Abstract
Aims
To describe and compare clinical characteristics, healthcare costs, and institutionalization/mortality outcomes among patients with and without agitation associated with Alzheimer’s dementia (AAD).
Methods
Data from the Reliant Medical Group database (01/01/2016-03/31/2020) were used, including claims, electronic medical records, and clinical information/physician notes abstracted from medical charts. Patients aged ≥55 years with Alzheimer’s dementia (AD) were observed during a randomly selected 12-month study period after AD diagnosis. Using information recorded in medical charts, patients were classified into cohorts based on experiencing (agitation cohort) and not experiencing (no agitation cohort) agitated behaviours during the study period. Entropy balancing was used to create reweighted cohorts with similar characteristics. Study outcomes (patient demographic and clinical characteristics, treatments received, healthcare costs, institutionalization and death events) were compared between cohorts; agitation characteristics were described for the agitation cohort only.
Results
Among 711 patients included in the study, 240 were classified in the agitation cohort and 471 in the no agitation cohort. After reweighting, several comorbidities were more frequently observed in the agitation versus no agitation cohort, including infection, depression, and altered mental status. Use of antidepressants, anticonvulsants, antipsychotics, and antianxiety medications was more common in the agitation versus no agitation cohort. Common agitated behaviours included hitting (20.8%), pacing/aimless wandering (17.5%), and cursing/verbal aggression (15.0%). Total all-cause healthcare costs were $4287 per-patient-per-year higher in the agitation cohort versus no agitation cohort (p = 0.04), driven by higher inpatient costs. Death was more common and time to death and institutionalization were shorter in the agitation versus no agitation cohort.
Limitations
Results may not be generalizable to the US population with AD.
Conclusions
Among patients with AD, agitation was associated with shorter time to death/institutionalization and increased comorbidities, medication use, and healthcare costs, highlighting the additional clinical and economic burden that agitation poses to patients and the healthcare system.
Alzheimer’s dementia (AD) is a neurodegenerative brain disease that is estimated to affect 6.7 million adults aged ≥65 years in the United States (US) in 2023Citation1. While cognitive decline is a well-recognized sign of AD, neuropsychiatric symptoms, including apathy, depression, sleep disorders, hallucinations, delusions, and agitation, are also highly prevalent in dementia and are often the most distressing to caregivers and patientsCitation2,Citation3. Agitation is a particularly common neuropsychiatric symptom of AD that is observed in ∼45% to 50% of patients with AD/dementia, though this prevalence may be underestimated given the potential incompleteness and limitations of the clinical data used to derive these estimatesCitation4,Citation5. According to the International Psychogeriatric Association (IPA), agitation represents a wide range of emotionally distressing behaviours, including exaggerated motor activity and verbal and physical aggression, which significantly impair activities of daily living, interpersonal relationships, and/or other aspects of social functioningCitation6.
To manage agitation, the American Psychiatric Association (APA) recommends the use of non-pharmacological interventions as first-line approaches, while pharmacological agents may be recommended when symptoms are severe, dangerous, or cause significant distress to the patient and caregiverCitation7. However, until recently, there were no pharmacological treatments approved by the US Food and Drug Administration for the management of agitation associated with AD (AAD)Citation8. Therefore, pharmacological treatments employed included off-label use of typical and atypical antipsychotics, selective serotonin reuptake inhibitors (SSRIs), other antidepressants, anticonvulsants, anxiolytics, and mood stabilizersCitation9,Citation10.
Given the prior lack of effective pharmacological treatments and the negative impact of symptoms on daily life, agitation is associated with a substantial clinical burdenCitation11. A previous study demonstrated that agitation severity was correlated with the need for professional caregivers, institutionalization, hospitalization in a psychiatric ward, and treatment with antipsychotics and antidepressantsCitation12. Detailed information on the drivers of poor outcomes and costs associated with AAD is needed to provide a better understanding of the impact of agitated behaviours on patients and the healthcare system. However, the International Classification of Diseases, Tenth revision, Clinical Modification [ICD-10-CM] diagnostic codes for agitation were not specific to AD until recently, and the use of these codes was generally inconsistent. Given the natural delay in uptake of the new codes and subsequent availability of data, combining claims and electronic medical records (EMR)/medical chart abstraction provides an opportunity for the identification of patients with agitation and the assessment of economic outcomes in the meantime. Leveraging this type of data source, the current study aimed to describe and compare clinical characteristics, healthcare costs, and incidence of death and institutionalization among patients with and without AAD.
Data source
The Reliant Medical Group Database (1 January 2016 – 31 March 2020) was used to conduct this analysis. The Reliant Medical Group is an integrated delivery network in Massachusetts that covers over 1.2 million lives annually, with an average of 247,000 unique patients represented in the data. It combines administrative health insurance claims data with EMR for a portion of their patients. Importantly, the Reliant Medical Group has the capability to retrieve medical charts and abstract specific information from unstructured physician notes, allowing for the identification of patients experiencing specific agitated behaviours (i.e. information that is typically not documented in prepopulated, structured EMR fields or in claims data). The claims portion of the database provides detailed information on medical services received and associated costs.
This study received an exemption from the Western Copernicus Group (WCG) Institutional Review Board (IRB).
Study design and data collection
This was an observational study following a retrospective cohort design. The index date was defined as a randomly selected date after AD diagnosis that was followed by ≥12 months of continuous health plan enrolment (i.e. the study period). The baseline period comprised the 6-month period prior to the index date. Patients were identified in the administrative claims data and unique coded IDs were used to abstract information from EMR and medical charts using a defined electronic case report form (eCRF).
Study population
Patients were included in the study if they met the following inclusion criteria: (1) had ≥1 diagnosis of AD (ICD-10-CM codes: G30.xx); (2) had ≥12 months of continuous health plan enrolment following their AD diagnosis; (3) were aged ≥55 years as of the index date (). Patients were excluded if the documented information on their agitation status and symptoms was unclear (i.e. abstracted as having agitation but only neuropsychiatric symptoms were recorded; it is unclear whether their charts had additional information to denote agitation or whether these symptoms were misclassified) or they had missing costs for 100% of their pharmacy claims.
Figure 1. Sample selection flow chart. Abbreviations. AD, Alzheimer’s dementia; ICD-10-CM, International classification of diseases, Tenth revision, Clinical Modification. Note: 1. Patients who were identified as having agitation based solely on neuropsychiatric symptoms (i.e. hallucinations, delirium, paranoia) were excluded, as the true agitation state was not identifiable based on the information collected (i.e. unclear whether their charts had additional information to denote agitation or whether these symptoms were misclassified).
Included patients were classified into either the “agitation cohort” (i.e. experienced ≥1 agitated behaviour as documented in their medical chart during the study period) or the “no agitation cohort” (i.e. did not experience any agitated behaviours during the study period). The list of agitated behaviours was derived from agitation measurement instruments (e.g. Cohen-Mansfield Agitation Inventory), professional association definitions (e.g. IPACitation6) and clinical expert opinion.
Study measures
EMR data were used to obtain patient demographic information (i.e. age, gender, race, ethnicity, calendar year of index date) and smoking history at the index date and data on death (i.e. proportion of patients who died, time from AD diagnosis or index date to death) or institutionalization (i.e. proportion of patients who lived in a private or group residence and the proportion who subsequentially moved to an institutionalized setting; time from index date, AD diagnosis, or first onset of agitated behaviors to institutionalization) at any time. Death and institutionalization were evaluated from the index date until the earliest of the end of eligibility, death, or the end of data availability.
Data on comorbidities (i.e. Charlson Comorbidity Index [CCI], proportion of patients with selected comorbidities), prescription treatments (i.e. proportion of patients receiving selected treatments), and healthcare costs during the study period were collected from administrative claims. All-cause and AD-related healthcare costs (defined as costs for claims during which a diagnosis for AD was recorded) included medical and pharmacy components and were reported per-patient-per-year (PPPY) in 2021 US dollars (USD). Medical costs were those incurred for services covered by payers and included costs for inpatient (i.e. admissions in hospital settings and skilled nursing facilities), emergency room, outpatient, and other (e.g. telehealth, home health) visits. For AD-related healthcare costs, pharmacy costs represented only those associated with use of antidementia, antidepressant, antipsychotic, and antianxiety medications.
Data collected from medical chart abstraction included information on AD diagnosis (i.e. severity, year of diagnosis, time from first onset of AD symptoms to diagnosis, time from AD diagnosis or first onset of symptoms to index date), agitated behaviours (i.e. proportion of patients with aggressive behaviours, physically non-aggressive behaviours, verbally agitated behaviours, hiding and hoarding behaviours, other behaviours) during the study period, and living situation at the index date.
Statistical analysis
Entropy balancing was used to balance observed key characteristics that may have an impact on the differences in outcomes across the agitation and no agitation cohorts (i.e. age, gender, race, ethnicity, calendar year of index date, smoking history, time from first onset of AD symptoms to diagnosis, time from AD diagnosis to index date, time from first onset of AD symptoms to index date, severity of AD, primary residence, living situation, and CCI score)Citation13. Patients in the no agitation cohort were assigned weights such that the distribution of key characteristics had the same mean and standard deviation as the agitation cohort. Subject to rounding exceptions, weights were normalized so that the sum of weights was equal to the number of patients in each cohort. Standardized differences were used to assess the balance of patient characteristics, comorbidities, and death/institutionalization characteristics between cohorts, with differences <0.1 considered as balancedCitation14.
Incremental costs associated with agitation were estimated using weighted two-part models, where the first part was a logistic model with a binomial distribution and the second part was a generalized linear model with a log link and a gamma distribution. Adjusted cost differences were reported with their p-values and 95% confidence intervals (CIs).
Sensitivity analysis
A sensitivity analysis was conducted to provide a more conservative approach to identifying patients for inclusion in the agitation cohort, where patients were required to experience ≥2 agitated behaviours (excluding unspecific agitation), as opposed to ≥1 in the primary analysis. All-cause and AD-related healthcare costs were compared between the agitation and no agitation cohorts.
Patient characteristics
Among 711 patients included in the study, 240 were classified in the agitation cohort and 471 in the no agitation cohort (). Before weighting, several characteristics were unbalanced between the two cohorts (). Therefore, entropy balancing was used to balance observed key characteristics that may have an impact on the differences in outcomes across the two cohorts. After weighting, most characteristics were balanced between the cohorts (standardized difference <0.1; ). Mean ± SD age was 84.9 ± 7.7 and 84.8 ± 7.7 years in the agitation and no agitation cohorts, respectively, and 66.7% were female. Most patients were white (84.6% in the agitation cohort and 84.5% in the no agitation cohort). The largest proportion of patients lived in a group home (49.2% and 49.0%, respectively), such as assisted living, nursing home, or convent. AD was severe (based on clinical documentation from medical chart abstraction) in 32.9% of patients in both cohorts.
Table 1. Patient characteristics.
Comorbidities and treatments
Several comorbidities were more common in the agitation cohort compared to the no agitation cohort during the 12-month study period, including infection (55.0% vs 43.7%), depression (57.5% vs 37.2%), and altered mental status (42.9% vs 28.2%; ).
Table 2. Comorbidities and treatments during the study period.
Use of antidepressants (53.8% vs 40.6%), anticonvulsants (27.1% vs 18.3%), antipsychotics (29.6% vs 9.6%), and antianxiety medications (8.8% vs 5.1%) was more common in the agitation cohort compared to the no agitation cohort during the 12-month study period ().
Agitation characteristics
The 240 patients in the agitation cohort experienced a mean ± SD of 1.9 ± 1.2 documented agitated behaviours during the study period. Common aggressive behaviours included hitting (20.8%), cursing or verbal aggression (15.0%), and screaming (10.4%; ). Common physically non-aggressive behaviours included pacing/aimless wandering (17.5%), general restlessness (13.3%), and trying to get to a different place (10.4%; ).
Figure 2. Agitation characteristics during the study period.
Healthcare costs
Total all-cause healthcare costs were $4287 PPPY higher in the agitation cohort compared to the no agitation cohort ($25,032 vs $20,745; p = 0.04; ). Total AD-related healthcare costs were $6242 PPPY higher in the agitation cohort compared to the no agitation cohort ($17,259 vs $11,017; p < 0.01; ). Inpatient costs were a major driver of this incremental cost, with an adjusted cost difference of $4941 PPPY ($13,182 vs $8241; p = 0.02).
Figure 3. Healthcare costs incurred during the study period. Abbreviations. AD, Alzheimer’s dementia; CI, confidence interval; PPPY, per-patient-per-year. *Significant at the 5% level. Notes: 1. IP costs includes admissions in hospital settings and skilled nursing facilities. 2. For AD-related healthcare costs, pharmacy costs represented only those associated with use of antidementia, antidepressant, antipsychotic, and antianxiety medications.
Death and institutionalisation
The mean (median) duration of follow-up for death and institutionalization outcomes was 16.8 (7.6) months. A higher proportion of patients in the agitation cohort died any time after the study period compared to the no agitation cohort (64.6% vs 57.3%; ). Moreover, the mean ± SD time from index date to death was shorter in the agitation cohort than the no agitation cohort (1.8 ± 0.9 vs 2.0 ± 0.9 years).
Table 3. Death and institutionalization characteristics at any time.
Among patients living in a private or group residence at any time (47.5% in the agitation cohort and 47.6% in the no agitation cohort), a higher proportion of patients in the agitation cohort moved to an institutionalized setting compared to the no agitation cohort (39.5% vs 26.3%; ). The mean ± SD time from the index date to institutionalization was shorter in the agitation cohort than the no agitation cohort (1.3 ± 0.9 vs 1.8 ± 1.2 years).
Sensitivity analysis
Using a more conservative definition of agitation, 108 patients were classified in the agitation cohort and 471 in the no agitation cohort. Patient characteristics were generally similar to the primary analysis cohorts. Total all-cause healthcare costs were $5449 PPPY higher in the agitation cohort compared to the no agitation cohort ($27,576 vs $22,127; p = 0.02; Table S1). Total AD-related healthcare costs were $8,682 PPPY higher in the agitation cohort compared to the no agitation cohort ($20,288 vs $11,606; p < 0.01).
In this real-world study, patients with AD and agitation had a larger clinical burden, including more comorbidities and higher pharmacological treatment utilization, than those without agitation. The combined use of insurance claims data with abstractable physician notes allowed for a unique evaluation of the specific economic burden associated with agitation. Patients with agitation incurred $4287 PPPY higher healthcare costs than those without agitation, primarily driven by hospitalizations. Furthermore, death and institutionalization were more common and occurred sooner among patients with agitation than without. The sensitivity analysis using a more conservative definition of agitation produced similar trends, with higher all-cause and AD-related healthcare cost differences than the main analysis.
Though there is limited evidence on the burden of agitation, our findings are consistent with prior literature demonstrating the large clinical burden associated with behavioural disturbances, which include agitation. In a claims-based study of Medicare beneficiaries with dementia, behavioural disturbances (as a proxy for agitation) were associated with significantly more medical and psychiatric comorbidities (e.g. urinary tract infection, depression, altered mental state) and greater medication use (e.g. antidepressants, antipsychotics) compared to not having behavioural disturbancesCitation15. Similarly, these comorbidities and medications were also more common in the agitation cohort than the no agitation cohort in the current study. In a separate physician survey analysis of patients with dementia, a greater number and severity of agitation symptoms was significantly correlated with a larger proportion of patients that were institutionalized and receiving antipsychotic and antidepressant treatmentCitation12. These findings are aligned with those of the current study and together, they highlight the particularly substantial comorbidity and treatment burden associated with agitation among patients with dementia.
Given the historic lack of an ICD-10-CM code for AAD, evaluating the economic burden associated with agitation is challenging. Therefore, a major strength of the current study is its use of the Reliant Medical Group Database, whereby patients experiencing agitated behaviours could be identified from physician notes and linked to insurance claims. Jones et al. evaluated costs based on a physician survey analysis and found that mean total healthcare costs in the 12 months prior to data collection were higher for patients with agitation ($20,041) than without ($9243), but these estimates were derived from calculations using unit costs from the literatureCitation12. Similarly, Cloutier et al. used aggregated unit costs from a literature review to estimate a total incremental cost of agitation-related institutionalization of $4.3 billion among patients with AD from the National Alzheimer’s Coordinating Centre Uniform Data Set Citation16 Literature-based costs may not represent the nuances of the AAD population; therefore, the current study provides a more refined estimate of the incremental costs of agitation.
Importantly, the burden of agitation can be influenced by many factors, including severity of AD. Since agitation is known to be associated with disease severityCitation17, the latter was included as a variable in the entropy balancing; however, the lack of complete and validated severity data may have limited the ability to fully adjust for this variable. Given the association between AD severity and higher costs of careCitation18, further research is warranted to evaluate the effect of severity on the costs of agitation specifically.
The large clinical and economic burden of agitation demonstrated in this study signals a need for more effective management strategies, including improved patient monitoring, documentation methods (e.g. education on the use of new ICD-10-CM codes to improve uptake), and awareness of the condition, which would contribute to a better understanding of agitation and its consequences. Moreover, the lack of pharmacological treatments approved for the management of these commonly occurring agitated behaviours in AD during the study period highlights the unmet need for additional treatment options. Indeed, the elevated use of antidepressants, antipsychotics, and anticonvulsants by patients with agitation may represent off-label use of medications for the management of agitation, even if these treatments are associated with mixed clinical outcomes and potential safety concernsCitation19,Citation20. Improved management of agitation has the potential to reduce the financial burden from both the payer’s and societal perspectives, while providing additional benefits for patient and caregiver quality of life and family engagement during patients’ later years of life.
Limitations
The study findings should be interpreted in the context of some limitations. The sample size was limited due to the requirement for the combination of claims and EMR data. Since the study population only included commercially insured patients from an integrated delivery network in Massachusetts, with most patients being white, results may not be generalizable to the US population with AD. Moreover, information captured by the study’s eCRF was limited to information available in the patients’ medical records. As such, assessment of some characteristics was not possible if the associated data were not completely available (e.g. severity of AD). If agitated behaviours were not systematically recorded or did not occur during the study period, the burden of agitation may have been underestimated. Additionally, physician documentation of agitated behaviours may have been biased towards more noticeable or damaging behaviours. Healthcare costs were limited to those that were available in the claims data (i.e. were reimbursed by covered insurance plans); as such, healthcare costs incurred during institutionalization may not have been captured. Relatedly, costs incurred outside of the Reliant network or covered by other insurance plans may not have been captured in the data. Lastly, despite balancing observable key characteristics between the agitation and no agitation cohorts, there may have been residual confounding due to unobservable confounders.
Among patients with AD in this study, agitation was associated with increased comorbidities, medication use, and healthcare costs, highlighting the additional burden that agitation poses for the management of AD. Strategies focusing on improving agitation management have the potential to mitigate the burden associated with AD for payers and the healthcare system.
Transparency
Declaration of funding
This study was funded by Otsuka Pharmaceutical Development & Commercialization, Inc. and Lundbeck LLC.
Declaration of financial/other relationships
GG served as a clinical consultant to Otsuka Pharmaceutical Development & Commercialization, Inc. and Lundbeck LLC at the time of study completion. AU is an employee of Lundbeck LLC. JS and JA are employees of Otsuka Pharmaceutical Development & Commercialization, Inc. RB, MC, MGL, DC, and AG are employees of Analysis Group, Inc., a consulting company that has provided paid consulting services to Otsuka Pharmaceutical Development & Commercialization, Inc., which funded the development and conduct of this study and manuscript.
Author contributions
RB, MC, MGL, DC, and AG contributed to study conception and design, collection and assembly of data, and data analysis and interpretation. GG, AU, JS, and JA contributed to study conception and design, data analysis and interpretation. All authors reviewed and approved the final content of this manuscript.
Reviewer disclosures
A reviewer on this manuscript has disclosed that they are a consultant at IQVIA. Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.
Previous presentations
Part of the material in this manuscript was presented at the International Society for Pharmacoeconomics and Outcomes Research 2023 Annual Conference, held on May 7–10 in Boston, MA, USA, as a poster presentation.
Supplemental Material
Download MS Word (28.7 KB)Acknowledgements
Medical writing support was provided by a professional medical writer, Christine Tam, MWC, an employee of Analysis Group, Inc.
Data availability statement
Data is not available due to legal restrictions. Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available. Therefore, restrictions apply to the availability of these data, which are not publicly available.
References
- Alzheimer’s Association. 2023. Alzheimer’s disease facts and figures: Special report [cited 2023 June 16]. Available from: https://www.alz.org/media/Documents/alzheimers-facts-and-figures.pdf
- Phan SV, Osae S, Morgan JC, et al. Neuropsychiatric symptoms in dementia: considerations for pharmacotherapy in the USA. Drugs R D. 2019;19(2):93–115. doi: 10.1007/s40268-019-0272-1.
- Knopman DS, Amieva H, Petersen RC, et al. Alzheimer disease. Nat Rev Dis Primers. 2021;7(1):33. doi: 10.1038/s41572-021-00269-y.
- Fillit H, Aigbogun MS, Gagnon-Sanschagrin P, et al. Impact of agitation in long-term care residents with dementia in the United States. Int J Geriatr Psychiatry. 2021;36(12):1959–1969. doi: 10.1002/gps.5604.
- Halpern R, Seare J, Tong J, et al. Using electronic health records to estimate the prevalence of agitation in alzheimer disease/dementia. Int J Geriatr Psychiatry. 2019;34(3):420–431. doi: 10.1002/gps.5030.
- Sano M, Cummings J, Auer S, et al. Agitation in cognitive disorders: progress in the international psychogeriatric association consensus clinical and research definition. Int Psychogeriatr. 2023; 7:1–13. doi: 10.1017/S1041610222001041.
- Reus VI, Fochtmann LJ, Eyler AE, et al. The American psychiatric association practice guideline on the use of antipsychotics to treat agitation or psychosis in patients with dementia. Am J Psychiatry. 2016;173(5):543–546. doi: 10.1176/appi.ajp.2015.173501.
- US Food and Drug Administration. FDA Approves First Drug to Treat Agitation Symptoms Associated with Dementia due to Alzheimer’s Disease 2023 [cited 2023 June 16]. Available from: https://www.fda.gov/news-events/press-announcements/fda-approves-first-drug-treat-agitation-symptoms-associated-dementia-due-alzheimers-disease#:∼:text=Today%2C%20the%20U.S.%20Food%20and,treatment%20option%20for%20this%20indication.
- Magierski R, Sobow T, Schwertner E, et al. Pharmacotherapy of behavioral and psychological symptoms of dementia: state of the art and future progress. Front Pharmacol. 2020;11:1168. doi: 10.3389/fphar.2020.01168.
- Porsteinsson AP, Antonsdottir IM. An update on the advancements in the treatment of agitation in alzheimer’s disease. Expert Opin Pharmacother. 2017;18(6):611–620. doi: 10.1080/14656566.2017.1307340.
- Carrarini C, Russo M, Dono F, et al. Agitation and dementia: prevention and treatment strategies in acute and chronic conditions. Front Neurol. 2021;12:644317. doi: 10.3389/fneur.2021.644317.
- Jones E, Aigbogun MS, Pike J, et al. Agitation in dementia: real-world impact and burden on patients and the healthcare system. J Alzheimers Dis. 2021;83(1):89–101. doi: 10.3233/JAD-210105.
- Hainmueller J. Entropy balancing for causal effects: a multivariate reweighting method to produce balanced samples in observational studies. Polit Anal. 2012;20(1):25–46. doi: 10.1093/pan/mpr025.
- Austin P. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Communications in Statistics - Simulation and Computation. 2009;38(6):1228–1234. doi: 10.1080/03610910902859574.
- Aigbogun MS, Stellhorn R, Hartry A, et al. Treatment patterns and burden of behavioral disturbances in patients with dementia in the United States: a claims database analysis. BMC Neurol. 2019;19(1):33. doi: 10.1186/s12883-019-1260-3.
- Cloutier M, Gauthier-Loiselle M, Gagnon-Sanschagrin P, et al. Institutionalization risk and costs associated with agitation in alzheimer’s disease. Alzheimers Dement (N Y). 2019;5(1):851–861. doi: 10.1016/j.trci.2019.10.004.
- Anatchkova M, Brooks A, Swett L, et al. Agitation in patients with dementia: a systematic review of epidemiology and association with severity and course. Int Psychogeriatr. 2019;31(9):1305–1318. doi: 10.1017/S1041610218001898.
- Marešová P, Dolejs J, Mohelska H, et al. Cost of treatment and care for people with alzheimer’s disease: a meta- analysis. Curr Alzheimer Res. 2019;16(14):1245–1253. doi: 10.2174/1567205017666200102144640.
- Kongpakwattana K, Sawangjit R, Tawankanjanachot I, et al. Pharmacological treatments for alleviating agitation in dementia: a systematic review and network meta-analysis. Br J Clin Pharmacol. 2018;84(7):1445–1456. doi: 10.1111/bcp.13604.
- Mühlbauer V, Möhler R, Dichter MN, et al. Antipsychotics for agitation and psychosis in people with alzheimer’s disease and vascular dementia. Cochrane Database Syst Rev. 2021;12(12):CD013304. doi: 10.1002/14651858.CD013304.pub2.