Agitation in Dementia: Real-World Impact and Burden on Patients and the Healthcare System

INTRODUCTION

Behavioral or neuropsychiatric symptoms are experienced by more than 90%of patients with dementia at some point in the course of their disease [1, 2]; these include agitation, psychotic symptoms (hallucinations and delusions), apathy, depression, and sleep disturbances [3]. The prevalence of agitation has been reported to be 60%in patients with mild cognitive impairment (CI) and 76%in those with Alzheimer’s disease (AD) [4]. A systematic review reported that the proportion of patients with agitation symptoms increase slightly or remain stable over the course of dementia, but that agitation severity ratings increase over time [5].

Despite the high prevalence of agitation in patients with dementia, its consideration as a distinct entity has been poor [6]. Until recently, there was no widespread agreement on what elements comprise agitation or its definition. In 2015, the International Psychogeriatric Association (IPA) published a provisional consensus on the definition of agitation: 1) occurring in patients with a CI or dementia syndrome; 2) exhibiting behavior consistent with emotional distress; 3) manifesting excessive motor activity, verbal aggression, or physical aggression; and 4) evidencing behaviors that cause excess disability and are not solely attributable to another disorder [7].

Agitation and aggression in AD have been shown to be associated with reductions in cholinergic and serotonergic markers [8]; and while treatments that are known to affect relevant pathways are commonly used, there is currently no Food and Drug Administration-approved pharmacologic treatment for agitation in dementia [9]. Pharmacologic treatments currently prescribed for dementia-associated agitation include antipsychotics, antidepressants, anticonvulsants, anxiolytics, and acetylcholinesterase inhibitors [10]. The 2016 American Psychiatric Association Practice Guidelines recommend a comprehensive personalized treatment plan, including non-pharmacologic and pharmacologic interventions, with antipsychotics recommended when agitation symptoms are severe [9].

Agitation has been reported to adversely affect patient care, increasing the chances of institutionalization and impacting the quality of life of patients [6, 7]. Furthermore, a systematic review of literature revealed a paucity of data on the economic burden of agitation in dementia [11]. Therefore, the objective of this study was to assess the impact and burden of agitation in dementia on patients and the healthcare system.

MATERIALS AND METHODS

RESULTS

Caregiver presence

Physician-reported data on caregiver presence were available for 1,341 patients; 442 (67.8%) non-agitated and 594 (86.2%) agitated patients required a caregiver (professional or non-professional); of these, 100 (15.3%) non-agitated patients and 248 (36.0%) agitated patients needed a professional caregiver.

Regression analysis showed no difference in the proportion of patients with a caregiver based on agitation score (p = 0.329), but the proportion of patients with a professional caregiver increased with increasing agitation score (p = 0.006).

Sensitivity analysis using≥2 symptoms to define an agitated patient, following propensity score matching, confirmed that agitated patients were more likely than non-agitated patients to have a professional caregiver (p = 0.015).

HCRU

Based on descriptive data, physicians reported agitated patients requiring more hours/week of caregiving (professional or informal/non-professional) than non-agitated patients. The number of consultations with all types of HCPs in the 12 months prior to data collection was higher for agitated than non-agitated patients, although the number of memory specialist consultations was very low for both groups of patients and did not differ between agitated and non-agitated patients. A higher proportion of agitated than non-agitated patients was institutionalized and more hospitalizations, for any cause or due to CI, were reported for patients demonstrating agitation symptoms compared with those not demonstrating agitation symptoms in the 12 months prior to data collection (Table 2).

A higher proportion of non-agitated than agitated patients were currently receiving no treatment, while a higher proportion of agitated than non-agitated patients were receiving acetylcholinesterase inhibitors, memantine, antipsychotics, antidepressants, and benzodiazepines (Table 3).

Based on regression analysis (Fig. 1), the level of agitation experienced by patients was observed to be associated with the use of a number of healthcare resources. Within the 12 months prior to data collection, as agitation score increased, so did the number of consultations with psychiatrists (p = 0.001), psycho-geriatricians (p < 0.001), and all HCPs (p < 0.001). The proportion of patients institutionalized increased with increasing agitation score (p = 0.004). The number of hospitalizations in the 12 months prior to data collection increased with increasing agitation score, whether for any reason or for CI (both p = 0.001). The likelihood of spending time on a psychiatric ward in the previous 12 months was also associated with agitation score (p = 0.035). The proportion of patients receiving no treatment decreased with increasing agitation score (p = 0.016), while the proportion of patients receiving either an antipsychotic or an antidepressant increased with increasing agitation score (both p < 0.001).

Fig. 1

Impact of agitation score on HCRU –regression analysis. CI, confidence interval; HCP, healthcare professional; IRR, incidence rate ratio; OR, odds ratio.

On propensity score analysis, the number of consultations in the past 12 months with psychiatrists (p = 0.024), psycho-geriatricians (p = 0.006) and all HCPs (p < 0.001) was higher in agitated than non-agitated patients. Propensity score analysis also indicated a higher proportion of agitated than non-agitated patients was institutionalized (p =0.019), and that the proportions of patients receiving antipsychotics (p = 0.016), antidepressants (p < 0.001), and benzodiazepines (p = 0.017) were all higher in agitated than non-agitated patients.

Healthcare costs

Mean (standard deviation) total direct healthcare costs in the 12 months prior to data collection were $9,243 ($17,725) for non-agitated patients and $20,041 ($26,495) for agitated patients (Supplementary Table 2).

Based on regression analysis (Fig. 2), the level of agitation experienced by patients was observed to be associated with the cost of hospitalizations (p =0.003), HCP consultations (p < 0.001), scans (p =0.001), and total direct healthcare costs (p < 0.001) in the 12 months prior to data collection. The costs of professional caregiving (p = 0.880), nursing homes (p = 0.528), and treatment (p = 0.266), however, were not correlated with agitation score.

Fig. 2

Impact of agitation score on healthcare costs –regression analysis. CI, confidence interval; GLM, generalized linear modeling; HCP, healthcare professional.

Propensity score sensitivity analysis indicated that the cost of professional caregiving, HCP consultations and scans in the previous 12 months were higher in agitated than non-agitated patients when agitation was defined as≥2 symptoms (all p < 0.001).

DISCUSSION

This study demonstrated the incremental impact and burden of agitation in patients with dementia on patients, HCRU, and healthcare costs using real-world US data. Patients showing symptoms of agitation were perceived by their physicians to be at a more severe stage of dementia and were more likely be diagnosed with AD rather than mild CI compared with patients without agitation symptoms. This reflects findings reported in the literature, including observational studies from the US, the Netherlands, Australia, and Japan, showing agitation to be associated with dementia severity [21–24].

There was clear evidence of agitation adding to the presence of a caregiver (professional or informal/non-professional), with the proportion of patients with a caregiver and the number of caregiving hours higher in agitated than non-agitated patients, and the proportion of patients with professional caregivers correlated with agitation score.

We also observed HCRU and healthcare costs in patients with dementia to be associated with agitation. The total number of consultations with all HCPs, psychiatrists, and psycho-geriatricians in the previous 12 months were correlated with the agitation score. The likelihood of being institutionalized was more than three times higher in agitated than non-agitated patients, and the proportion of institutionalized patients increased with increasing agitation level. The number of hospitalizations (for any cause or due to CI) in the previous 12 months was related to agitation score, as was the proportion of patients who spent time on a psychiatric ward.

There are a few published studies on the impact of agitation on HCP consultations, institutionalizations, or hospitalizations in patients with dementia. A cross-sectional study of patients in nursing homes in the Netherlands indicated that agitation might increase consultations with psychiatrists, as we observed, as agitation was reported to be the primary reason for a psychiatric consultation [25]. Agitation in adults aged≥71 years with CI was also shown to be associated with a significantly increased risk of institutionalization in a Japanese study [26].

We observed the likelihood of receiving no treatment was lower in agitated than non-agitated patients and decreased with increasing agitation score, that more agitated than non-agitated patients were receiving treatment with acetylcholinesterase inhibitors, memantine, antipsychotics, antidepressants, and benzodiazepines, and that the proportion of patients receiving an antipsychotic or antidepressant increased with increasing agitation score. There is limited evidence of efficacy and a risk of adverse effects for pharmacologic agents in the management of agitation symptoms in dementia; consequently, drug therapy should be considered only if symptoms are severely distressing for patients or caregivers [9, 27]. However, a Cochrane review reported the antidepressants sertraline and citalopram to be associated with a reduction in symptoms of agitation when compared to placebo [28]; meanwhile, brexpiprazole, an atypical antipsychotic, has been shown to have the potential to improve agitation in AD compared with placebo [29]. A sequential treatment algorithm with risperidone, aripiprazole or quetiapine, carbamazepine, citalopram, gabapentin, and prazosin has been proposed [30].

We observed increasing healthcare costs to be associated with increasing agitation. Cross-sectional data from a large patient cohort from eight European countries showed a significant difference in cost of care between dementia patients experiencing and not experiencing agitation, in both home care and institutional long-term care settings [31]. A prospective cohort study of data from UK patients with AD showed health and social care costs to vary significantly between patients with and without agitation, with the authors calculating an excess cost associated with agitation of £2 billion/year for all AD patients in the UK [32]. An analysis of a population-based sample of patients with dementia in the US also showed an increase in agitation/aggression led to increased costs of informal care [33].

Given the association of agitation with more severe/advanced dementia that we observed, and that has been reported in the literature [21–24], it might be speculated that underlying disease severity is driving the higher HCRU and healthcare cost that we noted in patients with agitation. However, as both disease severity (assessed by the MMSE) and time since diagnosis were covariates in the regression analyses, it would appear that the presence of agitation is an independent factor in the higher economic burden seen in the group of patients experiencing agitation symptoms.

Our study had a number of potential limitations. The DSP data collection included the next nine consecutively consulting patients, resulting in a pseudo-random, rather than truly random, study sample. As with all surveys of this type, the methodology relied on accurate reporting by physicians; missing data were to be expected and may have influenced results. This was a point-in-time rather than a longitudinal survey and data may have been used to assess the association between factors but not causality. The survey was restricted to physician office and hospital settings, excluding residential care facilities; more severe dementia cases might not have been included in the physicians’ consulting practices, and hence would not have been reflected in this sample. It is likely that healthcare costs varied by region, but costs were calculated based on a number of national data sources; this might have resulted in the impact of agitation on healthcare costs being under- or over-estimated. In our analysis, agitation was defined to reflect elements of the IPA consensus on the definition of agitation [7] as closely as possible given this was a retrospective analysis, and therefore utilized data available in the data source. The inclusion of additional symptoms, or the exclusion of any of the symptoms used in our analysis, might result in the re-classification of some patients from agitated to non-agitated, or vice versa. Finally, it is important to consider that the agitation scoring method gave equal weighting to all symptoms and assumed a linear relationship between mild, moderate, and severe symptom severity. While acknowledging these limitations, a substantial body of data from a representative population of almost 1,500 patients with varying degrees of dementia in the real-world setting in the US was included in the analysis, providing valuable insight into the incremental burden of agitation in dementia.

Our data demonstrated an association between agitation in patients with dementia with increased HCRU and healthcare costs. There remains a largely unmet need for effective therapeutic interventions to address neuropsychiatric symptoms, including agitation, in dementia, with the potential to alleviate patient suffering and societal costs.