Potentially Inappropriate Medication Use in Community-Dwelling Older Adults Living with Dementia

Abstract

Background:

The Beers Criteria identifies potentially inappropriate medications (PIMs) that should be avoided in older adults living with dementia.

Objective:

The aim of this study was to provide estimates of the prevalence and persistence of PIM use among community-dwelling older adults living with dementia in 2011-2017.

Methods:

Medicare claims data were used to create an analytic dataset spanning from 2011 to 2017. The analysis included community-dwelling Medicare fee-for-service beneficiaries aged 65 and older who were enrolled in Medicare Part D plans, had diagnosis for dementia, and were alive for at least one calendar year. Dementia status was determined using Medicare Chronic Conditions Date Warehouse (CCW) Chronic Condition categories and Charlson Comorbidity Index. PIM use was defined as 2 or more prescription fills with at least 90 days of total days-supply in a calendar year. Descriptive statistics were used to report the prevalence and persistence of PIM use.

Results:

Of 1.6 million person-year observations included in the sample, 32.7% used one or more PIMs during a calendar year in 2011-2017. Breakdown by drug classes showed that 14.9% of the sample used anticholinergics, 14.0% used benzodiazepines, and 11.0% used antipsychotics. Conditional on any use, mean annual days-supply for all PIMs was 270.6 days (SD = 102.7). The mean annual days-supply for antipsychotic use was 302.7 days (SD = 131.2).

Conclusion:

Significant proportion of community-dwelling older adults with dementia used one or more PIMs, often for extended periods of time. The antipsychotic use in the community-dwelling older adults with dementia remains as a significant problem.

Keywords

Alzheimer’s disease anticholinergics antipsychotics benzodiazepines community-dwelling dementia potentially inappropriate medication

INTRODUCTION

Dementia is a syndrome characterized by symptoms of impaired cognition, such as difficulties with memory, language, problem solving and other cognitive skills [1]. As the disease progresses, the deterioration of cognitive function gradually affects a person’s ability to perform everyday activities leading to loss of independence [1]. It is estimated that about 11.3% of age 65 and older living in the United States (US) have a diagnosis for dementia, while 22.7% live with some form of mild cognitive impairment [2]. In 2022, the total national cost of caring for individuals with dementia is estimated to be $321 billion [1]. Individuals with dementia will ultimately progress into severe loss of function requiring 24-hour assistance, necessitating institutionalization in a nursing home (NH).

The Beers Criteria, which was adopted by the American Geriatric Society in 2011, is regarded as the gold standard for assessing the appropriateness of medication use in older people [3, 4]. Potentially inappropriate medications (PIMs) are generally defined as medications that: 1) have no clear evidence-based indication, 2) carry a substantially higher risk of adverse effects than benefits, or 3) are not cost effective [5, 6]. Recently updated Beers Criteria (2019) identify specific medications that should be avoided in older adults with dementia or cognitive impairment, which largely fall into categories of: 1) drugs with strong anticholinergic properties, 2) benzodiazepines and benzodiazepine receptor agonist hypnotics, and 3) antipsychotics (Supplementary Table 1). [7] Studies have shown that use of medications identified as potentially inappropriate for people with dementia (hereafter referred to as “PIM use”) is a significant risk factor for NH placement among community-dwelling older adults living with dementia [8, 9]. This is mainly due to the increased risk for accidental injurious falls and development of behavioral symptoms, which can lead to premature institutionalization [8 –12]. For antipsychotics, the recommendation is to avoid use even in a short-term basis except in cases where nonpharmacological options have failed and the patient is threatening substantial harm to self or others [7]. Numerous studies reported about the dangers of using antipsychotics for treatment of behavioral symptoms related to dementia, demonstrating strong links between their uses and adverse events such as cognitive disturbances, cardio- and cerebrovascular events and increased mortality [15 –18]. Clinical guidelines instead recommend nonpharmacological interventions and environmental measures for reducing psychosis and agitation [13, 19].

Many research studies have reported high rates of PIM uses in nursing home settings, which led to a national response to reduce the use of PIMs and antipsychotics among institutionalized older adults. These efforts contributed to a significant reduction in antipsychotic use; the Centers for Medicare & Medicaid Services (CMS) reported that the percentage of long-stay NH residents who received an antipsychotic medication decreased from 23.9% in 2011 to 14.5% in 2020 [20]. While this signifies that the quality of care is moving in a positive direction for older adults living with dementia, the evidence is limited to NH settings. There is a paucity of studies examining the prevalence of PIM use in community settings and whether prescribing changes in NH settings have affected use in community settings. A few existing studies have estimated 15% to 27% of community-dwelling older adults with dementia were taking at least one PIM on a regular basis [14 , 21–23]. However, these estimates were either limited by recall bias due to the nature of data sources or had limited sample size. Additionally, those reports used data prior to the national efforts to reduce antipsychotic medication use in NHs, therefore, precluding any potential spillover effects of the national response in community settings.

Therefore, the current study’s objective was to provide estimates of the prevalence and persistence of PIM use among community-dwelling older adults living with dementia using fee-for-service (FFS) Medicare sample from 2011 to 2017.

MATERIALS AND METHODS

Study design and data

This was a retrospective cohort study using 20% Medicare claims data from January 1, 2011, to December 31, 2017. Beneficiary characteristics were derived from Medicare enrollment and claims data and linked to the Rural-Urban Commuting Area (RUCA) to obtain rurality information. RUCA is a publicly available dataset released by US Department of Agriculture classifying US census tracts and provides information on the level of rurality at ZIP code area-level. Areas with RUCA codes 7-10 (small towns and rural regions) were classified as rural and linked with study sample using the reported ZIP code data. The final analytic dataset was created by linking beneficiary characteristics, prescription claims data, and inpatient and outpatient visit data at person-year level.

This study was reviewed and deemed exempt from requiring ethics approval by the University of Southern California Institutional Review Board (IRB) because the risk to participants was considered lower than minimal risk.

Study sample

The study sample included community-dwelling adults aged 65 and older who were alive and enrolled in FFS and Medicare Part D plan for 12 months in a calendar year between 2011 and 2017. Long-Term Care Minimum Data Set 3.0 (MDS 3.0) was used to create an indicator for nursing home stays at person-year level. MDS 3.0 is a standardized assessment tool that is required by CMS for all Medicare- and Medicaid-certified nursing homes to evaluate all residents upon admission and discharge and at least on a quarterly on-going basis. Because MDS can be linked with Medicare claims data, beneficiaries who stayed in a nursing home at any point during a calendar year can be identified and excluded from the sample for that year. Those who entered a nursing home but were discharged and dwelled in community for an entire calendar year were included in the sample for such years. Figure 1 shows a flow diagram describing the sample selection process at person-year level. Approximately 61% of Medicare enrollees were enrolled in a Part D plan over the study period (2011-2017). Of those beneficiaries, 40% were community-dwellers, aged 65 years or older, and enrolled in FFS plan (total 19,588,187 person-year observations). Among those enrollees, 8.2% (1,605,256 person-year observations) met the criteria for having dementia and were included in the sample.

Fig. 1

Flow diagram of sample selection*. *Numbers shown are at person-year level.

Presence of dementia was determined based on indication of disease in the Chronic Conditions Data Warehouse (CCW) Chronic Conditions algorithms for Alzheimer’s Disease, Related Disorders, or Senile Dementia (Supplementary Table 2) or Charlson Comorbidity Index algorithms for dementia [24 –27]. The primary and secondary ICD-9 or ICD-10 diagnosis codes from inpatient and outpatient records were used to create a Charlson Comorbidity Index score for each beneficiary. The inpatient and outpatient records were also searched to identify beneficiaries who had other comorbid conditions that may have warranted use of antipsychotics for treatment. Beneficiaries with diagnosis of schizophrenia, Tourette’s syndrome, or Huntington’s disease were flagged so that their use of antipsychotics were not counted as potentiallyinappropriate.

Outcome measures and data analysis

The Medicare Part D Event file was used to identify prescription drug use that are considered potentially inappropriate for older adults with dementia based on the definition used by the 2019 American Geriatrics Society (AGS) Beers Criteria of Potentially Inappropriate Medications and Classes to Avoid in Older Adults with Dementia [7]. While there are broader categories of drugs considered potentially inappropriate in general older adults, Beers Criteria identifies PIMs that should particularly be avoided in people with dementia consisting of three categories: drugs with strong anticholinergic properties, benzodiazepines or hypnotics (nonbenzodiazepine and benzodiazepine receptor agonist), and antipsychotics (Supplementary Table 2). In previous versions of Beers Criteria published in 2012 and 2015, an additional drug class (H2-receptor antagonists) was included in PIMs to be avoided in dementia, but it was removed in the most recent update (2019) while no additional drug class was added [28, 29]. This study used the definition of PIMs in dementia from the 2019 update so that the result of the study may be meaningful and applicable to the current practice and policy considerations. The prescription drug claims records were searched for the generic names of PIMs to capture all drugs containing the drug ingredient regardless of the variations in respective brand names. If there was a combination of two or more drug ingredients, each ingredient was considered as a distinct drug and counted accordingly. Some PIMs that would not be considered potentially inappropriate in certain dosage forms were identified and unflagged (e.g., topical diphenhydramine, scopolamine in ophthalmic suspension, etc.). All identified PIMs were classified into their respective drug classes and merged with analytic dataset containing beneficiary information by calendar year to create person-year level data across 2011-2017. Outcome measures were the prevalence of any PIM use and the persistence of use among those with one or more PIMs as measured by the total days-supply at the person-year level. To exclude temporary PIM use and rule out false positive records from being included in the analysis, the prevalence measure only included PIM use with: 1) the total annual days-supply of 90 days or more within same drug ingredient regardless of dosage change and 2) two or more prescription fills within a year. The persistence measure was described as mean of overall PIM use and by each drug class (conditional on any use of drugs in each class). A sensitivity analysis estimated the outcome measures using a less stringent definition of PIM use—any number of prescription fills within a year with at least 90 days of total days-supply—to test the robustness of the findings.

Descriptive statistics were used to describe the sample characteristics and to report the prevalence and persistence of PIM use in the sample. Continuous variables were described as aggregate means with standard deviation and categorical variables were shown as percentages. Student’s t-test was used to compare continuous data across groups and Chi-square test was used for binary data. All statistical analyses were performed using Stata version 16.1.

RESULTS

Sample characteristics

The sample contained on average 229,322 Medicare beneficiaries per year during the study period. Demographic characteristics of beneficiaries included in the sample based on 2017 data are shown in Table 1. On average, sample beneficiaries were 81.4 years of age (SD = 8.1) and 66.6% female. Majority of the sample were white (81.0%) and a small proportion were residing in rural areas (12.6%). Twenty-nine percent had Medicare-Medicaid dual eligibility and 2.9% were enrolled in a low-income subsidy program. The sample characteristics of beneficiaries with and without 1 or more PIM use were compared (Table 1). Those with PIM use were more likely to be female (70.1% versus 61.9%), white (84.0% versus 79.7%), and dually eligible (34.9% versus 26.3%).

Table 1

Sample characteristics by PIM use status (represented by 2017 data)

		All (n = 259,291)	1 or more PIM use (n = 82,006)	No PIM use (n = 177,285)	p
Age, Mean (SD)		81.4 (8.1)	80.3 (8.2)	81.9 (8.0)	<0.01
Female, %		64.5%	70.1 %	61.9%	<0.01
Race/ethnicity, %	White	81.0%	84.0%	79.7%	<0.01
	Black	9.1%	7.0%	10.0%
	Asian	3.5%	2.6%	3.9%
	Hispanic	3.4%	3.7%	3.2%
	Other/unknown	3.0%	2.7%	3.2%
Living in rural areas, %		12.6%	12.9%	12.5%	<0.01
CCI, %	0-2	71.8%	69.8%	72.8%	<0.01
	3-5	19.4%	20.9%	18.7%
	>6	8.8%	9.3%	8.5%
Dual eligibility, %		29.0%	34.9%	26.3%	<0.01
LIS enrollment, %		2.9%	3.1%	2.9%	<0.01

CCI, Charlson Comorbidity Index; LIS, low-income subsidy; SD, standard deviation.

Patterns of PIM use: prevalence and persistence

The prevalence and persistence of PIM use are presented in Table 2. At the person-year level, 32.7% of the sample used one or more medications considered potentially inappropriate for persons with dementia by Beers Criteria. Of those who used one or more PIMs, 21.1% were taking two PIMs and 6.6% were taking three or more. The breakdown by drug class showed that 14.9% of beneficiaries used at least one drug with strong anticholinergic properties and 14.0% used benzodiazepines or hypnotics. Antipsychotics were used by 11.3% of those in the sample. The mean annual days-supply of PIM use are shown in the second panel of Table 2. Among those who used one or more PIMs, the mean duration of use was 270.6 (SD = 102.7) days per year. At the drug class-level, the mean annual days-supply was 266.8 days (SD = 99.4) for strong anticholinergic medications and 249.5 days (SD = 102.0) for benzodiazepines. The greatest annual days-supply was observed among antipsychotics users, which averaged 302.7 days (SD = 131.2) per year.

Table 2

Prevalence and persistence of PIM use, unadjusted

(1) Prevalence of Any Use
	n=1,605,256
Any PIM	n (%)	525,436 (32.7)
Number of PIMs used, if any	n (%)
	1	379,672 (72.3)
	2	110,901 (21.1)
	3 or more	34,863 (6.6)
Any anticholinergics	n (%)	238,894 (14.9)
Any benzodiazepines/hypnotics	n (%)	224,768 (14.0)
Any antipsychotics	n (%)	181,432 (11.3)
(2) Persistence as in Days-Supply Per Year
Any PIM		n=525,436
	Mean (SD), days	270.6 (102.7)
Anticholinergics		n=238,894
	Mean (SD), days	266.8 (99.4)
Benzodiazepines/Hypnotics		n=224,768
	Mean (SD), days	249.5 (102.0)
Antipsychotics		n=181,432
	Mean (SD), days	302.7 (131.2)

PIM, potentially inappropriate medication; SD, standard deviation; min, minimum; max, maximum.

Estimates of PIM use by year suggested that the use of anticholinergics had been on a decreasing trend (from 17.2% to 13.4%) while antipsychotics use had been relatively steady (from 11.4% to 10.7%) during the study period (Fig. 2). There was a steep increase in the prevalence of benzodiazepine use in 2013 (from 6.6% to 17.7%), which declined over the next four years to 14.9%. The annual days-supply for anticholinergics and benzodiazepines did not show any apparent trends over the study years. However, the annual days-supply for antipsychotics use steadily increased from 293.3 days to 310.5 days in 2011-2017.

Fig. 2

Prevalence and mean annual days-supply of antipsychotic medication use over time (2011 to 2017). PIM, potentially inappropriate medication.

Sensitivity analysis

A sensitivity analysis examined the outcome measures with a less stringent definition for PIM use—only requiring one or more prescription fills instead of two with at least 90 days-supply within a calendar year. As shown in Table 3, the results did not change significantly with the lax definition of PIM use. The prevalence of any use of PIM was 33.6% in the sensitivity analysis compared to 32.7% with the original definition and the annual days-supply for any PIM was 264.5 days (SD = 104.8) compared to 270.6 days (SD = 102.7). Other estimates of outcome measures at drug class-level similarly did not differ significantly from the main analysis (Table 3).

Table 3

Sensitivity analysis results: Prevalence and persistence of PIM use defined as any number of prescription fills with total 90 days-supply per year or more, unadjusted

(1) Prevalence of Any Use
		n=1,605,256
Any PIM	n (%)	539,312 (33.6)
Anticholinergics	n (%)	249,742 (15.6)
Benzodiazepines/Hypnotics	n (%)	229,513 (14.3)
Antipsychotics	n (%)	184,590 (11.5)
(2) Persistence as in Days-Supply Per Year
Any PIM		n=539,312
	Mean (SD), days	264.5 (104.8)
Anticholinergics		n=249,742
	Mean (SD), days	257.5 (103.4)
Benzodiazepines/Hypnotics		n=229,513
	Mean (SD), days	245.8 (103.3)
Antipsychotics		n=184,590
	Mean (SD), days	298.6 (132.9)

PIM, potentially inappropriate medication; SD, standard deviation.

DISCUSSION

This study estimated the prevalence and persistence of medication use considered potentially inappropriate for people with dementia among community-dwelling Medicare FFS beneficiaries who were enrolled in Part D plans in 2011-2017. Those who used one or more PIMs were characterized by higher percentage of female, white race, and dually eligible beneficiaries compared with those who did not. Nearly one-third of beneficiaries in the sample used one or more PIMs for 271 days per calendar year on average. For those who used antipsychotics, the mean days-supply per year was 302.7 days, suggesting that these medications were taken nearly on a daily basis.

Examination of prevalence of use by year showed an overall decreasing trend in total PIM use with an except in 2013 when benzodiazepines had a spike in their use (from 6.6% to 17.7%) likely due to the change in Medicare Part D plans to cover benzodiazepines starting 2013. The increase likely was not entirely attributed to new prescriptions in benzodiazepine-naïve beneficiaries because many previously cash-paying users, therefore not captured in the claims data, would then have started submitting claims for the coverage. The decreasing trend of PIM use appeared to be primarily driven by the steady decrease in anticholinergic drugs users over the study period, from 17.2% in 2011 to 13.4% in 2017. This trend was consistent with findings of previous European studies, which reported general trends of decrease in medications with anticholinergic properties among general older patients [30, 31]. Additionally, a previous study that examined the general older adult population in the US reported a trend of decrease in anticholinergic medication use over 2006–2010 [32]. Although the current study examined only people with dementia, our results provide a meaningful update to the previous report that suggests sustained decline in anticholinergic medications in the following years. The annual days-supply for anticholinergics increased over the study period from 259.2 days to 270.9 days. Potential underlying reasons may include steady expansion of mail order pharmacies over the study period. Mail order pharmacies typically allow dispensing of 90-day-supply prescriptions instead of 30-day-supply to reduce work load associate with delivery and for patient convenience [33]. Additionally, there has been uptake of 90-day-supply practice among retail pharmacies following reports on its association with increased medication adherence in patients with chronic conditions [33, 34].

The fraction of antipsychotic medication users in the sample remained relatively steady with slight decrease from 2015 to 2017 (11.7% to 10.7%). This is in contrast with the changes that took place in regard to antipsychotics prescribing practice in the NH setting following the CMS National Partnership Initiatives with NHs nationwide. CMS reports showed that the prevalence of off-label antipsychotic medication use reduced significantly from 23.9% to 15.7% among people residing in nursing homes from 2011 to 2017 [20]. However, our analyses suggested that there was no apparent spillover impact in the community-dwelling population of people living with dementia (Fig. 2). The mean annual days-supply for antipsychotics in the study sample had increased from 293.3 days to 310.5 days over the study period, which raises concerns about the potentially significant negative impact on both the patients and the health systems. Community-dwelling dementia patients are more likely to be in earlier stages of disease and more functional in their activities of daily living compared to those in institutional settings such as nursing homes [35]. However, extended use of antipsychotic medications can accelerate the rate of their cognitive deterioration, exerting great burden and distress on the patients and their informal caregivers [36]. Health systems will also be faced with bearing the strain of increased demand for care and patient’s premature institutionalization as patient’s cognitive and functional abilities decline faster. The value of innovative and effective solutions to improve the quality of medication management in this vulnerable patient population is significant to all stakeholders in dementia care.

The results of this study should be interpreted and applied while considering its limitations. First, the study sample was restricted to Medicare beneficiaries who were enrolled in a Part D plan, which was approximately 60% of FFS Medicare beneficiaries during the study period and, therefore, limits our ability to generate nationally representative estimates of prevalence of PIMs. The estimates presented in this study are likely higher than the national estimates as people who did not enroll in a Part D had fewer comorbidities and likely had less needs for prescription medications (Supplementary Table 3). Another limitation of this study is that, due to lack of inpatient and outpatient records for Medicare Advantage (MA) beneficiaries, the study sample only included FFS beneficiaries. The FFS beneficiaries were approximately 70% of available data, with enrollment in MA increasing over the study period. Several studies reported that healthier and younger beneficiaries are more likely to select themselves into MA enrollment [37 –39]. This suggests that our estimates may be higher than that of the full population of Medicare beneficiaries since the beneficiaries in the study sample may have been relatively sicker and likely to be taking more medications. Finally, the change in Medicare Part D coverage benefit for benzodiazepine in 2013 likely introduced bias in estimates of prevalence for years prior. As noted earlier, the change in coverage policy led to a spike in the estimates of use but the increase likely was not entirely due to a sudden increase in the true prevalence of use. This suggests that the overall PIM use prior to 2013 (30.2% in 2011, 29.4% in 2012) may be underestimates.

It is important to note that the current study examined years prior to COVID-19 pandemic and does not inform how COVID-19 may have impacted PIM use among people with dementia. In institutional settings, there have been reports of increased antipsychotic use because of worsened behavioral symptoms associated with dementia, attributable to social isolations and shortage of nursing staff during the pandemic [40, 41]. It is possible that community-dwelling individuals living with dementia may have experienced similar impacts due to COVID-19. Additional research is needed to examine how COVID-19 has impacted the patterns of PIM use in community settings.

As the age of general population increases and more people spend their later years with varying stages of dementia, it is more important than before to ensure that the quality of dementia care meets the acceptable standards. Our study results suggest that the overprescribing of PIMs is still prevalent among community-dwelling older adults living with dementia despite the existence of evidence-based practice guidelines and decision aids such as Beers Criteria. Qualitative reports suggest a few potential explanations such as prescribers having limited knowledge regarding PIM and lacking time to explore other options. These may be addressed by improving dissemination of evidence at the health system level [42]. Other identified barriers for prescribers such as patient-related factors that seem to warrant PIM use and individual prescriber’s beliefs regarding PIM and its impact may warrant targeting specific normative beliefs to shift the decision paradigm [43].

The overuse of antipsychotics particularly raises concerns that there may not be adequate oversight on medication prescribing for people with dementia in community settings. Policy efforts such as those that took place in the NH settings with CMS National Partnership Initiatives in 2011-2012 are also needed in community settings to ensure the safety and well-being of older adults with dementia and to reduce serious downstream complications associated with PIM use.

Conclusions

About one-third (32.7%) of older adults living with dementia in community settings used one or more PIMs that should be avoided in people with dementia. The mean annual days-supply for PIMs was 270.6 days per year. For those who used antipsychotic medications (11.3% of the sample), the mean annual days-supply was 302.7 days per year suggesting patterns of long-term, almost daily use. The mean annual days-supply for antipsychotics had steadily increased from 293.3 days to 310.5 days over the study period. The steady rate of antipsychotic use suggested there was no apparent impact of the CMS National Partnership initiatives with NHs in the community settings and warrant policy efforts and innovation solutions to improve the quality of medication management for this vulnerable patient population.

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

The authors have no funding to report.

CONFLICT OF INTEREST

Yuna Bae-Shaaw, Victoria Shier, Neeraj Sood, Seth Seabury, and Geoffrey Joyce declare that they have no financial, personal, or other potential conflicts of interest to disclose. Seabury has a consulting arrangement with Bristol Myers Squibb that is unrelated to the submitted work.

DATA AVAILABILITY

The datasets generated and analyzed during the current study contain identifiable patient health information and cannot be made publicly available. However, the original data from which the analytic datasets were generated are available at the Research Data Assistance Center (ResDAC) and can be accessed by submitting a Research Identifiable File (RIF) request at: resdac.org.

The supplementary material is available in the electronic version of this article: .

References

(2022) 2022 Alzheimer’s disease facts and figures., Alzheimers Dement 18, 700–789.

Rajan

, Weuve

, Barnes

, McAninch

, Wilson

, Evans

(2021) Population estimate of people with clinical Alzheimer’s disease and mild cognitive impairment in the United States (2020–2060), Alzheimers Dement 17, 1966–1975.

Beers

(1997) Explicit criteria for determining potentially inappropriate medication use by the elderly. An update, Arch Intern Med 157, 1531–1536.

Fick

, Cooper

, Wade

, Waller

, Maclean

, Beers

(2003) Updating the Beers Criteria for potentially inappropriate medication use in older adults: Results of a US consensus panel of experts, Arch Intern Med 163, 2716–2724.

O’Mahony

, Gallagher

(2008) Inappropriate prescribing in the older population: Need for new criteria, Age Ageing 37, 138–141.

Hanlon

, Schmader

, Ruby

, Weinberger

(2001) Suboptimal prescribing in older inpatients and outpatients, J Am Geriatr Soc 49, 200–209.

By the 2019 American Geriatrics Society Beers Criteria^®Update Expert Panel (2019) American Geriatrics Society 2019 updated AGS Beers Criteria^® for potentially inappropriate medication use in older adults, J Am Geriatr Soc 67, 674–694.

Ziere

, Dieleman

, Hofman

, Pols

, van der Cammen

, Stricker

(2006) Polypharmacy and falls in the middle age and elderly population, Br J Clin Pharmacol 61, 218–223.

Hartikainen

, Lonnroos

, Louhivuori

(2007) Medication as a risk factor for falls: Critical systematic review, J Gerontol A Biol Sci Med Sci 62, 1172–1181.

10.

Banaszak-Holl

, Fendrick

, Foster

, Herzog

, Kabeto

, Kent

, Straus

, Langa

(2004) Predicting nursing home admission: Estimates from a 7-year follow-up of a nationally representative sample of older Americans, Alzheimer Dis Assoc Disord 18, 83–89.

11.

Boudreau

, Hanlon

, Roumani

, Studenski

, Ruby

, Wright

, Hilmer

, Shorr

, Bauer

, Simonsick

, Newman

(2009) Central nervous system medication use and incident mobility limitation in community elders: The Health, Aging, and Body Composition study, Pharmacoepidemiol Drug Saf 18, 916–922.

12.

Colloca

, Tosato

, Vetrano

, Topinkova

, Fialova

, Gindin

, van der Roest

, Landi

, Liperoti

, Bernabei

, Onder

(2012) Inappropriate drugs in elderly patients with severe cognitive impairment: Results from the shelter study, PLoS One 7, e46669.

13.

Centers for Medicare and Medicaid Services (CMS), HHS (2016) Medicare and Medicaid programs; reform of requirements for long term care facilities, Fed Regist 81, 68688–68872.

14.

Lau

, Mercaldo

, Harris

, Trittschuh

, Shega

, Weintraub

(2010) Polypharmacy and potentially inappropriate medication use among community-dwelling elders with dementia, Alzheimer Dis Assoc Disord 24, 56–63.

15.

Schneider

, Tariot

, Dagerman

, Davis

, Hsiao

, Ismail

, Lebowitz

, Lyketsos

, Ryan

, Stroup

, Sultzer

, Weintraub

, Lieberman

(2006) Effectiveness of atypical antipsychotic drugs in patients with Alzheimer’s disease, N Engl J Med 355, 1525–1538.

16.

Maust

, Kim

, Seyfried

, Chiang

, Kavanagh

, Schneider

, Kales

(2015) Antipsychotics, other psychotropics, and the risk of death in patients with dementia: Number needed to harm, JAMA Psychiatry 72, 438–445.

17.

Tampi

, Tampi

, Balachandran

, Srinivasan

(2016) Antipsychotic use in dementia: A systematic review of benefits and risks from meta-analyses, Ther Adv Chron Dis 7, 229–245.

18.

Ralph

, Espinet

(2018) Increased all-cause mortality by antipsychotic drugs: Updated review and meta-analysis in dementia and general mental health care, J Alzheimers Dis Rep 2, 1–26.

19.

Michel

, Zekry

, Mulligan

, Giacobini

, Gold

(2001) Economic considerations of Alzheimer’s disease and related disorders, Aging (Milano) 13, 255–260.

20.

Centers for Medicare & Medicaid Services, National Partnership to Improve Dementia Care in Nursing Homes: Antipsychotic Medications Use Data Report (October 2021), http://www.cms.gov/files/document/antipsychoticmedication-use-data-report-2020q4-updated-07302021.pdf, Last updated October, 2021, Accessed on February 8, 2022.

21.

Weston

, Weinstein

, Barton

, Yaffe

(2010) Potentially inappropriate medication use in older adults with mild cognitive impairment, J Gerontol A Biol Sci Med Sci 65, 318–321.

22.

Kachru

, Carnahan

, Johnson

, Aparasu

(2015) Potentially inappropriate anticholinergic medication use in older adults with dementia, J Am Pharm Assoc (2003) 55, 603–612.

23.

Kales

, Zivin

, Kim

, Valenstein

, Chiang

, Ignacio

, Ganoczy

, Cunningham

, Schneider

, Blow

(2011) Trends in antipsychotic use in dementia 1999-2007, Arch Gen Psychiatry 68, 190–197.

24.

Grodstein

, Chang

, Capuano

, Power

, Marquez

, Barnes

, Bennett

, James

, Bynum

, Lipsitz

(2021) Identification of dementia in recent medicare claims data, compared to rigorous clinical assessments, J Gerontol A Biol Sci Med Sci 77, 1272–1278.

25.

Chen

, Tysinger

, Crimmins

, Zissimopoulos

(2019) Analysis of dementia in the US population using Medicare claims: Insights from linked survey and administrative claims data, Alzheimers Dement 5, 197–207.

26.

Gorina

, Kramarow

(2011) Identifying chronic conditions in Medicare claims data: Evaluating the Chronic Condition Data Warehouse algorithm, Health Serv Res 46, 1610–1627.

27.

Thunell

, Ferido

, Zissimopoulos

(2019) Measuring Alzheimer’s disease and other dementias in diverse populations using Medicare claims data, J Alzheimers Dis 72, 29–33.

28.

American Geriatrics Society 2012 Beers Criteria Update Expert Panel (2012) American Geriatrics Society updated Beers Criteria for potentially inappropriate medication use in older adults, J Am Geriatrics Society 60, 616–631.

29.

By the American Geriatrics Society 2015 Beers Criteria Update Expert Panel (2015) American Geriatrics Society 2015 updated beers criteria for potentially inappropriate medication use in older adults, J Am Geriatr Soc 63, 2227–2246.

30.

Hovstadius

, Petersson

, Hellström

, Ericson

(2014) Trends in inappropriate drug therapy prescription in the elderly in Swfrom 2006 to 2013: Assessment using national indicators, Drugs Aging 31, 379–386 eden.

31.

Drusch

, Le Tri

, Ankri

, Zureik

, Herr

(2021) Decreasing trends in potentially inappropriate medications in older people: A nationwide repeated cross-sectional study, BMC Geriatrics 21, 621.

32.

Davidoff

, Miller

, Sarpong

, Yang

, Brandt

, Fick

(2015) Prevalence of potentially inappropriate medication use in older adults using the 2012 beers criteria, J Am Geriatr Soc 63, 486–500.

33.

Duru

, Schmittdiel

, Dyer

, Parker

, Uratsu

, Chan

, Karter

(2010) Mail-order pharmacy use and adherence to diabetes-related medications, Am J Manag Care 16, 33–40.

34.

Liberman

, Girdish

(2011) Recent trends in the dispensing of 90-day-supply prescriptions at retail pharmacies: Implications for improved convenience and access, Am Health Drug Benefits 4, 95–100.

35.

Olsen

, Pedersen

, Bergland

, Enders-Slegers

, Jøranson

, Calogiuri

, Ihlebæk

(2016) Differences in quality of life in home-dwelling persons and nursing home residents with dementia - a cross-sectional study, BMC Geriatrics 16, 137.

36.

Vigen

, Mack

, Keefe

, Sano

, Sultzer

, Stroup

, Dagerman

, Hsiao

, Lebowitz

, Lyketsos

, Tariot

, Zheng

, Schneider

(2011) Cognitive effects of atypical antipsychotic medications in patients with Alzheimer’s disease: Outcomes from CATIE-AD, Am J Psychiatry 168, 831–839.

37.

Layton

(2017) Imperfect risk adjustment, risk preferences, and sorting in competitive health insurance markets, J Health Econ 56, 259–280.

38.

Montz

, Layton

, Busch

, Ellis

, Rose

, McGuire

(2016) Risk-adjustment simulation: Plans may have incentives to distort mental health and substance use coverage, Health Aff (Millwood) 35, 1022–1028.

39.

Ndumele

, Elliott

, Haviland

, Burkhart

, Orr

, Gaillot

, Cleary

(2019) Differences in hospitalizations between fee-for-service and Medicare advantage beneficiaries, Med Care 57, 8–12.

40.

Paananen

, Rannikko

, Harju

, Pirhonen

(2021) The impact of Covid-19-related distancing on the well-being of nursing home residents and their family members: A qualitative study, Int J Nurs Stud Adv 3, 100031.

41.

Levere

, Rowan

, Wysocki

(2021) The adverse effects of the COVID-19 pandemic on nursing home resident well-being, J Am Med Dir Assoc 22, 948–954.e2.

42.

Voigt

, Gottschall

, Köberlein-Neu

, Schübel

, Quint

, Bergmann

(2016) Why do family doctors prescribe potentially inappropriate medication to elderly patients? BMC Fam Pract 17, 93.

43.

Gill

, Almutairi

, Donyai

(2019) “The Lesser of Two Evils” versus “Medicines not Smarties”: Constructing antipsychotics in dementia, Gerontologist 59, 570–579.