Delirium in Nursing Home Residents: A Narrative Review

Delirium is an important component of the geriatric syndromes and has been recognized to negatively influence the prognosis of older people in hospital and in a post-acute setting. About 2–5% of older people world-wide live in nursing homes and are characterized by functional impairment, cognitive decline, dementia, comorbidities, and polypharmacotherapy, all factors which influence the development of delirium. However, in this setting, delirium remains often understudied. Therefore, in this narrative review, we aimed to describe the latest evidence regarding delirium screening tools, epidemiology characteristics, outcomes, risk factors, and preventions strategies in nursing homes.


Introduction
The Latin word delirare means to deviate from a straight line, and the term delirium was initially mentioned by Celsus during the first century to describe mental disorders during head trauma or fever [1]. Delirium is a complex neurocognitive disorder characterized by acute modification of attention, awareness, cognitive functions, and behavioral abnormalities caused by an underlying medical condition [2,3]. Delirium results from a combination of several specific factors such as neuroinflammation, oxidative stress, endocrine dysfunction, and circadian rhythm dysregulation, which lead to a breakdown in cerebral network connectivity and failure of interaction between processing sensory signals and motor effectors [4]. The prevalence of delirium varies considerably by setting: the overall prevalence was 23% in a medical setting [5], 35% in palliative care [6], 1-2% in the community, and up to 70% in a long-term care setting [7]. Delirium has been recognized to negatively influence the prognosis of people with acute illness, and a meta-analysis study which investigated the association of delirium with mortality, institutionalization, and dementia, in hospitalized or post-acute care people aged 65 years or older, revealed that delirium is an independent predictor of poor outcome [8]. Other important negative consequences of delirium are related to the emotional distress and burden among healthcare providers and caregivers. Indeed, a study described that the subjective burden that nurses experience when caring for people with delirium was high and the hyperactive/hyperalert subtype was the most challenging to deal with [9]. The abovementioned studies generally did not focus on nursing home residents (NHRs), and little is known about delirium in a nursing home setting.
About 2-5% of older people worldwide live in nursing home settings [10], and NHRs are characterized by functional impairment, cognitive decline, dementia, multicomorbidities, and polypharmacotherapy [11], factors which influence the development of delirium, increasing the risk of negative outcomes and the emotional distress of older people and healthcare providers. Nevertheless, delirium remains often undetected and underdiagnosed. Many reasons such as lack of education and training in healthcare providers, lack of implementation of validated detection tools, and negative attitudes hinder the detection and implementation of preventive and therapeutic measures. Hence, it is crucial to have in-depth knowledge regarding delirium in NHRs. In this narrative review, we

Delirium Prevalence and Incidence in Nursing Homes
From our review, we identified 28  articles with relevant data regarding the prevalence of delirium, and eight articles [11,26,[39][40][41][42][43][44] described the incidence of delirium. Total population mean age ranged from 76.2 years to 88.5 years, and 46.8% to 90.8% of the participants in the studies were female.

Delirium Prevalence
The prevalence of delirium ranged widely from 1.4% to 70.3%. Diagnostic Statistical Manual of Mental Disorders, third-fifth edition (DSM-III to DSM-V), Confusion Assessment Methods (CAM), Nursing Home Confusion Assessment Methods (NH-CAM), and the Neelon and Champagne Confusion Scale (NEECHAM) were the most frequently used diagnostic criteria and screening tools.
Morichi et al. [30] and Bo et al. [29] described a point prevalence of about 37% in a multicenter Italian observation study where the four As test (4AT) tool was performed for the detection of delirium. In contrast, a cross-sectional study including six nursing homes in Belgium revealed that the prevalence of delirium assessed by the delirium observational screening scale (DOSS) was no more than 15% [31]. A validation pilot study which compared the results of DSM-V criteria with Informant Assessment of Geriatric Delirium (I-AGeD) reported a prevalence of 5.9% and demonstrated that I-AGeD was suitable for the detection of delirium in Swiss NHR [34]. NHRs affected by COVID-19 disease had a high prevalence of delirium from 49.2% to 57.9%, mainly presented as hypoactive delirium [33,38].

Delirium Incidence
The incidence of delirium also ranged widely from 10% to 60%, and the main diagnostic criteria and screening tools applied were DSM-IV, DSM-V, and CAM. Two studies reported higher incidence of delirium among NHR, at 40% and 60% [40,43], compared to others included in our review. The authors suggested that high prevalence of dementia and need for continued care were possible explanations. Table 1 reports the characteristics of the abovementioned studies regarding the prevalence and the incidence of delirium in NHR. Delirium occurred frequently as a complication of acute illness, and was associated with cognitive function decline.

Demographic and Social Risk Factors
Most of the studies were in agreement about risk factors; however, some contradictions were identified. Voyer et al. [19], Zucchelli et al. [35], and Sepulveda et al. [44] reported that age was a significant risk factor for delirium onset in NHR. Other studies included in our review did not describe a significant association regarding age and delirium onset ( Table 1). Female gender was identified as a risk factor in one study [39]; in contrast, another study [27] reported that older people with delirium were less likely to be females. Widowhood was also reported as a risk factor.

Comorbidities
Most of the studies concluded that dementia was an independent risk factor with the exception of one study [44], where data analysis after correction for other risk factors failed to find that dementia was associated with delirium, probably because many people living with dementia could have been recently treated for delirium symptoms without a specific diagnosis of delirium. Parkinson's disease and depression were other comorbidities reported as risk factors. Of interest, diabetes resulted as a protective risk factor, probably because the frequent glycemic control results in closer contact between nurses and older people, which may avoid the development of delirium [41].

Malnutrition
Nutritional status and other elements that may reflect malnutrition such as low lean muscle mass, dehydration, and electrolyte imbalance have been described as significant predisposing and/or precipitating risk factors [13,15,17,25,35]. Other indicators of malnutrition such as low albumin or low protein level were higher among older people who developed delirium compared to them without but without reaching statistical significance [17].

Infections
Infections were characterized by higher odds of delirium onset [28]. Furthermore, urinary and respiratory infections were mentioned as important precipitating risk factors [25,39]. Delirium was also described in NHRs with SARS-CoV-19 infection [33]. Furthermore, a case study series study reported that, the day after COVID-19 vaccination, delirium or subsyndromal delirium of mild to moderate severity was identified in 10% of NHRs with no potential competing explanation [42].

Other Factors
Hearing deficits, visual impairment, and functional dependency resulted as significant and independent risk factors in several studies. Chair and physical restraints increased the odds for delirium development [26,44]. Falls showed a strong association with the onset of delirium in NHR [13,41]. Data from a retrospective cohort study including 1571 persons from 12 nursing homes found pain to be significantly associated with delirium [40]. Table 1 summarizes the characteristics of studies that described significant risk factors associated with delirium.

Adverse Outcomes Associated with Delirium in NHRs
An overall increased risk for mortality in NHRs which experienced delirium was reported by different studies [15,21,23,27,36]. In contrast, one study [25] did not find an association between delirium and overall mortality, but other factors such as infections, comorbidities, advanced age, and low plasma albumin level resulted as independent predictors. Rehospitalizations were higher among older people with delirium [27], and an important functional decline was reported as a consequence of delirium [21,27]. Of note, delirium was associated with long-term cognitive decline and dementia development [39]. Studies reporting information regarding the impact of delirium regarding outcomes in NHR are summarized in Table 1.

Delirium Screening Tools
The diagnosis of delirium should be guided by the standard criteria of DSM-V and International Classification of Diseases, 10th Revision (ICD-10). DSM-V criteria for delirium include acute and fluctuating disturbance of attention and awareness, disturbance in cognition, the absence of a pre-existing neurocognitive disorder, and the presence of medical conditions, withdrawal, exposure to toxins, or multiple etiologies [45]. ICD-10 criteria for delirium, similar to DSM-V, include disturbance of attention, awareness, memory deficit, rapid onset, and fluctuations of the symptoms, in addition to psychomotor deficits, sleep disturbance, or sleep/awake cycle disturbance [46]. It should be mentioned that both DSM-V and ICD-10 do not provide clear methodology regarding the evaluation of attention and awareness or the determination of pre-existing neurocognitive disorders [47]; over the last four decades, the detection of delirium has remained stable [5]. However, it has been reported that less strict DSM-V criteria regarding attention and orientation make the diagnosis of delirium more inclusive [48]. Furthermore, ICD-10 and DSM-V represent the best expert consensus and available evidence, and their application is encouraged in both clinical and research contexts [49].
During the recent years, different screening tools for the detection of delirium have been developed. Here, we report the most frequent screening tools in the context of NHRs.
CAM is a widely used standardized delirium instrument for clinical and research purposes, designed to allow nonpsychiatric healthcare providers to detect delirium accurately [50,51].
CAM is based on DSM-III revised criteria and includes the following delirium features: (a) Fluctuating course, (b) Inattention, (c) Disorganized thinking, (d) Altered level of consciousness.
A further algorithm is applied for the classification of three delirium levels: (1) Subsyndromal delirium with only one of the four features present, (2) Subsyndromal delirium with two of the four features are present, (3) Full delirium: features a and b are present, along with features c and/or d [50].
At least 12 studies [11,15,17,18,25,27,28,[39][40][41][42][43] in the context of nursing homes applied CAM, from 2003 to 2022 [52]. It should be mentioned that the prevalence of delirium still ranged widely, but CAM was helpful in the prediction of outcome (Table 1). This instrument has been validated from different studies applying DSM-III to IV, ICD-10, or consensus diagnosis including psychiatrist, geriatricians, and advanced practice nurses, with an overall sensitivity of 82% ranging from 69% to 91%, and a specificity of 99% ranging from 87% to 100% [52]. Of note, in the nursing home setting, the validity of CAM remains understudied.
NH-CAM is based on the four CAM features, which are modified using variables and items from the Minimum Dataset (MDS) Resident Assessment Protocol (RAP). NH-CAM includes the following features [36]:

1.
Mental function varies over the course of the day or mood decline over the last 90 days; 2.
Periods of altered perception or awareness of surrounding or episodes of disorganized speech or cognitive decline over the last 90 days; 4.
Periods of restlessness or periods of lethargy or behavior decline over the last 90 days.
Subsyndromal delirium level 1 is defined if any of the features is present, level 2 if any two of the four features are present, and full delirium if features 1 and 2 are present together with features 3 or 4.
Dosa et al., in a large cohort of 35,721, concluded that NH-CAM is a useful tool in estimating delirium related prognosis among NHR [36]. Other studies estimated the prevalence, incidence, and risk factors of delirium using NH-CAM or an adapted version to their population [22,26]. However, a clinical validation of NH-CAM is still lacking and future studies should evaluate if application of NH-CAM may result in the improvement of care and outcome in NHR.
NEECHAM is a nine-item instrument based on daily nursing practice, which evaluates the level of processing information, the level of behavior, and the physiological condition [53]. Compared to DSM-IV, NEECHAM covers 13 of the criteria for delirium with a total score that ranges from 0 to 30. A score of 24 or less indicates possible delirium. It has been demonstrated that NEECHAM scale is reliable for the detection of delirium by nurses in long-term care [54,55]. Regular daily administration may be difficult or burdensome; however, NEECHAM may be administrated when an acute change in mental status is present, as well as for a more in-depth assessment of delirium.
The 4AT is a brief screening tool including four items: 1. Alertness, 2.
Acute change or fluctuating course.
Its score ranges from 0 to 12 points, where a score ≥ 4 suggests possible delirium. The 4AT tool is fast and does not require specific training for the implementation. It was initially validated and used in hospital settings and [56], and further large studies applied the 4AT tool in NHR settings [29,30]. A point prevalence analysis revealed that almost one to three persons in nursing homes are affected by delirium [30], and the 4AT score was not associated with the use of indwelling urinary catheters in NHRs [29]. It should be mentioned that a score of 4 points may be present in chronic cognitive impairment and cases where delirium is superimposed on dementia cannot be excluded. However, the adverse consequences of delirium non-detection are greater than delirium overdiagnosis.
The Delirium Diagnostic Tool-Provisional (DDT-Pro) is a brief scale designed to allow accurate delirium diagnosis by evaluating vigilance, comprehension, and sleep/awake cycle. The scores range from 0 to 9 points, and the recommended cutoff for delirium is a score ≤ 6. It has been reported that DDT-Pro sensitivity ranges from 88% to 100% and specificity ranges from 85.3% to 94.4% [57]. Data from a recent study reported that DDT-Pro is a valid tool to detect delirium when used by skilled nurses showing 77.2% sensitivity and 84% specificity; for a cutoff score ≤ 7, DDT-Pro sensitivity increased to 84.8% [37].

Prevention Strategies for Delirium in Nursing Home Settings
Dehydration was significantly associated with delirium onset [15,20,25], and persons in nursing homes may be highly vulnerable to dehydration due to the presence of other factors such as swallowing difficulties, lack of thirst, cognitive impairment, use of restraints, or substandard care. However, a randomized study where weight-based intervention of hydration was performed did not show reduction in delirium incidence during 4 weeks follow-up [58]. It should be mentioned that hydration status was controlled by bioelectrical impedance analysis, and older people with cardiovascular disease and diuretic therapy should be frequently monitored. In addition, hydration is one among the multiple risk factors which influence delirium, and modification of only one factor may not be sufficient for significant changes.
Despite the multifactorial characteristics of delirium, a multicomponent 16 month enhanced educational package to support care home staff to address key delirium risk factors did not show effectiveness in nursing home settings [59].
Furthermore, another multicomponent intervention study, targeting delirium risk factors (e.g., cognition, immobility, hydration, and undernutrition) in acutely ill long-term nursing home residents did not show significant results in the prevention of delirium. Differences regarding cognitive impairment prevalence at baseline between intervention and usual care group, resistance to some interventions because of physical impairment and disability, and lack of a formal interdisciplinary team were reported as possible explanations of the results [60].
The Transfusion Requirements in Frail Elderly (TRIFE) study was a prospective, assessor-blinded, randomized controlled trial focused on the role of red blood cell transfusion strategies on physical recovery or survival in frail older people with hip fracture after surgery. A post hoc analysis of this study, including NHR with hip fracture, revealed that liberal transfusion (Hb levels ≥ 11.3 g/dL) compared to the restrictive transfusion strategy (Hb levels ≥ 9.7 g/dL) reduced the occurrence of delirium. Results from this study indicated that, in older people with hip fractures, maintaining hemoglobin level above 11.3 g/dL reduced the rate of delirium [61].
Application of a software which correlates medication effects with physical, functional, and cognitive decline to foster early recognition of potential adverse drug effects showed that newly admitted NHRs experienced a lower rate of potential delirium compared to usual care NHR [62]. Another study focused on the implementation of an educational program directed toward nursing home physicians in reducing inappropriate prescription and improving health outcomes revealed that this intervention improved the inappropriate prescription of drugs and the occurrence of delirium [63].
A recent randomized control trial reported that doll therapy was more effective in reducing agitation, aggressiveness, professional caregiver burden, and the incidence of delirium compared to standard treatment in people living with dementia [64].
Of interest, other nonpharmacological interventions such as music therapy in people living with dementia improved the psycho-behavioral profile [65], and bright-light therapy was effective in reducing daytime sleep in NHRs with dementia [66]. Future studies may consider exploring the application of these interventions as a preventive strategy regarding the occurrence of delirium in nursing home settings. Table 2 summarizes the characteristics of studies that explored interventions and prevention strategies for delirium in NHR. Delirium increased in the control group (from 0.04 to 0.14 per resident, p = 001) but was reduced significantly in the intervention group (0.08 at baseline, 0.03 at the end of the study p = 0.035)

Discussion
Delirium is a common condition in the healthcare system with important health and socioeconomic costs. The majority of delirium research has been focused on the hospital setting, and NHRs remain an understudied field. According to a previous review which analyzed data from 15 studies, the prevalence of delirium in long-term care was between 1.4% and 70% [7]. However, long-term care has a broad definition and does not include only NHRs. In addition, since then, other studies focused on delirium have been published. In this review, we initially summarized the evidence referring to the prevalence and incidence of delirium in NHR. Although we focused on NHR and reviewed data from a relevant number of studies, the prevalence of delirium in NHR still ranged widely from 1.4% to 70.3%, while the incidence ranged between 10% and 60%. Differences related to diagnostic criteria, screening tools, baseline population characteristics such as the prevalence of dementia, and acute events are possible explanations for the wide range of prevalence and incidence.
Recognition of delirium may be difficult when dementia is present because of the overlapping of clinical characteristics such as hypoactivity and fluctuations of symptoms. Moreover, delirium superimposed on dementia is common in aged populations, and the present diagnostic criteria and screening tools provide a suboptimal performance [67]. Dementia is mainly characterized by memory deficits and characteristics of delirium such as alternation of attention, language, motor function, and sleep/awake cycle may also be present in dementia. However, these alternations in dementia without delirium are less accentuated and are characterized by an insidious and progressive onset [68][69][70]. It should be mentioned that both ICD-10 and DSM-V do not provide specific indications when dementia coexists.
Age, dementia, depression, and restraints were risk factors mentioned from different studies conducted in NHRs [71]. From our review a variety of demographic and social risk factors, comorbidities, conditions related to aging and diseases, malnutrition, and drugs were identified as risk factors for delirium onset in NHR. Age, dementia, physical restraints, and falls are common and significant risk factors in nursing homes. Not all the studies conducted in NHRs reported age as a significant risk factor. The majority of studies in nursing homes included populations at advanced age, with a more homogeneous age distribution compared to other settings. In NHRs, the severity of dementia was measured using different scales such as the Mini-Mental State Examination or Montreal Cognitive Assessment Scale. However, dementia resulted as an independent risk factor in most of the NHR studies. This result is line with other studies in different settings [72]. The association of falls and restrains with delirium resulted significant after adjustment for confounders.
Furthermore, NHRs with COVID-19 infection experienced a higher prevalence rate of delirium ranging from 49.2% to 57.9% compared to hospitalized people aged over 65 years, where an overall prevalence of 28% was reported [73]. Clinical presentation of COVID-19 disease may be masked by delirium, and SARS-CoV-2 infection should be considered in people experiencing delirium. In addition, cases of delirium after COVID-19 vaccination have been reported. It should be mentioned that delirium resolved without complications [42]. However, healthcare providers should monitor for delirium after vaccination in this population.
Delirium experienced in NHR is associated with poor outcome regarding survival, rehospitalizations, and functional decline. In other settings, delirium experienced during hospitalization also increases likelihood of nursing home placement and is associated with about twofold increased mortality risk within 2 years [8,74]. In persons newly admitted into post-acute care facilities from acute care hospitals, delirium is associated with worse activities of daily living (ADL) and instrumental activities of daily living (IADL) recovery, indicating that delirium may have important consequences for functional performance and dependence [75]. Of note, the relationship between delirium and dementia seems bidirectional; delirium may be associated with long-term cognitive decline and dementia development [76]. CAM, NH-CAM, and NEECHAM are the most frequent screening tools applied in NHR. Despite their confirmed validity in different settings, the clinical validity of these tools in nursing homes is still understudied. Furthermore, routine use of delirium screening tools in nursing homes is not common [77]. It should be mentioned that screening tools may be time-consuming and not always easy to use; for these reasons, an "ideal" delirium screening tool in the context of NHR should be fast, easy to use, and able to be integrated in everyday work. Recently, new screening tools such as DDT-Pro, 4AT, and I-AGeD have shown good performance in nursing homes, and further validation studies should be performed with larger populations. In addition, the impact of these tools in the prognosis and management of NHR should be evaluated in-depth.
Previous studies have demonstrated that interventions for delirium prevention are highly effective in decreasing occurrence of delirium, with more than 50% odds reduction, and approximately one million cases of delirium in the hospital settings could be preventable by multicomponent nonpharmacologic interventions [78]. However, limited evidence has been identified on interventions for preventing delirium in a long-term care setting [79]. Hydration, multicomponent interventions, and educational strategies neither reduced the incidence of delirium nor influenced the outcome [58][59][60]. Application of software and educational programs which are helpful in the identification of possible adverse drug effects seems promising for the reduction in delirium incidence [62,63]. In older people with specific conditions such as hip fracture, transfusion resulted in a reduction in delirium rate [61]. Furthermore, other interventions focused on psychologic behaviors such as doll, bright-light, and music therapy should be further investigated regarding their role in reducing the incidence of delirium.
The identification and the correction of life-threatening causal factors and the underlying causes of delirium are crucial once delirium has been detected. Adaptation of nonpharmacological interventions such as avoiding unnecessary catheters and physical restraints, approaching people at the bedside, and tools to facilitate orientation may be useful to improve mild-to-moderate agitation. It is also recommended to avoid poor communication and inadequate staff attention to promote adequate nutrition and fluid intake. Exclusion of pain, constipation, and urinary retention should be considered as important causes of agitation in older people. Several guidelines offer practical recommendations regarding delirium management, as well as indications regarding antipsychotic use [80,81].
This study had some limitations; we provided a review of the published studies and not a systematic review. It is possible that other studies could provide important information regarding delirium epidemiology, risk factors, outcome, and prevention in NHR. Furthermore, being a narrative review, the quality of studies was not evaluated.

Conclusions
The prevalence and the incidence of delirium in nursing homes are frequent but range widely. Different diagnostic criteria, a variety of detection tools, and delirium superimposed on dementia may in part explain the wide range of delirium occurrence in NHR. A considerable number of risk factors associated with delirium onset have been described including demographic, social, and nutritional factors, as well as comorbidities and therapy. In addition to well-known risk factors such as age and dementia, physical restraints and falls also need to be considered as potential risk factors in nursing home settings. SARS-CoV-2 infection should be considered in NHRs with delirium, since delirium may mask COVID-19 disease. Different delirium detection tools have been developed; however, the clinical validity of most of them is still understudied in nursing homes. Prevention strategies such as software interventions to monitor therapy, educational programs applied among healthcare providers, delirium routine screening by nurse staff, and other interventions regarding anemia, behavioral, and psychological symptoms give promising results; however, further studies should investigate their effectiveness in reducing delirium incidence in NHR.
Author Contributions: Conceptualization, K.K. and C.F.; methodology, K.K. and G.G.; investigation, C.F. and F.A.; resources, K.K.; writing-original draft preparation, K.K., G.G., F.A. and C.F. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.