Journal Logo

Original Articles

Duration of Antipsychotic Medication Use by Aged Care Facility Residents With Dementia

Lind, Kimberly E. PhD, MPH; Raban, Magdalena Z. PhD; Georgiou, Andrew PhD; Westbrook, Johanna I. PhD

Author Information
Alzheimer Disease & Associated Disorders: October–December 2019 - Volume 33 - Issue 4 - p 331-338
doi: 10.1097/WAD.0000000000000336


In 2015, an estimated 46.8 million people had dementia worldwide; this is projected to increase to 74.7 million by 2030.1 Behavioral and psychological symptoms of dementia (BPSD) are common across all stages of dementia; agitation and irritability are most frequent.2 Antipsychotic medications are sometimes used for BPSD, after behavioral interventions have been unsuccessful. However, antipsychotic medications have modest effects on BPSD symptoms, and carry extrapyramidal side effects, and increased risk of cerebrovascular events and death.3–9 Because of these risks, if antipsychotics are used to manage BPSD symptoms, consensus guidelines state that use should be for the short term (<12 wk) with regular review.10–15 Some guidelines allow use for over 12 weeks if alternative treatments are not effective, and there have been at least 2 unsuccessful discontinuation attempts of the antipsychotics.10

Antipsychotic medication use for BPSD is highest among people in residential aged care facilities (RACFs; long-term care facilities). A meta-analysis of international studies estimated antipsychotic medication use among people with dementia to be 12.3% in the community and 37.5% in RACFs worldwide.16 Despite the emphasis on the duration of use in clinical guidelines, little is known about the duration of antipsychotic medication use; only a handful of studies measured proxies for the duration. In Norway, antipsychotic medication use assessed at multiple time points suggested that duration was likely much longer than 90 days.17 The few studies that have attempted to evaluate the duration of antipsychotic medication use have used prescribing or dispensing data, which are subject to measurement error (if the medication taken was unknown).18–20 Consistent with other countries, Australia has reported a high prevalence of antipsychotic medication use among people with dementia. Among community-dwelling people with dementia, 11% were prescribed antipsychotic medication between 1998 and 2000.21 In RACFs, estimates of the prevalence of use have been in the order of 20% of all residents22,23 and 44% of residents taking antidementia medication.18

Measuring the duration of antipsychotic medication use and factors associated with potentially inappropriate use is critical to determining whether clinical guidelines, interventions, and medication coverage policies are sufficiently limiting the use of potentially harmful medications. This evidence is necessary for tailoring policies to reduce excessive antipsychotic medication use. The aims of this study were to estimate the: prevalence, duration, and time to initiation of antipsychotic medication use among residents with dementia. We also aimed to assess the relationships between antipsychotic use and a range of sociodemographic, health status, and facility factors. We hypothesized that the prevalence of antipsychotic use would be similar to prior studies conducted in Australia, and that duration and time to initiation of antipsychotic medication use would vary by age, sex, and psychiatric comorbidities.



We conducted a secondary data analysis using electronic health record (EHR) data from a large not-for-profit residential aged care provider with facilities across New South Wales and the Australian Capital Territory. Medications administered to residents in 68 facilities during the period 2014 to 2017 were analyzed. We included permanent residents (respite care excluded) aged 65 years or above. Residents with diagnosed dementia were included for the prevalence estimates. For the duration analysis, the sample is limited to residents with dementia who used antipsychotic medication. For the analysis of time to initiation, the sample is further limited to those who were not using antipsychotic medication at admission to the facility. Our data represent incident medication administration events for a dynamic unbalanced longitudinal cohort of older people with dementia; residents entered and permanently left the sample at different times and tended to have long stays. In Australia, medications are prescribed by physicians who are external to the RACFs and are administered by the RACF staff.


Dementia and other chronic conditions were identified using Aged Care Funding Instrument (ACFI) data and by searching for relevant text strings in 2 free-text fields containing resident conditions in the EHR. The ACFI is an assessment of health and functional status used in Australia to determine the level of care and Federal Government funding for people receiving aged care services.24 We included the following common chronic comorbidities, and psychiatric conditions that may be associated with higher antipsychotic medication (including off-label use): (1) psychoses (includes paranoid states), schizophrenia or bipolar disorder (PSBD),25 (2) depression, mood, and affective disorders (depression hereafter),26 and (3) anxiety and stress-related disorders (anxiety hereafter).27 To determine these groupings, we consulted facility nurses and also considered the categories used to report these conditions in the ACFI. Because of substantial overlap between PSBD, depression, and anxiety, we classified residents by PSBD, depression without PSBD, and anxiety without PSBD. Sociodemographic covariates included in the analysis were age, sex, marital status, country of origin, and whether English was the resident’s primary language. We also included an indicator for high care needs and length of stay (LOS) in years. Facility-level covariates included facility remoteness28 (ranked from most to least urban; these relative remoteness categories are as follows: major cities, inner regional, outer regional, remote, and very remote—the last 2 categories are not represented in our data), staff per bed (a proxy for staff per resident), and social deprivation score (IRSAD,29 an area-level scale of income and proportion of skilled workers) quintile.

Medication Outcomes

Medication administration data provided a longitudinal record of every medication dose administered to residents. We identified antipsychotic medication administrations. Antipsychotic medications (Table S1, Supplemental Digital Content 1, were classified as typical (first generation) or atypical (second generation).30 Although risperidone is the only antipsychotic subsidized for short-term use in BPSD in Australia, we included all other short-acting antipsychotics to capture any potential off-label use. Residents who used long-acting injectable preparations of antipsychotic medications (n=17) were excluded, because the use of these medications is variable, making episodes of use difficult to define in a meaningful way, and the small number of them prohibited a subanalysis. Examples of the 3 medication duration definitions used are presented in Figure S1 (Supplemental Digital Content 2, For the primary duration analysis, we defined an antipsychotic medication administration episode as continuous days of administration, allowing 1-day gaps and allowing time away from the facility (eg, hospitalizations during which medication administrations were not observed). The counting of days included in the episode resumed once the resident reentered the facility. Allowing 1-day gaps in use assumed that a single day lapse in use followed by resumption was not a true endpoint and was likely due to other events (eg, residents not taking their medications due to other reasons such as dose tapering31). Time away from the facility was identified between the time when all medications stopped and time when medication administrations resumed; as most residents are on multiple medications, a lapse in all medications typically indicates that the resident is away. Right censoring of medication duration was not common (<6% of episodes), and it occurred when medication use continued until the following occurred: (1) resident died, (2) resident was discharged or transferred to a facility outside of the provider network, and (3) observation time ended. Episodes were left censored when the true start point was not observed because the resident was already on antipsychotic medication at admission to the facility. Censoring was controlled for, as right-censored episodes were likely to create a downward bias in days of use, whereas left censoring had the potential to also generate downward bias, but may also be associated with longer use, as long-term users may be more likely to be admitted while already on antipsychotics.

Statistical Analysis

Prevalence of antipsychotic medication use among residents with dementia was calculated annually and for 2014 to 2017. For residents who used antipsychotic medication, we calculated the number of days that antipsychotic medications were administered in each episode, and number of days between episodes (for residents with >1 episode). To generate descriptive statistics, we created a person-level (summarized) data set with the total number of antipsychotic medication episodes, total number of episodes that exceeded 90 days, and an indicator for having at least 1 episode that exceeded 90 days.

To evaluate the importance of resident and facility characteristics on antipsychotic medication duration, we fit a multilevel mixed effects generalized linear (gamma) model with a random intercept for resident and a random intercept for facility. As predictors, we included resident sociodemographics, psychiatric comorbidities (PSBD, depression, and anxiety), health status, LOS, and facility characteristics, indicators for left and right censoring, and an interaction term for left×right (ie, double) censoring. Quadratic terms for age and LOS were included to accommodate potential curvilinear effects. We included interactions between sex, psychiatric comorbidities, and age (including age2) to allow psychiatric comorbidity effects to vary by sex and age, and also interactions between psychiatric comorbidities and censoring; interactions were retained at a significance level of ≤20,32 or if they behaved as confounders. In the presence of a higher order interaction, lower order terms were retained. We estimated marginal effects that can be interpreted as the number of days of medication use associated with a change in a given covariate from the base/reference level.

To evaluate resident and facility factors associated with initiating antipsychotic medications, we calculated the time to initiation as the number of days from the facility entry date to the starting date for their first antipsychotic medication administration (excluding left-censored episodes, ie, when a resident entered a facility already on an antipsychotic). We used a regression approach similar to that used in the duration analysis, but only included a random intercept for facility and used a Poisson distribution. Model-based estimated days to initiation by sex, age, and psychiatric comorbidities were produced.

Sensitivity Analysis

We conducted 2 sensitivity analyses to evaluate how sensitive our estimates were to the medication episode definition used. In the first sensitivity analysis, we considered a naive definition of continuous use (no gaps allowed within episodes). We postulated that the naive definition would be a reasonable, albeit simple, approach and would indicate whether the more complex algorithms of episode duration were necessary in this work and future studies. In the second sensitivity analysis, we allowed 2-day gaps to occur within administration episodes and accounted for time away from the facility. The descriptive statistics and regression models described in the main text were repeated for each of the sensitivity analyses using the alternative episode definitions (naive and 2-day gaps allowed).

A type I error rate of 0.05 was used. Analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC) and Stata 15 (Stata Corp., College Station, TX). This research was approved by our Institution’s Human Research Ethics committee.


Across 68 facilities, there were 10,480 permanent residents, and 5825 had dementia. The period prevalence of antipsychotic medication use among residents with dementia was 37.8% during the years 2014 to 2017; annual prevalence ranged from 27.6% to 32.6% (Table S2, Supplemental Digital Content 1, Duration analyses included 1730 residents with dementia who used an antipsychotic medication at any time from January 1, 2014 to September 28, 2017 (Fig. S2, Supplemental Digital Content 2, sample size and exclusions), of which 1316 (76.1%) had left-censored antipsychotic use. A majority (65.3%) were female individuals, and 76.3% had Alzheimer-type dementia (Table 1). The median [interquartile range (IQR)] age at first observed antipsychotic medication use was 87 (81 to 91) for female individuals and 83 (77 to 88) for male individuals. Frequencies of specific antipsychotic medications are presented in Table S1 (Supplemental Digital Content 1,; risperidone was most commonly used). Among residents with dementia who used antipsychotic medication, 298 (17.2%) had comorbid PSBD (detail on case identification reported in Fig. S3, Supplemental Digital Content 2, and Table S3, Supplemental Digital Content 1,, 962 (55.6%) had comorbid depression, and 477 (27.6%) had comorbid anxiety.

Sample Characteristics: Residents With Dementia Who Used an Antipsychotic Medication (N=1730)


A total of 9242 antipsychotic medication administration episodes (primary definition: 1-day gaps and time away allowed) were identified (Table S4, Supplemental Digital Content 1, A total of 65.0% of residents with dementia who used antipsychotic medication had at least 1 episode that exceeded 90 days. Those with comorbid PSBD had a higher proportion of episodes exceeding 90 days (77.2%) compared with residents without any of the 3 psychiatric comorbidities (59.6%). Figure 1 presents plots of model-based estimates of mean duration by age, sex, and psychiatric comorbidities for the primary episode definition (Fig. 1A), and sensitivity analyses (naive definition, Fig. 1B; 2-day gaps allowed, Fig. 1C). In, Fig. 1A it can be seen that, although variation by age and psychiatric comorbidities decreases with age, the groups have overlapping confidence intervals (CIs). At median ages, mean (95% CI) duration overall was 212.74 (170.24, 255.25) days for female individuals and 216.10 (165.31, 266.89) days for male individuals. Stratified by psychiatric comorbidities, the mean (95% CI) duration at median ages were as follows: 197.08 (149.18, 244.98) days for female individuals without psychiatric comorbidities, 258.40 (177.17, 339.63) days for female individuals with PSBD, 216.24 (157.94, 274.54) days for male individuals without psychiatric comorbidities and 180.77 (103.50, 258.04) days for male individuals with PSBD (Table S5, Supplemental Digital Content 1, Regression analysis identified several comorbidities associated with shorter antipsychotic medication duration including cerebrovascular disease [31.89 (95% CI: 54.98, 8.80) d shorter] and liver disease [82.33 (95% CI:133.59, 31.08) d shorter] (Table 2). Atypical antipsychotics were used longer than typical antipsychotics by 108.82 days (95% CI: 89.91, 127.74). Age had a complex relationship with medication duration; there was more variation in mean duration at younger ages (70) compared with older ages (95), but, depending on sex and psychiatric comorbidity, the peaks in duration occurred at the younger or middle part of the age distribution—in other words, the intercepts for each group differed at younger ages, and the slopes differed across ages (Fig. 1). There was a 3-way interaction between age2, sex, and PSBD, wherein younger male individuals and female individuals with PSBD had the longest durations of use, and duration decreased with age, but this decrease was more pronounced for male individuals with PSBD compared with female individuals with PSBD. Interactions for age and age2, with depression, and age and age2, with anxiety, were included per our confounding criteria. LOS was associated with longer duration of use by 23.94 (95% CI: 14.74, 33.14) days per year, but these effects are smaller as LOS increases (small negative estimate for LOS2).

Estimated mean and 95% confidence intervals for duration (days) of antipsychotic medication use in residents with dementia by age group and comorbid psychological and behavioral conditions including psychoses (including paranoid states), schizophrenia, or bipolar disorder (PSBD), depression without PSBD, anxiety without PSBD, and none of the 3 conditions. The dashed horizontal line represents 90 days. A, Episodes using the primary definition allowing 1-day gaps and time away within episodes. B, The first sensitivity analysis, which uses a naive definition of days of continuous use, with no gaps and no time away allowed. C, The second sensitivity analysis, which allows 2-day gaps and time away.
Multilevel Regression: Marginal Effects for the Outcome of Antipsychotic Medication Duration

Time to Initiation

Time to medication initiation analysis included 414 residents. Estimated days to initiation of antipsychotic medication by psychiatric comorbidities and sex are plotted in Figure 2. Age, sex, and psychiatric comorbidities had complex relationships; in Figure 2, the slopes, curvature of the slopes, and the intercepts differed by age depending on sex and psychiatric comorbidities. Regression analyses found that most resident and facility factors were associated with time to initiation of medication (Table 3); age (and age2), sex, and psychiatric comorbidities had 3-way interactions. Medication initiation occurred earlier for male individuals compared with female individuals—on average after 308.4 days (95% CI: 252.6, 364.2) after admission for female individuals, and after 173.2 days (95% CI: 141.8, 204.6) after admission for male individuals (unadjusted).

Estimated mean and 95% confidence intervals for time to antipsychotic medication initiation after entry into aged care facility by sex, psychiatric comorbidities, and age. The sample is limited to residents who entered facilities after medication reporting began and did not have antipsychotic medications administered on their first day in the facility, or did not have antipsychotic medications administered on their first day of having any medications administered. PSBD indicates psychoses (includes paranoid states), schizophrenia or bipolar disorder.
Marginal Effects: Time to Initiation of Antipsychotic Medication

Sensitivity Analyses

The proportion of episodes exceeding 90 days was sensitive to the episode definition used. Applying the naive definition of continuous days of use (no gaps allowed), the proportion of episodes with durations exceeding 90 days decreased to 51.7%, and the total number of episodes increased to 23,977. Applying the 2-day gap definition, the proportion of residents with at least 1 episode exceeding 90 days increased to 65.7%, and the number of total episodes decreased to 6342. This trend was due to the time between episodes. Using the primary definition allowing for 1-day gaps, the median time between episodes was 3 (IQR: 2 to 6) days, which increased to 5 (IQR: 3 to 11) days when allowing for 2-day gaps. For the outcome of duration, sensitivity analyses found that effect sizes and significance of some covariates were sensitive to the episode definition used, and the results were more sensitive to the naive definition than to the definition allowing 2-day gaps (Table S6, Supplemental Digital Content 1, and Figs. 1B, C).


This study provides a more detailed analysis of usage trends building on existing international studies that have demonstrated a high prevalence of antipsychotic medication use by people with dementia. Among residents with dementia who use antipsychotic medications, use lasts roughly twice as long as the recommended maximum, and most residents are on these medications when they enter the facility. Although psychiatric comorbidities explained long-term antipsychotic use to some degree, duration >3 months was common among people without these conditions, and, across ages duration of use was similar for people with dementia alone compared with people with dementia and other psychiatric comorbidities. Our results are broadly consistent with the few longitudinal studies of duration conducted in Australia, the United States, Finland, and Sweden.18–20,33 Our findings add to the concerns with regard to the risks associated with the use of antipsychotic medications in people with dementia.

At younger ages, duration of use was shortest for male individuals and female individuals without psychiatric comorbidities, but not at older ages, wherein duration differences became less pronounced across psychiatric comorbidity and sex-specific groups. Other factors besides residents’ health status influenced the duration of use and time to initiation, including sociodemographic factors and facility-related factors. General practitioners initiate and manage treatment on the basis of resident symptoms, but their preferences and family preferences may also influence prescribing patterns.

Risperidone is the only antipsychotic subsidized by the Australian Medicare Pharmaceutical Benefits Scheme for management of BPSD (specifically aggression and psychosis) and was the most commonly used. The current criteria for prescribing risperidone under the Pharmaceutical Benefits Scheme funding system require that a patient has Alzheimer-type dementia, has been unresponsive to nonpharmacological methods of BPSD management, and that treatment must be limited to a maximum duration of 12 weeks (implemented in 2015).34 Given our findings, these restrictions may not be sufficient to limit the use of antipsychotic medications in this population.

Our estimates were sensitive to the definition of medication administration episode used; the median time between episodes of continuous use was only 1 day when we applied the naive definition of continuous medication use. Creating episode definitions that allowed for 1- and 2-day gaps in use (and correcting for time away from the facility) reduced the total number of episodes and greatly increased the episode length. Episode definition is clearly an important consideration for other studies measuring the duration of use, and our results suggest that using a naive definition overlooks important factors impacting measurement, which can generate substantial underestimation of duration.

Our study had a number of strengths including the use of a large longitudinal sample of people with dementia living in residential care, and use of medication administration data that reduced measurement error relative to prescribing and filling data. Sensitivity analyses strengthened the study by demonstrating the importance of medication episode definitions. Our findings may be generalizable to other RACFs in Australia, as the demographic profile of our sample is generally representative of the Australian residential aged care population.35 Antipsychotic medication use period prevalence in our sample was 37.9%—nearly identical to that estimated among facility-dwelling people with dementia worldwide (37.5%).16 This raises the possibility that the day to day use of these medications is similar in other countries, or that, on average, other facilities are (1) the same with respect to duration and number of episodes, (2) have longer duration but a lower number of episodes, or (3) have shorter duration but more episodes.

Our results should be interpreted with consideration of the limitations of the study design. We relied on EHR data for identifying comorbidities, with underidentification being a possible issue. Our estimates likely represent an underestimation of the true duration, considering that (1) we did not count the 1-day gaps in administration as a day of use, (2) 2-day gaps may not represent a cessation of antipsychotic medication, and (3) most residents were using antipsychotic medication before admission. Cumulative use of antipsychotic medications will be longer, as residents often had >1 episode of use. However, it is possible that some residents had valid clinical reasons for long durations of use, as guidelines allow long-term use for those who have not responded to other treatment and have had unsuccessful withdrawal attempts.10 Multiple episodes with small gaps in some residents may indicate withdrawal attempts, which may be consistent with guidelines. Results should be interpreted with these possibilities in mind.

Our findings suggest that recommendations for short-term use of antipsychotic medications for BPSD are not being followed by prescribers, despite restrictions on subsidization. Medication management in RACFs is complex, involving prescribers, aged care facility staff, the residents, and their families.36 Understanding the factors driving antipsychotic medication use in BPSD is critical to creating targeted interventions to reduce inappropriate medication prescribing and reducing harm. Interventions to reduce antipsychotic medication use have produced a modest improvement in Australia and internationally. Further strategies that reach prescribers and support RACF staff are needed. Using EHR data may be a useful approach to tracking medication use patterns and providing feedback to prescribers as a strategy to reduce antipsychotic medication use.


1. Alzheimer’s Disease International. World Alzheimer Report 2015: The Global Impact of Dementia. London, UK: Alzheimer’s Disease International; 2015.
2. van der Linde RM, Dening T, Stephan BCM, et al. Longitudinal course of behavioural and psychological symptoms of dementia: systematic review. Br J Psychiatry. 2016;209:366–377.
3. Ballard C, Waite J. The effectiveness of atypical antipsychotics for the treatment of aggression and psychosis in Alzheimer’s disease. Cochrane Database Syst Rev. 2006;1:CD003476.
4. Liperoti R, Pedone C, Corsonello A. Antipsychotics for the treatment of behavioral and psychological symptoms of dementia (BPSD). Curr Neuropharmacol. 2008;6:117–124.
5. Gareri P, De Fazio P, Manfredi VGL, et al. Use and safety of antipsychotics in behavioral disorders in elderly people with dementia. J Clin Psychopharmacol. 2014;34:109–123.
6. Muench J, Hamer AM. Adverse effects of antipsychotic medications. Am Fam Physician. 2010;81:617–622.
7. DynaMed. Antipsychotic medication overview. 2017. Available at: Accessed February 14, 2018.
8. Tampi RR, Tampi DJ, Balachandran S, et al. Antipsychotic use in dementia: a systematic review of benefits and risks from meta-analyses. Ther Adv Chronic Dis. 2016;7:229–245.
9. Schneider LS, Tariot PN, Dagerman KS, et al. Effectiveness of atypical antipsychotic drugs in patients with Alzheimer’s disease. N Engl J Med. 2006;355:1525–1538.
10. Zuidema SU, Johansson A, Selbaek G, et al. A consensus guideline for antipsychotic drug use for dementia in care homes. Bridging the gap between scientific evidence and clinical practice. Int Psychogeriatr. 2015;27:1849–1859.
11. Choosing Widely Australia. Psychiatry—antipsychotics to treat dementia symptoms. 2015. Available at: Accessed February 14, 2018.
12. Choosing Widely Australia. Medicinewise NPS; 2016. Available at: Accessed August 7, 2018.
13. Choosing Widely Australia. The Society of Hospital Pharmacists of Australia: treatments pharmacists and consumers should question; 2016. Available at: Accessed March 13, 2018.
14. National Institute for Health and Care Excellence. Dementia: supporting people with dementia and their carers in health and social care. Guidance and guidelines; 2016. Available at: Accessed May 22, 2018.
15. Guideline Adaptation Committee. Clinical Practice Guidelines and Principles of Care for People with Dementia. Sydney, NSW: NHMRC Partnership Centre for Dealing with Cognitive and Related Functional Decline in Older People; 2016.
16. Kirkham J, Sherman C, Velkers C, et al. Antipsychotic use in dementia. Can J Psychiatry. 2017;62:170–181.
17. Helvik A-S, Šaltytė Benth J, Wu B, et al. Persistent use of psychotropic drugs in nursing home residents in Norway. BMC Geriatr. 2017;17:52.
18. Shin H-Y, Gadzhanova S, Roughead EE, et al. The use of antipsychotics among people treated with medications for dementia in residential aged care facilities. Int Psychogeriatr. 2016;28:977–982.
19. Kim HM, Chiang C, Weintraub D, et al. Treatment changes among older patients with dementia treated with antipsychotics. Int J Geriatr Psychiatry. 2015;30:1238–1249.
20. Koponen M, Taipale H, Tanskanen A, et al. Long-term use of antipsychotics among community-dwelling persons with Alzheimer’s disease: a nationwide register-based study. Eur Neuropsychopharmacol. 2015;25:1706–1713.
21. Bridges-Webb C. Dementia care in general practice. What can the BEACH survey tell us? Aust Fam Physician. 2002;31:381–383.
22. Westbury JL, Jackson S, Peterson GM. Psycholeptic use in aged care homes in Tasmania, Australia. J Clin Pharm Ther. 2010;35:189–193.
23. Westbury JL, Gee P, Ling T, et al. RedUSe: reducing antipsychotic and benzodiazepine prescribing in residential aged care facilities. Med J Aust. 2018;208:398–403.
24. Australian Government Department of Human Services. Aged Care Funding Instrument; 2017. Available at: Accessed May 23, 2018.
25. Carnahan RM, Letuchy EM. Bipolar disorder in nursing homes: impact on antipsychotic use, diagnosis patterns, and new diagnoses in people with dementia. Am J Geriatr Psychiatry. 2018;26:2–10.
26. Gerhard T, Stroup TS, Correll CU, et al. Antipsychotic medication treatment patterns in adult depression. J Clin Psychiatry. 2018. Doi: 10.4088/JCP.16m10971.
27. Weber SR, Wehr AM, Duchemin A-M. Prevalence of antipsychotic prescriptions among patients with anxiety disorders treated in inpatient and outpatient psychiatric settings. J Affect Disord. 2016;191:292–299.
28. Australian Bureau of Statistics. Australian Statistical Geography Standard (ASGS): volume 4—significant urban areas, urban centres and localities, section of state; 2017. Available at: Accessed May 10, 2018.
29. Australian Bureau of Statistics. Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia; 2018. Available at: Accessed May 10, 2018.
30. Australian Medicines Handbook. 2019. Available at: Accessed February 19, 2018.
31. Ipswitch and East Suffolk Clinical Commisioning Group—National Health Service. Guidance for antipsychotic use and withdrawal in dementia; 2014. Available at: Accessed May 11, 2018.
32. Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989;129:125–137.
33. Gustafsson M, Karlsson S, Lövheim H. Inappropriate long-term use of antipsychotic drugs is common among people with dementia living in specialized care units. BMC Pharmacol Toxicol. 2013;14:10.
34. Australian Government Department of Health. Pharmaceutical Benefits Scheme (PBS). 2015. Available at: Accessed February 19, 2018.
35. Pont LG, Raban MZ, Jorgensen ML, et al. Leveraging new information technology to monitor medicine use in 71 residential aged care facilities: variation in polypharmacy and antipsychotic use. Int J Qual Health Care. 2018;30:810–816.
36. Tariq A, Georgiou A, Westbrook J. Medication errors in residential aged care facilities: a distributed cognition analysis of the information exchange process. Int J Med Inform. 2013;82:299–312.

BPSD; dementia; antipsychotic medication

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved