomized or controlled in design; the rest were uncontrolled before and after studies. Beneficial effects were noticed in 9 from the 11 studies with robust study design with reductions in antipsychotic prescribing degrees of between 12% and 20%. Little empirical information was provided on the sustainability of interventions.


Interventions to relieve inappropriate prescribing of antipsychotic medications to people who have dementia resident in care homes could be effective in the short term, but longer more robust studies are needed. For prescribing levels being reduced eventually, the culture and nature of care settings along with the availability and feasibility of nondrug alternatives needs being addressed.

Antipsychotic medications are often prescribed to control the behavioral and psychological signs and symptoms of dementia (BPSD). However, several large research has demonstrated an obvious association between treatment with antipsychotic drugs and increased morbidity and mortality in those with dementia. Treatment guidelines suggest that the first-line control over BPSD must be detailed assessment to recognize any treatable reason for symptoms (eg, hunger, thirst, pain, infection, loneliness). Furthermore, underlying causes ought to be treated and alternative nonpharmacological interventions explored prior to initiation of antipsychotics. Risperidone is the only antipsychotic licensed within the United Kingdom because of this indication, after which only for short-term use. Nevertheless, other antipsychotic agents in many cases are prescribed and applied to a long-term basis with infrequent medication review. BPSD might cause significant carer stress to family and care home staff that, without intervention, may rapidly lead to acute hospital admission and/or transfer to a more intensive care setting. Antipsychotic medication could be viewed as an easier option than nonpharmacological alternatives, along with the risks are hardly ever discussed or documented. In 2013, the American Medical Directors Association was involved in identifying the most notable 5 items which physicians and patients should question inside long-term care setting as part from the American Board of Internal Medicine Foundation's Choosing Wisely Campaign. Item 4 with this list was “Don't prescribe antipsychotic medications for behavioral and psychological the signs of dementia (BPSD) in those that have dementia lacking any assessment for an underlying cause with the behavior.

The most recent UK audit of primary care data showed a decline in antipsychotic prescribing to those that have dementia from approximately 17% in 2006 to 7% in 2011. The audit showed widespread and significant variation used across the country, which range from approximately 3% of people who have dementia receiving antipsychotic medication on the time in the audit in London and also the southeast to approximately 13% within the northwest. The audit provided no facts about duration of prescription or about the residential setting of individuals with dementia and represents data from approximately 50% of general practices inside the United Kingdom. Audit studies operating out of nursing homes have generally reported a greater prevalence of antipsychotic prescription among people who have dementia.

Anecdotally, we're aware of a variety of interventions being employed to assess, evaluate, and evaluate the prescription of antipsychotic medications in care homes. These include education and raising staff awareness, development and employ of decision-making pathways, medication checklists, mood, pain and behavioral charts, tips on nondrug-based alternatives, regular medication review by pharmacists, community or hospital-based psychiatrists and general practitioners, interdisciplinary education programs, and pharmacist-led strategies.

The intent behind this systematic review was to look at the effectiveness of interventions used to reduce inappropriate prescribing of antipsychotic medications to people with dementia resident in care homes to aid to inform the availability of services. We also were enthusiastic about published accounts of the views and experiences of prescribers of included interventions to focus on barriers and facilitators for the successful implementation of these interventions.


The systematic review was conducted following a general principles published with the NHS Centre for Reviews and Dissemination (CRD). A predefined protocol was made following consultation with topic and methods experts and is also registered with PROSPERO (PROSPERO 2012:CRD42012003425).

Literature Search and Eligibility Criteria

A comprehensive search syntax using MeSH and free text terms was developed by an info specialist (M.R.) in consultation using the review team. The strategy was made for MEDLINE and adapted as befitting the other searched databases (EMBASE, Social Policy and Practice including AgeInfo, and PsycINFO via OVID, CDSR and CENTRAL via The Cochrane Library, CINAHL via EBSCOhost, AMED and British Nursing Index via NHS Evidence, Science Citation Index Expanded and Social Science Citation Index via Web of Science). All databases were searched from inception to November 2012. Update searches were run in November 2013. No date, study design, or language restrictions were imposed. The reference lists coming from all included articles and identified review articles were checked for additional relevant studies. Forward citation seeking each included article was conducted using ISI Web of Knowledge.

We were interested in the effectiveness of interventions (eg, staff training, regular medication review) designed to lessen inappropriate prescription of antipsychotic medications to people who have dementia in community residential care settings. Interventions had being aimed at professionals (eg, general practitioners, community psychiatrists, pharmacists) accountable for prescription of these medications of these settings. We also were thinking about reports from the views and experiences of prescribers with all the included interventions.

All quantitative studies reporting comparative data were included. Qualitative studies using recognized strategies to qualitative data collection (eg, focus groups, interviews, and observation) and analysis (grounded theory, narrative analysis, thematic analysis, discourse analysis) were sought.

Study Selection

The listings were uploaded to reference management software (Endnote X5, V5; Thomson Reuters, Philadelphia, PA). Titles and abstracts were screened for relevance independently by 2 reviewers (J.T.C., M.R., or R.A.), with any disagreements being resolved by discussion and involvement of a third reviewer (J.T.C., M.R., or R.A.) where necessary. The full text of potentially relevant articles was retrieved and screened inside the same way with all the prespecified inclusion and exclusion criteria. All duplicate articles were double-checked and excluded.

Data Collection

For each study, details in the intervention, the functions of those receiving it, the characteristics of the patient population involved, the setting, the analysis methods, and outcomes concerning medication use were recorded. Data were extracted by one reviewer (J.T.C. or M.R.) in to a data extraction form based around the Cochrane Effective Practice and Organisation of Care Review Group Data Collection Checklist,16 that was piloted on several studies and refined. The Cochrane Effective Practice and Organisation of Care Review Group Data Collection Checklist features a taxonomy of intervention components, which has been completed for each and every trial included in this process. Data were collected from published articles only; manuals just weren't requested from trial authors. All data extraction was checked with a second reviewer (J.T.C. or M.R.) with discrepancies resolved by discussion and involvement of an third reviewer (R.A.) where necessary.

Risk of Bias

The quality coming from all included studies was appraised by one reviewer (J.T.C.) and checked by way of a second (M.R., R.A., or R.W.). In an amendment on the published protocol, all articles were appraised using the Effective Public Health Practice Project tool17 allow assessment of most study designs while using same rubric. Appraisal considered the strategy of sample selection, potential for bias linked with study design, differences between groups at baseline and just how these were dealt with inside analysis, assessment of outcome measures, description in the flow of patients through the study, and employ of a valid and reliable primary outcome measure.

Data Synthesis

Changes in medication use were reported in most included studies. However, the plethora of different formats where the data were provided along with the range of included study designs precluded formal pooling from the data. For example, one of the randomized studies, medication use was variously reported as psychoactive drug use score, proportion of residents who had antipsychotic medications discontinued, variety of days of antipsychotic therapy per patient each month, proportion of residents taking antipsychotic medications, and dose of antipsychotic medication. Data were therefore tabulated, grouped in accordance with study design and outcome, and discussed narratively.


The electronic searches retrieved a complete of 5071 unique citations. Screening of title and abstracts up against the inclusion and exclusion criteria resulted inside the retrieval of the full text of 80 articles. Fifty-nine articles were excluded because the following aspects in the article did not meet the inclusion criteria: population (n = 3), intervention (n = 14), reported outcomes (n = 1), and look design (n = 32). Six articles were published as conference abstracts just with insufficient information provided and now we were unable to find a full-text publication despite exposure to authors, and 3 were duplicate publications. One additional article was located through hand searching with the bibliographies of identified systematic review articles. The update search identified one more 985 articles, of which 7 were retrieved completely text and 1 article met the inclusion criteria. A total of 23 articles were included, describing 22 studies. Figure 1 shows the flow of studies from the review. Table 2 shows case study characteristics of most included articles. All the included studies provided quantitative data. We didn't identify any articles reporting the views and experiences of prescribers with specific interventions. Our search identified several qualitative articles exploring factors that influence prescribing practice in care homes; they're considered further in the discussion.

Study Characteristics

Six in the studies are randomized, 5 have a very controlled design, and 11 are uncontrolled before and after studies. The studies were published between 1987 and 2013 and were conducted inside the United States (n = 8), the United Kingdom (n = 5), Canada (n = 5), Australia (n = 2), Norway, and Sweden. Very little demographic information was provided concerning the people (physicians, nurses, pharmacists, and so forth) who received the interventions along with most studies it's not clear how many prescribers were involved. The studies ranged in proportions from 21 to 7000; approximately 19,300 people with dementia were a part of total (information not provided in all of the studies).

Intervention Characteristics

Descriptions with the interventions used inside the studies are shown in Table 3. We grouped studies as outlined by intervention type using 4 categories: educational programs (n = 11 studies), in-reach services (n = 2 studies), medication review (n = 4 studies), and multicomponent interventions (n = 5 studies). The EPOC Data Collection Checklist incorporates a taxonomy of intervention components grouped under 4 headings: professional, organizational, structural, and regulatory. The interventions within studies of educational programs consisted mainly of professional components, like educational meetings, distribution of educational materials, and educational outreach. In-reach services21 and 26 contained mainly organizational and structural components. Studies containing one of the most variety were those in the medication review and multicomponent intervention groups incorporating educational, organizational, structural, and regulatory interventions. In many cases, there were insufficient information provided inside the article to duplicate the intervention in another setting.

Using the EPOC Data Collection Checklist classification, the number of intervention components per study ranged from 1 to 7; most studies was comprised of 3. The most frequently employed intervention component was educational outreach (14 studies), and this was evident across all 4 types of intervention. Educational outreach was defined as the use of your trained one who met with providers within their practice settings to offer information while using intent of changing the provider's practice.

Study Quality

Assessment from the quality of each one included study is shown in Table 4. The global assessment of approximately a third with the studies was moderate or strong. The main regions of weakness were in the collection of primary outcome data and inside reporting of withdrawals and dropouts. In most from the studies, the results assessor was aware of the intervention status of participants and the study participants (prescribers) were aware in the research question. Although data on prescribing rates were taken from patient and pharmacy records on many occasions, the data-collection process was performed by one individual with no technique of checking accuracy. Furthermore, the data-collection tool was often not described, precluding judgment about the validity of the measure. In most studies, there were little information provided around the numbers of and reasons behind withdrawals and dropouts of either prescribers or patients. In Table 4 we now have assessed reporting of withdrawal and dropouts of patients; the reporting of the flow of prescribers was assessed as weak in most but 5 studies.

Medication Use

Educational programs (randomized and controlled study designs n = 7)

Despite considerable differences inside nature and implementation with the educational programs used, introduction of an program to boost the treating BPSD behaviors and improve appropriate prescribing of antipsychotic medications had beneficial effects in every 4 randomized studies and in 1 of the controlled studies. Four from the 5 showed home loan business medication use inside the intervention group compared with all the control gang of between 12% and 20%. Although Testad and colleagues reported no significant differences between groups inside change in proportion of residents taking antipsychotic medication, this became against an identification of reductions in restraint use and agitation.

The intervention did not influence prescription rates inside the 2 remaining studies. These are the largest studies inside review in terms from the number of patients how the intervention was ultimately directed at, although the volume of physicians receiving training was relatively low, and inside study by Ray and colleagues, training was not offered to nursing along with other care home staff. Explanations for your lack of effect offered from the authors of the articles include the simultaneous introduction and promotion with the use of atypical antipsychotics during the study period, a reflection of the wide variation in antipsychotic prescribing in care homes after a while, and barriers to reducing antipsychotic prescribing including the increased time commitment essential to implement alternative methods of behavior management. Educational programs (before and after study designs n = 4)

The is a result of these studies are harder to interpret, as it isn't clear how many other factors influenced prescription rates over case study period. Results showed similar trends to those affecting studies of a better quality design. These are smaller single or 2-center studies involving between 53 and 300 patients in addition to their associated care staff. The interventions ended in a reduction in antipsychotic use (variously reported) in 3 studies. The baseline degree of antipsychotic use inside study reported by Earthy and colleagues was low and little changed with the intervention (increased from 17% to 19%). However, the authors report improvements in documentation, enterprise administration of “as-needed” medication by nursing staff as well as a decrease inside the frequency of problem behaviors. In-reach services (randomized and controlled study designs n = 2)

Both of the studies involved improved multidisciplinary teamwork either with a psychiatric team or a pharmacist spending time working at care homes supporting the care home staff. In both studies, there have been statistically significant reductions in prescription rates associated using the intervention (19%; P = .007 21 and 16%; P < .0001 26); however, reductions also were affecting the control groups in studies partly 21 or wholly 26 negating the impact of the intervention.

Medication review (randomized n = 1 and before study designs n = 3)

The study reported by Patterson and colleagues provides essentially the most robust evidence in the effectiveness of this process to reducing inappropriate prescribing. The intervention used was also essentially the most sophisticated and used some in-reach in addition to medication review, with specially trained pharmacists visiting intervention homes monthly for 12 months to review prescribing information and guide prescribing decisions. The authors reported a substantial difference between intervention and control homes within the proportion of residents taking inappropriate antipsychotic medications (20% vs 50% odds ratio = 0.26; 95% confidence interval 0.14–0.49). The design in the remaining 3 studies permits the contemplation on trends in results only. Two used audit and feedback and reminders to examine medication needs on a regular basis and these ended in minimal alterations in prescribing rates. The final study was conducted against a background of adjustments to accommodation conditions for that residents in a way that they were moved in a specialized, secure dementia unit. Perhaps unsurprisingly, prescription rates were reduced from the extremely high (95% of residents receiving antipsychotic medication) to a much lower proportion (58%), although it's not possible to ascertain whether this is due for the change in accommodation or intervention.

Multicomponent interventions (controlled n = 1 and before study designs n = 4)

The 5 studies using multicomponent interventions ranged in complexity from a study involving 3 components, audit and feedback, continuity of care, and change for the site of service delivery36 to 7 components incorporating education, audit and feedback, and structural changes. Studies also varied widely in proportions, and were implemented among 1 and 25 homes. All studies showed reductions in prescription rates (including 5% to 66%) associated using the intervention, although only the research reported by Westbury and colleagues was controlled.

Long-term effects of interventions

Only 4 studies assessed whether changes to prescription levels achieved throughout the intervention period were maintained. Two studies reported a return to baseline antipsychotic prescription levels. Testad and colleagues18 reported that medication levels remained constant 6 months following the end with the intervention. Finally, Rovner and colleagues reassessed psychotropic drug use 9 months following the end in the study period and discovered the effects inside intervention on prescription rates ended up maintained. Detail is sparse since these follow-up visits were outside with the formal trial period, but it's likely that this extent which procedures used during the study continued to get used varied between sites both inside same trial and between trials. For example, Monette and colleagues commented that although staff on the long-term care centers had expressed an intention to take some from the program components, none were systematically adopted after the study. In contrast, Rovner and colleagues attribute the maintenance with the effect in the intervention of their study to a ongoing requirement of physicians to finish an “indications and side effects” document per resident receiving psychoactive medication.


Principal Findings

This could be the first systematic review to specifically synthesize evidence with the effectiveness of interventions to cut back inappropriate prescribing of antipsychotics to individuals with dementia resident in care homes. Irrespective from the nature from the intervention, in the studies with robust design, antipsychotic prescription rates were seen to fall as a result in the intervention. Although, more difficult to interpret, similar effects were also noticed in the less well-designed studies. There is little information in the included studies to help you understanding with the sustainability of the effects of interventions. Furthermore, one of the striking features of this body of literature is that it spans 27 years, with all the earliest trial reported in 1987. Over now, there have been a number of initiatives, including adjustments to regulations and widely disseminated guidance aimed at limiting the use of these agents, but evidently prescribers still find compelling why you should use them.

Results in Context

This work highlights 2 key problems that have been illustrated in the past systematic reviews of related areas: (1) troubles of changing practice within care homes and (2) the scarcity of good-quality research conducted on this setting. This body of literature spans a long time period through which research and reporting methods have improved considerably; however, 6 of the included before studies were conducted inside last 4 years. We specifically searched for qualitative information around the views and experiences of prescribers while using included interventions, but disappointingly were not able locate any articles meeting our inclusion criteria. Studies exploring factors that influence prescribing behavior more generally suggest various factors might be involved. These include shortfalls over time, staffing levels, and staff training that impact on nonpharmacological options to antipsychotic medication being considered viable, a pressure from loved ones and carers to prescribe as well as a misconception with the likelihood an individual might reap the benefits of antipsychotic medication. Other studies that have looked at implementation of interventions for other purposes in care home settings have identified the need for involving family in decision-making within the successful treating behavioral problems as well as the management of incontinence. A systematic review from the implementation of psychosocial interventions for individuals with dementia in care homes found that active engagement of care-home staff and members of the family played a crucial role in successful implementation. Similarly, systematic reviews for the more general topic of improving prescribing practice in care homes also are already unable to explain recommendations for future practice due on the varied nature from the design, interventions, outcomes, and results and the poor quality of included studies.

Strengths and Limitations of Our Study

This systematic review followed best practice guidelines for systematic reviews, is reported according to the PRISMA statement,54 and could be the first within this topic area. Extensive electronic searches that weren't limited by date, study design, or language were augmented with forward and backward citation searching of all included articles, and authors of conference abstracts were contacted for their data, where possible. We are, therefore, certain that this review encompasses most if not all the available data with this topic.

We focused the review on one outcome measure, change in medication use, but were not able perform a meta-analysis from the randomized many studies because from the variety of formats in which these data was presented. This is undoubtedly a limitation of the review but because of the uniformity in the direction in the effect in most in the studies, the small quantity of randomized numerous studies identified, and also the accompanying variation and complexity inside interventions used, it really is unlikely that a pooled result offers any more useful insight compared to the synthesis we present. Although the results of the before and after studies are hard to interpret, since there may are actually other influences on prescribing during the analysis period, they feature a full picture from the spectrum of interventions that are already evaluated and add weight towards the evidence, as interventions implemented in less tightly controlled conditions also may have produced positive results. We had hoped to explore in more depth whether specific attributes or implementation approaches impacted on the strength of interventions. Because of the relatively small variety of robust studies within each category along with the lack of reported detail, this became not possible, although we've got used an established method of characterizing the constituents of interventions16 to deliver the reader with the maximum amount of detail as you can.

Implications for Practice and Research

The overall picture is one where it would seem how the current guidelines to limit antipsychotic prescribing are tough to implement in the day-to-day reality of practice, whilst juggling ethical concerns, staffing levels, staff competence with nonpharmacological alternatives, as well as the wishes of distressed relatives and carers. Large, quality, well-reported, randomized research within the care home setting with accompanying process evaluations would enable a much better understanding of the environment and its impact on successful implementation of interventions. Further qualitative work to educate yourself regarding the barriers and facilitators for the appropriate prescription of antipsychotic medications will support efforts to achieve sustained change in the varying specific contexts of human care and nursing homes.


Interventions to reduce inappropriate prescribing of antipsychotic medications to people with dementia resident in care homes may be effective inside short term, but longer-term, more robust studies are needed. For prescribing levels to be reduced in the long term, the culture and nature of care settings and also the availability and feasibility of nondrug alternatives needs being addressed.




“Under the care of Leo J. Borrell, M.D. since December 2001, I have seen a remarkable improvement in my mother’s condition. She is responding dramatically to the new regiment Dr. Borrell has prescribed”

- Beth Rose


Feb 3, 2008

The Interdisciplinary Team; The Role of the Psychiatrist

by Dr. Leo J. Borrell, featured in Assisted Living Consult for November/December 2006. A HealthCom Media Publication