Psychobiological Monitoring of a Home-Based Dyadic Intervention for People Living with Dementia and Their Caregivers: Added Value to Evaluate Treatment Success and Understand Underlying Mechanisms

Abstract

Background:

Research concerning people living with dementia (PwD) and their informal caregivers (ICs) has recently begun to focus on dyadic aspects of psychosocial interventions.

Objective:

We adapted a dyadic psychosocial intervention and examined its effects on psychobiological stress in daily life.

Methods:

Twenty-four PwD-caregiver dyads were visited seven times at home by specialized nursing staff. Momentary subjective stress, salivary cortisol (sCort), and salivary alpha-amylase (sAA) were measured in PwD and ICs before and after each home visit as well as six times per day at two days each at the beginning and end of the intervention as part of an ambulatory assessment. Hair cortisol concentrations (HCC) were measured twice.

Results:

After each home visit session, ICs reported lower subjective stress. sCort was lower in both ICs and PwD, whereas sAA did not change. In daily life, area under the curve (AUCg) concerning sCort secretion indicated that PwD had lower sCort daily output at the end of the intervention, and AUCg concerning subjective stress indicated that both PwD and ICs reported lower subjective stress than at the beginning of the intervention. AUCg concerning sAA did not change over time in either group. HCC did not vary over time but increased with disease severity.

Conclusion:

The psychosocial intervention reduced psychobiological stress but affected psychobiological stress measures differently in PwD and ICs. In particular, the discrepancy between subjective and physiological markers of stress in PwD emphasizes the added value to evaluate treatment success and understand underlying mechanisms as a complement to self-reports.

Keywords

Cognitive-behavioral therapy coping dementia care home-based resilience

INTRODUCTION

People living with dementia (PwD) as well as their informal caregivers (ICs) are often exposed to chronic stress in their daily life. Chronic stress has deleterious effects on health, mediated by changes in stress-sensitive systems in the body [1], particularly increases in the activity of the hypothalamic-pituitary-adrenal (HPA) axis and dysfunctions in autonomic nervous system (ANS) activity. In this regard, both PwD [2] and their ICs [3] often show increased cortisol secretion. Moreover, autonomic dysfunctions have been observed in PwD [4] and ICs [5], and chronic stress is associated with adverse health effects in both.

Psychobiological stress exacerbates dementia pathology in PwD

Studies have recently begun to examine the feasibility of physiological stress markers in PwD due to probable disease-specific difficulties in self-reporting [6, 7]. There is mounting evidence that dysregulations in HPA axis activity are closely linked to disease onset, progression, and functioning in the daily life of PwD [8]. Such dysregulations in HPA axis activity have been reported cross-sectionally and longitudinally, and from both a short-term and long-term perspective. In this regard, higher cortisol secretion is associated with earlier disease onset and faster progression of dementia, presumably mediated by hippocampal degeneration in PwD [2]. Regarding short-term effects, daily profiles of cortisol have been associated with disrupted daily activity/rest cycles [9] and with behavioral and psychological symptoms of dementia (BPSD) [10]. Regarding long-term effects, chronic stress, as measured by hair cortisol concentrations (HCC), has been associated with poorer outcomes in PwD [11], and increases in HCC have been associated with disease progression [12]. Moreover, dysregulations in ANS activity, as mirrored by alpha-amylase activity, have been linked to the pathology of Alzheimer’s disease on a neural level, potentially mirroring reduced glucose metabolism [13].

Caregiver burden: ICs as a model for chronic stress

Caregiving is often conceptualized as a model for chronic stress [14]. Following a diathesis-stress model, stress is particularly harmful as it leads in combination with prior vulnerabilities to adverse health effects [15]. Reviews consistently point to increased cortisol secretion in ICs of PwD [3]. Both short-term and long-term dysregulations of HPA axis activity are reported, with ICs of PwD who present BPSD being at particular risk of pathological disruptions to daily profiles of salivary cortisol (sCort) [16]. Compared to non-caregiver controls, higher HCC were found in ICs of PwD, which were associated with depressive symptoms [17]. Overall, ICs of PwD show increased rates of depressive symptoms [18]. Furthermore, dysregulations in ANS activity are associated with characteristics of the caregiving situation. For instance, diurnal profiles of salivary alpha-amylase (sAA) were found to be beneficially affected by the use of adult day services [19].

Thus, the evidence consistently shows that from a mechanistic perspective, the activity in stress-sensitive systems of the body is closely related to the onset, manifestation and progression of dementia and dementia-related burden. As such, it is necessary to find interventions that reduce stress and target these psychobiological stress markers in both PwD and their ICs.

How caregivers cope with and adapt to the disease-related changes in daily life might set a switch in daily life to decrease stress and maintain health. Historically, most psychosocial interventions addressed either the PwD or the IC individually [20, 21]. Although these interventions often report beneficial effects for ICs [22] and PwD [23] separately, it is also hypothesized that these benefits might increase if both ICs and PwD are targeted together [24]. Particularly as interpersonal concepts such as social support and relationship quality are both important to PwD and ICs [25] and are generally associated with stress-reducing effects [26], dyadic interventions might form an avenue to reduce stress and increase coping in both PwD and their ICs.

Dyadic interventions for PwD and their ICs

Given that both PwD and their ICs are at increased risk of stress-related adverse health effects, research has recently begun to focus on dyadic aspects of psychosocial interventions in dementia, as dyadic factors have already been identified as highly relevant modulators of psychobiological stress responses in couples [24]. In healthy couples, there is evidence for coregulation of cortisol in daily life with cortisol levels within dyads being positively associated with each other [27]. In older couples, Pauly et al. [28] demonstrated that this synchrony increased with positive partner interactions. By means of a support intervention, Holt-Lunstad et al. [29] reported beneficial effects on sCort and sAA in daily life in couples.

As most PwD are cared for by ICs at home, the caregiver-PwD dyad deserves special attention. As a dyadic interplay is at work between ICs and PwD, with increases in stress endangering the stability of the care situation, a greater synergy might be expected when targeting both PwD and ICs for stress reduction purposes. This led to the rise of dyadic psychosocial interventions, as it is now acknowledged that dyadic aspects might form an important and additional layer in understanding dementia-related burden within such dyads [24, 30].

Dyadic interventions share the rationale of addressing both members of a dyad on eye level. Thus, in addition to intrapersonal benefits, advantageous effects on the interpersonal level might be expected [24]. As social support [31], high partnership quality [32] and successful dyadic coping [33] are associated with stress-buffering effects in PwD-caregiver dyads, dyadic interventions might be particularly promising in terms of reducing psychobiological stress and facilitating coping with dementia-related changes in daily life [24]. Furthermore, as most of the stress occurs in the daily life of PwD and ICs, it is plausible to assume that delivering a home-based intervention within the comfort zone of the dyads should have beneficial effects on the dyad too. This is particularly relevant in light of Fonareva et al.’s [34] finding that ratings of perceived stress differed between the research environment and the home environment of ICs.

Hypotheses

Therefore, the present study aims to investigate PwD and their ICs in a naturalistic context by combining home visits (intervention), ambulatory assessment (daily output), and long-term measures (HCC). First, we hypothesized that the intervention immediately reduces psychobiological stress (self-report, sCort, sAA) in both PwD and their ICs from beginning to end of each home visit. Second, we hypothesized that the intervention leads to intermediate and long-term stress-reducing effects in psychobiological stress markers in the daily life of both PwD and ICs as measured by sCort and sAA as part of an ambulatory assessment (daily output) and by cumulative cortisol secretion (HCC).

METHOD

Participants

All dyads were recruited from the outpatient memory clinic of Rheinhessen-Fachklinik Alzey, a rural area in Rhineland-Palatine, Germany. All participants were diagnosed with dementia according to guidelines [35]. A total of 70 dyads were screened for eligibility. Due to strict inclusion criteria, only 26 dyads were able to take part in the psychosocial intervention, with 24 dyads ultimately completing the intervention. Of these 24 dyads, 20 dyads were couples who had been married for 42.5±13.8 years on average. Spousal caregivers were predominantly female, with 16 female caregivers caring for a male PwD. All four of the child-parent dyads consisted of daughters caring for their mother.

Overall, caregivers (20 female) were 67.8±10.7 (range: 41– 85) years old, and eight caregivers were in employment, working for 22.6±17.7 hours per week. All PwD (16 male, age: 75.7±7.3 (range: 57– 87) years) had had a diagnosis of dementia for 19±19 months, and 22 had been diagnosed with Alzheimer’s disease dementia. The mean Mini-Mental State Examination (MMSE) sum score was 22±3 points (range 14– 26); 18 PwD scored in the range of mild dementia and six scored in the range of moderate dementia. The mean body-mass-index (BMI) was 24.71±3.86 (range 18.82 to 36.06) in the total sample.

The eligibility criteria for ICs were: ≥18 years of age; primary informal caregiver; cohabitation with the PwD; no intake of medication with immunological, autonomic or neuroendocrinological effects; no severe psychiatric or somatic disorder; no current psychotherapeutic or sociotherapeutic treatment; fewer than five cigarettes per week; fluency in the German language; hair length of at least 2 cm.

The eligibility criteria for PwD were:≥18 years of age; diagnosis of dementia; MMSE sum score in the range 26– 10; no intake of medication with immunological, autonomic or neuroendocrinological effects (with the exception of drugs to treat dementia and BPSD); if drugs to treat dementia or BPSD were taken: no changes in doses in last three months; fewer than five cigarettes per week; sufficient mastery of the German language; hair length of at least 2 cm.

Procedure

After completing standard diagnostic procedures and education on dementia diagnosis in the outpatient memory clinic, participants were invited to take part in a scientific evaluation of a home-based dyadic psychosocial intervention. Interested participants were informed about the study prior to providing informed consent. Participation was entirely voluntary and did not affect treatment in the memory clinic. The study protocol was approved by the ethics committee of the Landesärztekammer Rheinland-Pfalz (allowance number 837.198.17 (1103)).

At the beginning of the study, baseline questionnaires were completed and hair samples for the retrospective assessment of cumulative cortisol secretion of the past two months were collected from both PwD and ICs. Then, the first of a total of seven home visits was scheduled within the next seven days. The total number of home visits was scheduled to take part within 12 weeks. The time of day for each home visit was kept constant within each dyad thus automatically controlling for time of day in the analyses. Each home visit was scheduled with at least 1 h distance to the last meal and dyads were asked to drink only water or unsugared tea 1 h prior to the home visit. Additionally, dyads were asked to refrain from intensive physical activity (e.g., sports) in the hour prior to the home visit. Home visits were scheduled between 0900 h and 1600 h, with the majority of dyads being visited at either 1000 h, 1100 h, or 1400 h. Before and after each home visit, PwD and ICs were asked to rate their momentary subjective stress level on a visual analogue scale (VAS) and to provide a saliva sample for the later analysis of sCort and sAA. Additionally, dyads participated in a two-day ambulatory assessment at the beginning (directly after the first home visit) and end (directly before the final home visit) of the intervention in order to study daily output of the dyads. For the ambulatory assessment, both PwD and ICs provided six saliva samples over the course of the day (directly after awakening, 30 min after awakening, 1000 h, 1400 h, 1800 h, and 2100 h) on two consecutive days in addition to a rating of subjective stress levels at each of these time points. Furthermore, after the final home visit, participants completed questionnaires and provided a second hair sample. For all subjective ratings, the members of the dyads were asked to complete the scales individually without knowing what their partner rated. In the case of comprehension problems for the PwD, study personnel were available to assist in completing the questionnaires. ICs were specifically asked to refrain from assisting the PwD in completing the questionnaires in order to avoid biases. In Fig. 1, the study protocol is presented.

Fig. 1

Study design and elements of the psychosocial dyadic intervention adapted from Nordheim et al. [36]. PwD, people living with dementia; IC, informal caregiver; MMSE, Mini-Mental State Examination; NPI, Neuropsychiatric Inventory, 2 days ambulatory assessment: six saliva samples per day (directly after awakening, 30 min after awakening, 1000 h, 1400 h, 1800 h, and 2100 h) for the analysis of diurnal profiles and daily output of salivary cortisol and salivary alpha-amylase in addition to subjective stress levels, hair sample: for the analysis of hair cortisol concentrations.

We adapted the DYADEM manual [36] for this intervention study. The DYADEM intervention is a psychosocial, dyadic intervention for couples coping with dementia. Originally, the DYADEM trial is led by a psychotherapist and a social worker. It comprises a total of seven home visits in addition to two telephone calls. Each home visit is scheduled for a total of 120 min and the telephone calls are scheduled for 30 min. The first home visit is done by the social worker and the psychotherapist together, whereas the ensuing home visits are done by either the social worker or the psychotherapists depending on the content.

Two major changes are part of our adaptation: First, we decided to deliver the intervention by specialized nursing staff instead of a psychologist and social worker. There are two main reasons for this decision: On the one hand, we wanted to guarantee personal continuity, namely that there is one person that continually supports the dyad rather than two different persons. On the other hand, concerning the feasibility in this rural area, we considered it more feasible to have specialized nursing staff being supervised by a psychotherapist to deliver the home visits in person, as the catchment area of our intervention was in a rural area partly covering more than 50 km one way to the PwD and their ICs.

Second, we shortened the content of each session to a maximum of 50 min. This had practical reasons as we did not consider the attention span of PwD and their ICs sufficient for 120 min. Therefore, we shortened the content that each home visit focused on one theme and one intervention. In order to successfully limit the time of each home visit, our nursing staff was trained to be more direct in the structure of each session rather than to use many open-ended questions (especially in the opening of each session). We therefore follow the recommendations as described in Linnemann and Fellgiebel [37] who review adaptations of psychotherapeutic techniques and particularly recommend a more direct structuring of sessions by therapists for patients with cognitive impairments and dementia.

From a mechanistic point of view, the intervention includes psychoeducational elements that aim at increasing knowledge, skills, and coping with dementia-related changes in daily life. Particularly, psychoeducation programs are associated with stress-buffering effects in ICs that indirectly benefit the PwD as well [38]. Additionally, the DYADEM manual offers the innovation to deliver such a psychoeducational intervention to both members of the dyad on eye level acknowledging the particular role interpersonal factors play in stress and coping. Therefore, we particularly instructed the nursing staff to reinforce interpersonal resources and strengthen the relationship of the dyad whenever possible.

In Table 1 we present the title for each home visit together with the exercise/intervention that was done during that respective home visit. In Supplementary Table 1, we provide a detailed overview on the content of each home visit and telephone call accompanied by a description of the aims and major adaptations.

Table 1

Overview of each session’s content in adapted version of DYADEM Intervention

Home Visit #	Topic	Exercise/Intervention
1	Getting to Know the Dyad	Couple anamnesis
		Short recapitulation of psychoeducation dementia
2	Social Network &Pleasant Activity Scheduling	Social network map
		Pleasant activity scheduling (dyadic perspective: individual versus common activities)
3	Stress Management	Psychoeducation stress
		Psychoeducation relaxation techniques positive diary
4	Communication	Referral to positive diary
		Psychoeducation common communicative challenges
		Analysis of one difficult situation
5	Telephone Call	Common evaluation of ongoing intervention
6	Problem Solving	Referral to positive diary
		Selection of a problem in daily life that concerns both
		Introduction of 5-step problem solving approach
7	Everyday Practical Coping	Referral to positive diary
		Referral to pleasant activities and social contacts
		Compensation strategies for memory deficits in daily life; exploration of home for optimization
8	Telephone Call	Common evaluation of ongoing intervention
9	Conclusion	Looking back and forwards
		Reinforcement of dyad’s resources

Intervention

Measures

The analysis reported in this paper is part of a larger study on the evaluation and implementation of a home-delivered psychosocial intervention. Overall, there were three assessment periods (beginning of intervention, end of intervention, six months after intervention) in addition to the home visits and two-day ambulatory assessment. PwD and caregivers were asked to complete manifold questionnaires and scales repeatedly over time. However, to address the hypotheses in the following analyses, only the assessment periods which included psychobiological measures (beginning and end of intervention) are included. Furthermore, only questionnaires concerning the aforementioned hypotheses are presented.

Psychobiological stress

Stress was assessed using a multidimensional approach encompassing both subjective and physiological markers of stress.

Momentary subjective stress: Subjective momentary stress was measured using a one-item-approach in order to keep the burden for the participants to a minimum and in order to be as parsimonious as possible concerning the number of items. The VAS was printed on a sheet of paper with a length of 10 cm and ranged from 0 (‘not at all’) to 100 (‘entirely’). Before and after each home visit, both members of the dyad were asked during the collection of the saliva sample to mark the position of the VAS with a pencil that best represents their current stress level. During the ambulatory assessment at the beginning and end of the intervention, both members of the dyad rated their subjective stress level using this VAS six times per day (directly after awakening, 30 min after awakening, 1000 h, 1400 h, 1800 h, and 2100 h) on two consecutive days. The use of this single-item approach on stress in combination with the collection of saliva samples was done in accordance to other ambulatory assessment protocols from our work group [39 –41]. The only difference to these protocols is, that we printed the VAS and provided our participants with these sheets of paper, whereas in the ambulatory assessment studies the VAS was presented on a mobile diary device.

Physiological stress markers: The saliva samples were analyzed for sCort and sAA. The passive drool method was used for the collection of saliva samples: Participants were asked to not swallow for 2 min and then transfer all accumulated saliva into pre-labeled vials using SaliCaps (IBL International, a Tecan Group company, Hamburg, Germany). In order to check for compliance during the ambulatory assessment, participants had to fill in protocols detailing the code of the vial they used. Furthermore, they were asked to provide information on the consumption of food, drinks, and their level of physical activity two hours prior to saliva collection in order to control for confounding variables. However, these variables have > 50% of missing values and therefore we did not include these as control variables into our analyses. The saliva samples that were taken before and after each home visit were stored at – 80°C in freezers of the University Medical Center Mainz immediately after each home visit. The saliva samples for the daily outputs of sCort and sAA (ambulatory assessment) were stored as cool as possible in the participants’ fridge or freezer and subsequently stored, at the earliest convenience, at – 80°C in freezers of the University Medical Center Mainz. sCort levels were measured using a commercially available enzyme-linked immunoassay (IBL International, a Tecan Group company, Hamburg, Germany). sAA activity was measured using a kinetic colorimetric test and reagents obtained from DiaSys Diagnostic Systems (Holzheim, Germany). Interassay variance was 19.1% for sAA and 11.9% for sCort, and intraassay variance was 10.9% for sAA and 1.9% for sCort.

For the assessment of HCC, several thin hair strands were cut as close as possible to the scalp from the posterior vertex region of the head. To determine HCC, the first scalp-near 2 cm segment was used, which is thought to reflect the cumulative cortisol secretion of the past two months [42]. Hair-washing and cortisol extraction procedures were based on the laboratory protocol by Stalder et al. [43], with minor modifications. In brief, hair samples were washed twice for 3 min using 3 mL isopropanol. For cortisol extraction, 10±0.5 mg whole, finely cut hair was incubated in 1.8 mL methanol for 18 h at room temperature. After incubation, 1.6 mL was transferred into a glass vial. Then, 1.6 mL of the supernatant was evaporated at 50°C until samples were completely dried. Finally, the samples were resuspended with 225 μL HPLC ultra-pure water and vortexed for 20 s. A commercially available cortisol luminescence immunoassay was used for cortisol determination (LIA; IBL International, a Tecan Group company, Hamburg, Germany). Inter-assay variance was 7.64% and intra-assay was 1.38% for HCC.

Control variables

Concerning psychobiological stress measures, we added age, gender, and BMI as methodological control variables in all models due to their known association with HCC [44], sCort secretion [45, 46], and activity of sAA. From a content-based point of view, we included the following: we calculated the sum score of the MMSE [47], which is a common short test allowing the cognitive state of the patient to be graded. Additionally, we included the sum score of the Neuropsychiatric Inventory (NPI) [48] depicting BPSD. These variables were included as control variables in addition to form of relationship due to their known associations with caregiver burden [49].

Data analysis

Due to the nested structure of the dyadic data set, the data were analyzed by means of panel data econometrics using R, either using fixed effects or random effects specification. We used the plm package, as this provides efficient, consistent, and unbiased estimates based on ‘lm’ function run on transformed data [50]. For each hypothesis, we consider two-level models with timepoints (level 1) being nested in in dyads (level 2). We use fixed effects and random effects on the dyad level. As there is a 1-1 correspondence of individuals to dyads, we refrain from modelling three-level fixed effects models that would further distinguish a level for person in order to be as parsimonious as possible. All hypotheses were first modeled using a fixed effects model. Control variables were then included by using the random effects model specifications. B-values and robust standard errors are reported. The significance level α was set at 0.05. Fixed and random effects models were compared using a Hausman test, which was non-significant for all reported models.

With regard to hypothesis 1, immediate effects of each home visit on stress were modeled by predicting psychobiological stress measures (VAS, secretion of sCort in nmol/l and activity of sAA in U/ml) as dependent variable by time (before versus after each home visit) as independent variable. A positive b-value of the time variable thus indicates an increase in stress at the end of home visits whereas a negative b-values of the time variable indicates the hypothesized decrease in stress at the end of home visits. We estimated this in both PwD and caregivers while controlling for confounding variables according to the following models:

Fixed effects model:

Stress_ijt = a_i + Person_ij + Person_ij x time_ijt +

Session_ijt + age_ij + gender_ij + BMI_ij + E_ijt

Random effects model:

Stress_ijt = Person_ij + Person_ij x time_ijt + Session_ijt +

age_ij + gender_ij + BMI_ij + MMSE_ij +

relationship_ij + NPI total score_ij + E_i + E_ijt

where Person = 0 IC, 1 PwD; time = 0 before each home visit, 1 after each home visit, session = 0– 6 for each of the seven home visits, gender = 0 female, 1 male, BMI = body-mass-index, MMSE = baseline MMSE of PwD, relationship = 0 spousal dyad, 1 child-parent dyad, NPI total score = baseline sum score of NPI, a = intercept, i = index for dyads, j = index for person, t =index for time, E = error.

With regard to hypothesis 2, intermediate effects of the intervention were modeled by comparing characteristics of the daily output (ambulatory assessment) at the beginning of the intervention with characteristics of the daily output at the end of the intervention. The area under the curve with respect to ground (AUC_g) was calculated individually for each of the four days in accordance with the formula provided by Pruessner et al. [51]. All six time points per day (directly after awakening, 30 min after awakening, 1000 h, 1400 h, 1800 h, and 2100 h) were considered for calculation of AUCg. Then, the AUCg value (either for subjective stress levels, cortisol secretion, or alpha-amylase activity) was set as dependent variable. By including time with 0 = beginning of intervention and 1 = end of intervention and by including day with 0 = day 1 and 1 = day 2 as predictor variables, intervention-dependent changes in AUCg were modelled. HCC in pg/mg were compared between the beginning and end of the intervention as a marker of long-term effects of the intervention.

Missing values were not replaced but excluded pairwise in all analyses. Values for sAA were considered in the range of 4– 1440 U/ml, with values below or above this range being excluded from the analyses (52 out of 1,824 data points). Due to non-normality, values for sCort, sAA, and HCC were logarithmized using the natural logarithm ln.

RESULTS

Hypothesis 1: The dyadic psychosocial intervention immediately reduces stress in both PwD and their ICs

With regard to subjective stress levels (VAS), overall self-reported stress decreased with increasing number of home visits (b = – 1.61, p_{fixed effects model} =0.002; b = – 1.62, p_{random effects model} = 0.001). ICs reported lower subjective stress levels at the end of each home visit (b = – 10.16, p = 0.001), whereas for PwD, the subjective stress level remained the same from beginning to end of the home visit (b = – 3.03, p > 0.05). Regarding the secretion of sCort, both ICs (b = – 0.12, p = 0.02) and PwD (b = – 0.25, p≤0.001) showed lower sCort secretion at the end of each session. Higher NPI sum score was associated with lower sCort secretion (b = – 0.01, p = 0.03). Regarding the activity of sAA, there were no significant associations with time and person (all p > 0.05). In both the fix and random effects model, none of the control variables were associated with sAA activity (all p > 0.05) except for age (b = 0.04, p = 0.006) and MMSE (b = 0.04, p = 0.04) being positively associated with the activity of sAA. It is further possible to interpret significant results in terms of change scores $Δ_{subjetive Stress}^{IC} = - 10.16$ , $Δ_{sCort}^{IC} = - 2.37 \frac{nmol}{l}, Δ_{sCort}^{PwD} = - 5.54 \frac{nmol}{l}$ , which express the change in each variable relative to the score at the beginning of a home visit and with respect to the person.

All results can be found graphically in Fig. 2 and detailed in Supplementary Table 2.

To interpret the coefficient of “person”, we note the following. The model estimates on subjective stress indicate a difference of – 8.63 prior to the intervention (time = 0). Figure 2 is reflective of this difference with lower subjective stress for patients of about this magnitude. After the intervention (time = 1), we can calculate the difference between both groups using the respective interactions with time, – 8.63 + (– 3.05) – (– 10.15) = – 1.53. That is, the gap shrinks even further (though it was not significant even at time = 0). This is also reflected in Fig. 2, where the average difference between IC and PwD is reduced after each session. Note that the numerical values are about the same as in our fixed-effect model. However, there are subtle differences as the fixed-effect model includes fixed effects and other covariates.

Fig.2

Psychobiological stress measures before and after each session separately for caregivers and people living with dementia, averaged over all seven home visit sessions. sCort, salivary cortisol; sAA, salivary alpha-amylase; IC, informal caregiver; PwD, people living with dementia, error bars represent standard error of the mean, significance levels: ^***≤0.001, ^**0.01, ^*0.05.

Hypothesis 2: The dyadic psychosocial intervention leads to intermediate stress-reducing effects in psychobiological stress markers in both PwD and ICs

Concerning AUC_g (VAS), the interaction terms of person^*time were significant, meaning that ICs (b =– 5266.27, p_{fixed effects model} = 0.02; b = – 5260.68, p_{random effects model}≤0.01) and PwD (b = – 3076.31, p_{fixed effects model} = 0.14; b = – 3517.53, p_random _{effects model} = 0.04) reported lower subjective stress in their daily life at the end of the intervention compared to at the beginning. Overall, there was a significant main effect of person as PwD reported lower subjective stress in daily life than did ICs (b = – 12789.32, p_{fixed effects model} = 0.003; b =– 16185,55, p_{random effects model}≤0.001). Concerning control variables, NPI score was positively associated with AUCg_(VAS) (b = 228.46, p = 0.04) in addition to BMI (negatively correlated with AUCg_(VAS), b = – 1086.54, p = 0.01) and age (positively correlated with AUCg_(VAS), b = 399.46, p = 0.04) in the random effects model. In both the fixed and random effects models, the AUC_g (sCort) for sCort was lower in PwD at the end of the intervention compared to at the beginning (b = – 0.30, p_{fixed effects model} = 0.03; b =– 0.29, p_{random effects model} = 0.03. The AUC_g (sAA) for sAA activity did not differ significantly from pre- to post-intervention (p > 0.05), however was positively associated with age (b = 0.04, p_{fixed effects} _model = 0.01; b = 0.04, p_{random effects model} = 0.001).

All results can be found graphically in Fig. 3 and detailed in Supplementary Table 3.

Fig. 3 Mean diurnal profiles of psychobiological stress measures (AUCg) on two consecutive days at the beginning and end of intervention separately for caregivers (A-C) and people living with dementia (D-F). AUCg, Area under the curve with respect to ground, error bars represent standard error of the mean.

With regard to HCC, there were no significant differences from the beginning to the end of the intervention (p > 0.05) and no differences between PwD and ICs (p > 0.05). However, when examining control variables, a higher BMI was associated with higher HCC (b = 0.09, p_{fixed effects model} < 0.01; b = 0.06, p_{random effects model} < 0.01) in addition to a higher MMSE sum score being associated with lower HCC (p≤0.001). Post hoc testing of the interaction term revealed that higher MMSE was associated with lower HCC in both PwD (p≤0.001) and ICs (p = 0.01).

All results can be found graphically in Fig. 4 and detailed Supplementary Table 4.

Fig.4

Hair cortisol concentrations at the beginning and end of the intervention separately for caregivers and people living with dementia. PwD, people living with dementia; IC, informal caregiver, error bars represent standard error of the mean.

DISCUSSION

Summary of findings

The dyadic intervention differentially affected psychobiological stress measures over time in PwD and their IC. There was a discrepancy between subjective and physiological markers of stress, particularly for PwD. In terms of immediate effects of the intervention, ICs reported lower subjective stress after each home visit, whereas subjective stress reports remained the same for PwD. Both PwD and ICs showed lower sCort secretions after each home visit. sAA did not vary significantly. In terms of intermediate effects of the intervention as captured by ambulatory assessment, ICs reported lower subjective stress throughout the day at the end of the intervention whereas PwD showed overall lower cortisol secretion in addition to lower subjective stress reports at the end of the intervention. HCC did not differ over time and also did not differ between PwD and ICs, but were positively affected by the degree of the PwD’s cognitive impairment (MMSE sum score) in both PwD and caregivers. The inclusion of physiological stress measures was particularly relevant in PwD, as it allows their perspective to be objectively complemented.

Benefits of the dyadic intervention for stress reduction purposes

Caregivers benefitted in terms of stress reduction, as reflected both in their subjective stress levels and in sCort secretion from beginning to end of each home visit. However, the reduction in physiological stress markers was not significant in caregivers from beginning to end of intervention, although descriptively, the findings point to lower HCC. In ICs, intermediate effects (as measured by changes in psychobiological stress measures from beginning to end of each home visit) were stronger than intermediate effects (as measured by changes in AUCg for psychobiological stress measures, assessed during a two-day ambulatory assessment, from beginning to end of intervention and by HCC). Given the progressive nature of dementia, particularly spousal caregivers are at greater risk of increasing burden over time [52]. In this regard, stabilizing effects of interventions are worthwhile. For example, in a home-based intervention for ICs, Berwig et al. [53] found that caregiver burden only decreased very slightly in the intervention group, while it increased very strongly in the control group. Our findings add to this pattern, insofar as the results varied depending on the temporal scale considered: On the one hand, psychobiological stress measures decreased immediately after each home visit, and the extent of the stress reduction increased with the number of home visits. On the other hand, with respect to intermediate and long-term effects, rather stabilizing effects were found.

Regarding the control variables, at first glance, it appears contradictory that higher subjective reports of BPSD as measured by NPI total sum score were associated with lower sCort secretion surrounding the home visits. Nevertheless, these findings may be reconciled insofar as it is possible that the dyads being affected by BPSD experienced a greater reduction in stress than those dyads who are not affected by BPSD. Especially, as BPSD account for a major part of stress in both PwD and ICs, the home visits might have helped to alleviate the burden associated with these symptoms. Concerning ambulatory assessment data, higher extents of BPSD were associated with higher subjective stress mirroring findings on BPSD and subjective stress. However, physiological stress markers were not affected by BPSD in daily life. While studies consistently show that BPSD are associated with increased caregiver burden [54], the effects of BPSD on cortisol secretion in daily life remain unclear, and the number of studies in this regard is limited [55]. In line with this, it might not be BPSD per se that cause higher caregiver burden and higher physiological stress levels. Rather, particular symptoms might be more closely associated with caregiver health than others [56]. Furthermore, BPSD are not stable, but vary and interact with state characteristics of the dyad [57]. Thus, the inconsistent findings regarding the effects of BPSD on psychobiological stress in daily life might result from high within-variance over days in caregiver burden and the presentation of BPSD.

Disease severity as measured by the MMSE sum score was associated with HCC in that higher disease severity was linked to higher sCort secretion in both PwD and ICs. However, these scores did not correlate with the daily outputs of sCort. As HCC reflects long-term cortisol secretion, however, it might be plausible to assume that disease severity increases cortisol on a long-term scale, while it is inconsistently associated with short-term reactivity in cortisol secretion. In this regard, future studies need to investigate both short- and long-term associations among characteristics of the person with dementia, the caregiver, and the environment using a comprehensive assessment of psychobiological stress measures.

PwD benefitted from the intervention particularly regarding decreased sCort secretion after each home visit, as well as overall at the end of the intervention. Given that cortisol is associated with faster disease progression [58], it is an important finding that reductions in sCort secretion were found both immediately and intermediately. Furthermore, the discrepancy between subjective and physiological markers of stress is a relevant finding that deserves discussion. PwD subjectively reported lower stress particularly in daily life even though their sCort secretion was significantly higher than that of their respective caregiver. There are various possible explanations for this observation. One potential explanation is that due to the cognitive decline in dementia, PwD might have problems stating their subjective stress level, as has been reported in literature on the (questionable) validity of self-reports in PwD [59]. At the same time, the PwD investigated here showed only mild to moderate symptoms of dementia. It is also possible that they learned to hide their deficits and problems from others in daily life in order to avoid drawing attention from their caregiver. In particular, as relationship aspects quickly change, with a previously equal relationship transforming into an unequal one, the overall low subjective stress reports might be attributable to social desirability. Another underlying reason for this discrepancy between subjective and objective measures might be anosognosia which has been stated to be present in a majority of PwD [60]. It is even already common in patients with mild Alzheimer’s disease and increases with disease progression [61]. In fact, anosognosia concerning dementia symptoms was associated with anosognosia for perceived stress but not with sCort secretion in patients with Alzheimer’s disease [62]. This pattern of results fits our results well as it might indicate that subjectively PwD were not able to rate their stress level adequately, however the underlying physiology was unaffected by that. Thus, we discuss both motivational reasons or deficits in awareness of perceived stress underlying the discrepancy between subjective and objective stress markers. Irrespective of which is true, this finding emphasizes the need to assess the perspective of the person with dementia by means of objective markers. If the effectiveness of this intervention had only been measured by subjective stress reports, one conclusion would be that the intervention is beneficial for caregivers but not for PwD— a conclusion that would clearly neglect the beneficial effects of the intervention on both immediate and intermediate effects on cortisol secretion in PwD.

Overall, no effects of the intervention on sAA activity were found, either in PwD or in ICs. Concerning PwD, we found only one study investigating effects of an intervention on sAA activity. Rubia Ortí et al. [63] delivered a music therapy intervention for PwD and reported no changes in sAA from beginning to end of each session. The authors discussed that the notion of alpha-amylase as a potential biomarker in PwD is still preliminary. Concerning caregivers of PwD, Liu et al. [19] found that characteristics of daily sAA output varied depending on the use of adult day services and care-related stressors, thus emphasizing that care-related stressors in daily life can have a profound effect on stress physiology. Savla et al. [64] were able to show dissociations between daily outputs of sCort and sAA depending on stressors and negative marital interactions in spouses of persons with mild cognitive impairment. In general, the dynamics of sAA activity in PwD and their caregivers need more investigation in future studies.

Added value of psychobiological stress measures in the evaluation of dyadic psychosocial interventions

Relating the current findings to the existing literature, while no previous dyadic intervention studies have investigated psychobiological stress measures in both the PwD and IC [24], there are some studies reporting on psychobiological evaluations concerning the caregiver. Prick et al. [65] found no differences in cortisol awakening response from beginning to end of their home-based intervention, whereas Hirano et al. [66] found increased mean levels of cortisol from beginning to end of intervention. These two dyadic interventions had rather an activity-based approach, thus limiting the comparability to the study presented here. Concerning sAA, there is no available literature on effects of a dyadic intervention on the enzymatic activity of alpha-amylase.

Our findings suggest differential effects of this home-based dyadic intervention on psychobiological stress measures concerning person (PwD versus IC), concerning time scales (immediate versus intermediate and long-term effects), and concerning setting (home visit versus daily life).

In PwD, sCort secretion decreased both immediately after each home visit and intermediately in daily life (ambulatory assessment), as measured by AUCg. However, this effect decreased over time, which might be interpreted such that in daily life, regular repetitions are necessary to foster intermediate and long-term effects. In particular, as HCC reflects changes in HPA axis activity over months, it might be necessary to deliver more intense interventions in order to achieve a more profound longitudinal effect.

In ICs, the subjective stress levels corresponded to sCort secretion with regard to the immediate effects of each home visit. Concerning intermediate effects, only subjective stress levels were lower at the end of the intervention. No differences in HCC were found in ICs from beginning to end of intervention, which emphasizes the need for future research to examine how beneficial effects of each home visit can be transferred to the participants’ daily life in order to foster beneficial effects. At the same time, stabilizing effects in psychobiological stress measures are also worthwhile given the progressive nature of dementia.

Limitations

Although the comprehensive assessment of psychobiological stress measures is an important strength of the study, certain limitations warrant critical attention. Whereas the saliva samples before and after each session were collected in the presence of study staff, the saliva samples drawn in daily life were not as closely monitored. Although we included compliance checks (e.g., asking participants to write down the exact time of collection; providing them with a timer), the compliance with the protocol for saliva collection might be lower than the data surrounding the home visits. Furthermore, the assessment of saliva-related control variables covering eating, drinking, and physical activity by self-report was not feasible for PwD resulting in a high amount of missing values. On the one hand, it seems impossible to ask PwD to provide these memory-related facts and from a dyadic point of view it introduces an imbalance in relationship on the dyad as the ICs often take over for the patient. Therefore, we decided to refrain from assessing these control variables and relied on the scheduling of the home visits relative to the last meals and drinks instead. In this regard, we did not control for time since awakening or time since the last day-time nap. Future studies need to assess these variables as well to distinguish effects of an intervention from changes due to diurnal rhythm. Concerning the study sample, it has to be critically discussed that our sample was a convenience sample of predominantly spousal caregivers with wives caring for their husbands. This brings about an unequal gender distribution, as ICs were mostly female and PwD male, a common gender distribution of domestic care arrangements. As such, the ability to generalize the findings to male caregivers caring for female PwD is limited, although analyses controlled for gender and relationship within each dyad. Studies with a particular focus on male caregivers in domestic care situations are necessary. Finally, this study was exploratory in nature with a small sample size only and no control group, thus preventing conclusions on causality. Future randomized controlled trials are necessary to corroborate these findings in larger samples and to compare this intervention to a support group and no intervention to ultimately test which components of the intervention are beneficial.

Conclusion

The home-based psychosocial dyadic intervention differentially affected psychobiological stress markers in both PwD and their ICs. Both benefitted immediately from each home visit, with stress-reducing effects, as well as intermediately, as measured by daily outputs of psychobiological stress measures. In particular, PwD showed a discrepancy between subjective and physiological markers of stress. Furthermore, the stress-reducing effects decreased with the amount of time elapsing after each home visit. Two important conclusions can be drawn. First, objective physiological stress markers allow the perspective of PwD to be complemented and should be considered in the evaluation of interventions. Second, future studies are necessary to investigate how the beneficial effects of the intervention can be more sustainably transferred to daily life in the long term, although the stabilizing effects over time that we found in this study are also worthwhile given the progressive nature of dementia.

Footnotes

ACKNOWLEDGMENTS

We gratefully acknowledge the skilled work of Corinna Salbach and Hans-Joachim Bretscher in delivering the dyadic intervention. We also thank Benedikt Judmann, Sina Teifel, and Therese Steudter for their tremendous effort in analyzing saliva samples, as well as Ricarda Baake for her assistance in study administration. We further thank Ina Dhrami, who was involved in the analysis of HCC.

AF gratefully acknowledges funding by the Ministry of Social Affairs, Labor, Health, and Demography of Rhineland Palatinate, Germany. The Alfons Geib– Stiftung kindly supported the biochemical analyses.

Authors’ disclosures available online ().

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

References

McEwen

(1998) Stress, adaptation, and disease. Allostasis and allostatic load. Ann N Y Acad Sci 840, 33–44.

Csernansky

, Dong

, Fagan

, Wang

, Xiong

, Holtzman

, Morris

(2006) Plasma cortisol and progression of dementia in subjects with Alzheimer-type dementia. Am J Psychiatry 163, 2164–2169.

Allen

, Curran

, Duggan

, Cryan

, Chorcoráin

, Dinan

, Molloy

, Kearney

, Clarke

(2017) A systematic review of the psychobiological burden of informal caregiving for patients with dementia: Focus on cognitive and biological markers of chronic stress. Neurosci Biobehav Rev 73, 123–164.

da Silva

, Ramalho Oliveira

, Tavares Mello

, Moraes

, Deslandes

, Laks

(2018) Heart rate variability indexes in dementia: A systematic review with a quantitative analysis. Curr Alzheimer Res 15, 80–88.

Sakurai

, Onishi

, Hirai

(2015) Impaired autonomic nervous system activity during sleep in family caregivers of ambulatory dementia patients in Japan. Biol Res Nurs 17, 21–28.

Gaspar

, Ewing

(2020) Feasibility and usability of salivary alpha-amylase to measure stress among individuals with dementia. Res Gerontol Nurs 13, 120–124.

Pavlickova

, Russell

, Lightman

, McCabe

(2021) Feasibility of salivary cortisol collection in patients and companions attending dementia diagnostic meetings in memory clinics. BMC Res Notes 14, 30.

Ouanes

, Julius

(2017) Interrelationships between cortisol, cognition and dementia: A review of the literature and new own findings. Eur Psychiatry 41, S635–S636.

Hatfield

, Herbert

, van Someren

, Hodges

, Hastings

(2004) Disrupted daily activity/rest cycles in relation to daily cortisol rhythms of home-dwelling patients with early Alzheimer’s dementia. Brain 127, 1061–1074.

10.

Kovach

, Woods

, Logan

, Raff

(2011) Diurnal variation of cortisol in people with dementia: Relationship to cognition and illness burden. Am J Alzheimers Dis Other Demen 26, 145–150.

11.

Kim

, Bolkan

, Crespi

, Madigan

(2019) The relationship between hair cortisol, chronic stress, and well-being among older adults with dementia. Innov Aging 3, S468–S468.

12.

de la Rubia Ortí

, Prado-Gascó

, Sancho Castillo

, Julián-Rochina

, Romero Gómez

, García-Pardo

(2019) Cortisol and IgA are involved in the progression of Alzheimer’s disease. A pilot study. Cell Mol Neurobiol 39, 1061–1065.

13.

Byman

, Schultz

, Bank

, Fex

, Wennström

(2018) Brain alpha-amylase: A novel energy regulator important in Alzheimer disease? Brain Pathol 28, 920–932.

14.

Schulz

, Sherwood

(2008) Physical and mental health effects of family caregiving. Am J Nurs Sci 108, 23–27.

15.

Losada-Baltar

Stress and coping in caregivers, theories of. In Encyclopedia of Geropsychology, Pachana NA, ed. Springer Singapore, Singapore, pp. 2268–2276.

16.

de Vugt

, Nicolson

, Aalten

, Lousberg

, Jolle

, Verhey

(2005) Behavioral problems in dementia patients and salivary cortisol patterns in caregivers. J Neuropsychiatry Clin Neurosci 17, 201–207.

17.

Stalder

, Tietze

, Steudte

, Alexander

, Dettenborn

, Kirschbaum

(2014) Elevated hair cortisol levels in chronically stressed dementia caregivers. Psychoneuroendocrinology 47, 26–30.

18.

Cuijpers

(2005) Depressive disorders in caregivers of dementia patients: A systematic review. Aging Ment Health 9, 325–330.

19.

Liu

, Granger

, Kim

, Klein

, Almeida

, Zarit

(2017) Diurnal salivary alpha-amylase dynamics among dementia family caregivers. Health Psychol 36, 160–168.

20.

Van’t

Leven N

, Prick

A-E

, Groenewoud

, Roelofs

, de Lange

, Pot

(2013) Dyadic interventions for community-dwelling people with dementia and their family caregivers: A systematic review. Int Psychogeriatr 25, 1581–1603.

21.

Moon

, Adams

(2013) The effectiveness of dyadic interventions for people with dementia and their caregivers. Dementia 12, 821–839.

22.

Brodaty

, Green

, Koschera

(2003) Meta-analysis of psychosocial interventions for caregivers of people with dementia. J Am Geriatr Soc 51, 657–664.

23.

McDermott

, Charlesworth

, Hogervorst

, Stoner

, Moniz-Cook

, Spector

, Csipke

, Orrell

(2019) Psychosocial interventions for people with dementia: A synthesis of systematic reviews. Aging Ment Health 23, 393–403.

24.

Wuttke-Linnemann

, Baake

, Fellgiebel

(2019) Dyadic wind of change: New approaches to improve biopsychological stress regulation in patients with dementia and their spousal caregivers.1325-. J Alzheimers Dis 68, 1337.

25.

Rausch

, Caljouw

, van der Ploeg

(2017) Keeping the person with dementia and the informal caregiver together: A systematic review of psychosocial interventions. Int Psychogeriatr 29, 583–593.

26.

Cohen

, Wills

(1985) Stress, social support, and the buffering hypothesis. Psychol Bull 98(2), 310–357.

27.

Saxbe

, Repetti

(2010) For better or worse? Coregulation of couples’ cortisol levels and mood states. J Pers Soc Psychol 98, 92–103.

28.

Pauly

, Michalowski

, Drewelies

, Gerstorf

, Ashe

, Madden

, Hoppmann

(2020) Cortisol synchrony in older couples: Daily socioemotional correlates and interpersonal differences. Psychosom Med 82, 669–677.

29.

Holt-Lunstad

, Birmingham

, Light

(2008) Influence of a “warm touch” support enhancement intervention among married couples on ambulatory blood pressure, oxytocin, alpha amylase, and cortisol. Psychosom Med 70, 976–985.

30.

Braun

, Scholz

, Bailey

, Perren

, Hornung

, Martin

(2009) Dementia caregiving in spousal relationships: A dyadic perspective. Aging Ment Health 13, 426–436.

31.

Gellert

, Häusler

, Suhr

, Gholami

, Rapp

, Kuhlmey

, Nordheim

(2018) Testing the stress-buffering hypothesis of social support in couples coping with early-stage dementia. PLoS One 13, e0189849.

32.

Wuttke-Linnemann

, Henrici

, Müller

, Lieb

, Fellgiebel

(2020) Bouncing back from the burden of dementia: Predictors of resilience from the perspective of the patient, the spousal caregiver, and the dyad— An exploratory study. GeroPsych 33, 170.

33.

Häusler

, Sánchez

, Gellert

, Deeken

, Rapp

, Nordheim

(2016) Perceived stress and quality of life in dementia patients and their caregiving spouses: Does dyadic coping matter? Int Psychogeriatr 28, 1857–1866.

34.

Fonareva

, Amen

, Ellingson

, Oken

(2012) Differences in stress-related ratings between research center and home environments in dementia caregivers using ecological momentary assessment. Int Psychogeriatr 24, 90–98.

35.

DGN, DGPPN (2017) S3–Leitlinie Demenzen, DOI: 10.1007/978-3-662-53875-3.

36.

Nordheim

, Häusler

, Yasar

, Suhr

, Kuhlmey

, Rapp

, Gellert

(2019) Psychosocial intervention in couples coping with dementia led by a psychotherapist and a social worker: The DYADEM trial. J Alzheimers Dis 68, 745–755.

37.

Linnemann

, Fellgiebel

(2017) Psychotherapie bei leichter kognitiver Beeinträchtigung und Demenz. Nervenarzt 88, 1240–1245.

38.

Hepburn

, Lewis

, Kirk

, Narayan

, Center

, Bremer

, Tornatore

(2006) Partners in caregiving. Clin Gerontol 29, 53–69.

39.

Linnemann

, Ditzen

, Strahler

, Doerr

, Nater

(2015) Music listening as a means of stress reduction in daily life. Psychoneuroendocrinology 60, 82–90.

40.

Linnemann

, Kappert

, Fischer

, Doerr

, Strahler

, Nater

(2015) The effects of music listening on pain and stress in the daily life of patients with fibromyalgia syndrome. Front Hum Neurosci 9, 434.

41.

Linnemann

, Strahler

, Nater

(2016) The stress-reducing effect of music listening varies depending on the social context. Psychoneuroendocrinology 72, 97–105.

42.

Wennig

(2000) Potential problems with the interpretation of hair analysis results. Forensic Sci Int 107, 5–12.

43.

Stalder

, Steudte

, Alexander

, Miller

, Gao

, Dettenborn

, Kirschbaum

(2012) Cortisol in hair, body mass index and stress-related measures. Biol Psychol 90, 218–223.

44.

Stalder

, Steudte-Schmiedgen

, Alexander

, Klucken

, Vater

, Wichmann

, Kirschbaum

, Miller

(2017) Stress-related and determinants of hair cortisol in humans: A meta-analysis basic. Psychoneuroendocrinology 77, 261–274.

45.

Champaneri

, Xu

, Carnethon

, Bertoni

, Seeman

, DeSantis

, Diez Roux

, Shrager

, Golden

(2013) Diurnal salivary cortisol is associated with body mass index and waist circumference: The Multiethnic Study of Atherosclerosis.. Obesity 21, E56–E63.

46.

Uhart

, Chong

, Oswald

, Lin

, Wand

(2006) Gender differences in hypothalamic-pituitary-adrenal (HPA) axis reactivity. Psychoneuroendocrinology 31, 642–652.

47.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

48.

Cummings

, Mega

, Gray

, Rosenberg-Thompson

, Carusi

, Gornbein

(1994) The Neuropsychiatric Inventory: Comprehensive assessment of psychopathology in dementia. Neurology 44, 2308–2314.

49.

Dauphinot

, Delphin-Combe

, Mouchoux

, Dorey

, Bathsavanis

, Makaroff

, Rouch

, Krolak-Salmon

(2015) Risk factors of caregiver burden among patients with Alzheimer’s disease or related disorders: A cross-sectional study. J Alzheimers Dis 44, 907–916.

50.

Croissant

, Millo

, Tappe

, Toomet

, Kleiber

, Zeileis

, Henningsen

, Andronic

, Schoenfelder

, Croissant

(2021) Package ‘plm’. Choice 139, 227–240.

51.

Pruessner

, Kirschbaum

, Meinlschmid

, Hellhammer

(2003) Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology 28, 916–931.

52.

van

den Kieboom R

, Snaphaan

, Mark

, Bongers

(2020) The trajectory of caregiver burden and risk factors in dementia progression: A systematic review. J Alzheimers Dis 77, 1107–1115.

53.

Berwig

, Heinrich

, Spahlholz

, Hallensleben

, Brähler

, Gertz

H-J

(2017) Individualized support for informal caregivers of people with dementia - effectiveness of the German adaptation of REACH II. BMC Geriatr 17, 286.

54.

Allegri

, Sarasola

, Serrano

, Taragano

, Arizaga

, Butman

, Loñ

(2006) Neuropsychiatric symptoms as a predictor of caregiver burden in Alzheimer’s disease. Neuropsychiatr Dis Treat 2, 105–110.

55.

Cheng

S-T

(2017) Dementia caregiver burden: A research update and critical analysis. Curr Psychiatry Rep 19, 64.

56.

Feast

, Moniz-Cook

, Stoner

, Charlesworth

, Orrell

(2016) A systematic review of the relationship between behavioral and psychological symptoms (BPSD) and caregiver well-being. Int Psychogeriatr 28, 1761–1774.

57.

Campbell

, Rowe

, Marsiske

(2011) Behavioral symptoms of dementia: A dyadic effect of caregivers’ stress process? Res Gerontol Nurs 4, 168–184.

58.

Huang

C-W

, Lui

C-C

, Chang

W-N

, Lu

C-H

, Wang

Y-L

, Chang

C-C

(2009) Elevated basal cortisol level predicts lower hippocampal volume and cognitive decline in Alzheimer’s disease. J Clin Neurosci 16, 1283–1286.

59.

Bedard

, Squire

, Minthorn-Biggs

M-B

, Molloy

, Dubois

, O’Donnell

, Lever

(2003) Validity of self-reports in dementia research. Clin Gerontol 26, 155–163.

60.

Wilson

, Sytsma

, Barnes

, Boyle

(2016) Anosognosia in dementia. Curr Neurol Neurosci Rep 16, 77.

61.

Starkstein

(2014) Anosognosia in Alzheimer’s disease: Diagnosis, frequency, mechanism and clinical correlates. Cortex 61, 64–73.

62.

Arsenault-Lapierre

, Whitehead

, Lupien

, Chertkow

(2012) Effects of anosognosia on perceived stress and cortisol levels in Alzheimer’s disease. Int J Alzheimers Dis 2012, 209570.

63.

de la Rubia Ortí

, Pardo

MPG

, Pérez-Ros

, Rochina

(2018) Iga and alpha-amylase: New targets for well-being in Alzheimers disease patients, a pilot study with music therapy. Neuropsychiatry 8, 378–383.

64.

Savla

, Granger

, Roberto

, Davey

, Blieszner

, Gwazdauskas

(2013) Cortisol, alpha amylase, and daily stressors in spouses of persons with mild cognitive impairment. Psychol Aging 28, 666–679.

65.

Prick

, de Lange

, Twisk

, Pot

(2015) The effects of a multi-component dyadic intervention on the psychological distress of family caregivers providing care to people with dementia: A randomized controlled trial. Int Psychogeriatr 27, 2031–2044.

66.

Hirano

, Umegaki

, Suzuki

, Hayashi

, Kuzuya

(2016) Effects of leisure activities at home on perceived care burden and the endocrine system of caregivers of dementia patients: A randomized controlled study. Int Psychogeriatr 28, 261–268.