Table 2 Characteristics of studies included in the review

Resilience Scale (25 items)

  Bull [26]

(To explore the approaches family caregivers use to help them provide care for a family member with dementia; to describe the psychological distress and resilience of family caregivers)

N = 18 family caregivers. 39% were caring for spouses; the others were children of the person with dementia. Age range 37 to 86 years, m = 64 SD = 14.1. 67% were female (67%); 60% were white and 40% were African American. 89% identified themselves as Christian

Midwestern USA (English)

Cross sectional mixed-methods design

(Telephone interviews)

None specified

None reported

  Dias et al. [27]

(To investigate the relationship between caregivers’ resilience and the sociodemographic and clinical factors of people with dementia)

N = 58 carers (dyads). Most caregivers were female (79.3%), married (77.6%), with a mean age of 62.5 ± 13.4 years. 51.7% were the adult children of the person with dementia

Type of dementia: Mild to moderate Alzheimer’s disease, vascular dementia and mixed dementia

Rio de Janeiro, Brazil (Portuguese)

Cross-sectional design

(Mode of data collection unclear)

Caregivers’ resilience is a personality trait, independent from the clinical symptoms of the person with dementia

Reliability not reported in this study sample

No relationship was found between resilience and caregivers’ gender (p = 0.883), nor clinical (p = 0.807) and emotional problems (p = 0.420). There was no significant relationship between caregivers’ resilience and the sociodemographic and clinical characteristics of the person with dementia

Large correlations between resilience and caregivers’ depressive symptoms (r = -0.539; p < 0.01) and carer quality of life (r = 0.514; p < 0.01)

Medium correlation between resilience and anxiety (r = -0.334 p < 0.01)

Small correlation between resilience and burden (r = -0.27, p < 0.05). Of these significant univariate associations, depression, and carer QoL significantly predicted resilience. (Effect size for regression not available as authors used SPSS 22)

  Fitzpatrick & Vacha Haase [28]

(To explore the relationship between resilience and marital satisfaction in caregivers of spouses with dementia)

N = 30 caregivers (9 males and 21 females) of spouses with dementia Age range 64 to 90 years, m = 76.4 years, SD = 6.0. One African-American caregiver, the rest were Euro-American

Type of dementia: probable Alzeimer’s disease; with 10% of carers reporting MCI, vascular dementia or dementia unspecified

Colorado USA (English)

Cross-sectional design

(Telephone and face-to-face interviews)

Resilience would be most related to marital satisfaction when adjusting for caregiver burden,

age of caregiver, length of marriage in years, gender, and years since

dementia diagnosis

No hypotheses regarding expected differences in resilience between groups

Ideal internal consistency (α = 0.93)

The hypothesis that resilience would be related to marital satisfaction was not supported (r = -0.10; probability level not reported)

Not hypothesis driven – high correlation between resilience and self-efficacy (r = 0.52, p < 0.01)

  Kimura et al. [29]

(To assess whether the clinical symptoms of the person with young onset Alzheimer’s Disease might be associated with resilience in their carers)

N = 43 family caregivers; n = 34 were female; (n = 21 spouses, n = 15 children, n = 4 siblings, n = 3 ‘other’). 81.4% lived with care recipient. Mean age = 51.1, SD = 15.2

Rio de Janeiro, Brazil (Portuguese)

Cross sectional design

(Face to face interviews)

Carer resilience might be related to the presence of awareness of disease, neuropsychiatric symptoms, high levels of depression, and impairment on the functionality of the person with YOAD

Ideal internal consistency (α = 0.80)

No significant differences between carers resilience and the characteristics of the person with dementia, contrary to their hypothesis

Not hypothesis driven—there are medium negative correlations between carers resilience and their depressive symptoms (r = -0.40, p < 0.01); anxiety (r = -0.36, p < 0.05), and hopelessness (r = -0.33, p < 0.05). No significant correlations between carers resilience and carer quality of life

  Kimura et al. [30]

(To compare the quality of life, burden, and depressive symptoms of caregivers of individuals with young-onset dementia (YOD) and late-onset dementia (LOD))

N = 110 dyads of individuals with mild to severe Alzheimer disease and their caregivers (55 dyads of individuals with young-onset Alzheimer disease and 55 dyads of individuals with late-onset Alzheimer disease)

N = 85 females, mean age = 54.70 (SD = 14.4); N = 44 spouses, N = 54 children

Rio de Janeiro, Brazil

(Portuguese)

Cross sectional design

(Face to face interviews)

None specified

Reliability not reported for the study sample

Not hypothesis driven:

YOD caregivers’ resilience: medium correlation with QoL(r = -0.365; p < .05) and small correlation with depression (r = -0.297; P < .05) but not burden

LOD carers resilience: small correlation with depressive symptoms

(r = -0.269; P < .05) but not burden or QoL

  MacCourt, et al. [31]

(To report on the structure and effectiveness of a grief management coaching intervention with carers of individuals with dementia)

N = 200 Dementia caregivers Spouse (61.9%), parent (23%), other 10%. Mean age = 64.4 (range and SD not reported). 79% (n = 158) were female; 82% (n = 163) were married; 62% (n = 122) caring for spouse; 23% (n = 65) caring for parent; 5.1% (n = 10) caring for ‘other’

Type of dementia:

Alzheimer’s Disease or dementia

British Columbia, Canada (English)

Controlled mixed methods intervention study

(Not reported)

None specified

Ideal internal consistency (time 1 and time 2 α = 0.91)

Not hypothesis driven: Responsiveness: There was a significant improvement in resilience from T1 to T2 for the grief coaching intervention group compared to the control group (F = 10.70, df = 185, p = .009), and there was no change in the control group. Effect sizes not reported

  Monteiro et al. [32]

(To test the construct validity of the Resilience Scale through exploratory and confirmatory procedures, and to investigate the relationship between caregiver’s resilience and clinical status of people with Alzheimer’s disease)

N = 143 carers of people with AD; mean age = 58.8 (SD = 14.3); N = 118 females. No other demographic information is presented

Rio de Jeneiro, Brazil

(Portuguese)

Cross-sectional design

(Face to face interviews)

None specified

Adequate to ideal internal consistency α = 0.77

The factor analysis found a four-factor solution—sense of life and self-sufficiency, perseverance, self-confidence and equanimity, and meaningfulness. The authors say this demonstrates the construct validity of the measure, but they do not specify any hypotheses, and note earlier the measure reflects serenity, perseverance, self-confidence, sense of life and self-sufficiency. In view of this, evidence of ‘structural validity’ is difficult to confirm

There were no correlations between resilience and PWD clinical measures (functional activities, depression, psychosocial impact, MMSE, NPI, CDR) and between resilience and burden (ZBI)

  Pessotti et al. [33]

(To ascertain the impact of family caregivers quality of life, burden and resilience and religiosity; to relate these to the clinical and cognitive characteristics of older people with dementia)

N = 50 family caregivers. 88% female; mean age = 54.7 (SD = 11.1)

32% were wives and 54% daughters of the person with dementia

Type of dementia: 4 people (68%) with Alzheimer’s disease; n = 12 with vascular dementia; n = 2 with alcohol related; n = 2; Parkinson’s Disease related

Sao Paulo, Brazil (Portuguese)

Cross-sectional design

(Mode of data collection is unclear)

Perceived QoL and burden of carers is more related to aspects of religiosity and more resilient responses, and less associated with clinical aspects of the elder with dementia

No hypotheses regarding expected differences in resilience between groups

Reliability not reported in this study sample

Results supported the authors hypothesis; high positive correlation between resilience and carers quality of life (r = 0.56, p < 0.001); medium negative correlations between resilience depression (r = -0.36. P < 0.01); burden (r = -0.36, p < 0.01), and intrinsic religiosity, where lower scores indicate higher religiosity (r = -0.37, P < 0.01);

With the exception of a positive correlation between more severe dementia and higher resilience in a regression model (β = 11.15, p < 0.01), there were no associations between caregiver resilience and the socio-demographic and cognitive and disability characteristics of the person living with dementia

  Rosa et al. [34]

(To investigate the resilience of carers of people with mild and moderate Alzheimer’s, and the related socio-demographic and clinical characteristics)

N = 106 caregivers. 72% (n = 44) of the mild AD caregivers were female and 55.7% (n = 34) daughters. 75.4%, (n = 46) were married. Mean age = 57.9 ± 13.7 years. 68.9% (n = 42) were co-residing with the care recipient. 88.9% (n = 40) of the moderate AD caregivers women and 48.9%, (n = 22) daughters. 71.1% (n = 32) were married. Mean age = 59 ± 11.83. 77.7% (n = 35) were co-residing with the care recipient

Type of dementia: 34 cases (68%)

Alzheimer’s disease; n = 12 vascular dementia; n = 2 with alcohol related; n = 2 Parkinson’s disease

Rio de Jeniero Brazil (language not reported, but likely Portuguese)

Cross-sectional design

(Face to face interviews)

Disease severity may have a direct

influence on the resilience of caregivers of PwAD;

caregivers of moderate PwAD will have higher levels of burden and lower levels

of resilience

Co-residing with the PwAD and caregivers’ physical and emotional problems (anxiety and depression) directly contribute to lower levels of resilience among the caregivers of moderate PwAD

No hypotheses regarding expected differences in resilience between groups

Reliability not reported in this study sample

No differences in the resilience of dementia carers according to whether the person cared for had mild or moderate dementia, contrary to the hypothesis

For caregivers of those with moderate dementia, there were small correlations between resilience and the PwAD depressive symptoms (r = 0.293; p < 0.05), and whether the carer lived with the person with dementia (r = 0.299; p < 0.05)

Medium correlations between carers resilience and the PwAD delusions (r = 0.417; p < 0.05) and awareness of disease (r = -0.374; p < 0.05) and lower levels of carers depressive symptoms (r = 0.36; p < 0.05)

High correlations between higher levels of resilience correlated with carer quality of life (r = 0.519; p < 0.001)

Not hypothesis driven: there were small correlations between the resilience of caregivers of mild dementia and the person’s neuropsychiatric symptoms (r = 0.25; p < 0.05) and appetite abnormalities (r = 0.267; p < 0.05) and medium corrleations between resilience and quality of life (r = 0.34; p < 0.05)

Resilience was inversely correlated to the caregivers’ depressive symptoms (r =—0.33; p < 0.05) and anxiety symptoms (r = -0.259; p < 0.05)

  Svanberg et al. [35]

(To explore whether children of younger people with dementia can be compared to other young carers)

N = 12 caregivers (6 male/6 female) Ages ranged from 11 to 17 years (mean 14.6). Eleven were White British and one was mixed race (White-Other)

Type of dementia: young onset dementia (n = 5

Alzheimer’s disease, n = 2

Pick’s disease, n = 1 vascular dementia; n = 1 suspected Pick’s disease

UK (English)

Cross-sectional mixed methods design

( Face to face interviews)

None specified

No relevant data reported

  Scott [74]

(To explore whether resilience has a moderating effect between Alzheimer’s caregiver stressors and burden)

N = 110 caregivers (89 female, 22 male)

Ages ranged between 25 and 89 years old (M 63, SD 11)

57 (51.4%) = White, and 52 (46.8%) = Black; 2 (1.8%) = other (self-identified)

40 (36%) = spouse caregivers, 66 (59.5%) = adult–child caregivers

28.8% unemployed, 28.8% full time work, 9% part-time work, 33.3% retired

Caregivers provided care for an average of 5 years

Type of dementia: Alzheimer’s Disease

US (English)

Cross sectional design

(Surveys)

Resilience moderates the relationship between identified caregiver stressors and caregiver burden

Reliability not reported in this study sample

Contrary to the hypothesis, resilience did not have a moderating effect between caregiver stressors and caregiver burden

Not hypothesised: there were no difference in resilience in regards to ethnicity, gender or caregiver type

Main effect for resilience (p = .001, accounting for approximately 10.2% of the variance in caregiver burden scores. As resilience increased, caregiver burden decreased as demonstrated in Pearson correlation (-.320, p = 0.001) and multiple regression (b = -.299, t =—4.099, p < .001) analyses

  Garity [73]

(To explore the relationship between stress, learning style, resilience and ways of coping in Alzheimer’s caregivers)

N = 76 (29% male, 71% female)

Mean Age = 61.5 (SD = 14.1)

49% not employed, 31% employed full-time 20% employed part-time

80% married 8% single 7% widowed 5% divorced

43% spouses 33% daughters 9% sons 8% sisters 7% granddaughters

17% 0–1 year care provision, 56% 2–4 years, 19% 5–7 years, 7% more than 7 years

Dementia type: Alzheimer’s disease

US (English)

Cross sectional design?

(Surveys administered in support group)

None specified. Conceptual model of stress used as framework for study

Reliability not reported in this study sample

Small correlations were found between resilience and

coping style on the emotion-focused subscale of escape avoidance (r = -.26, p < .05) e.g. less resilience = higher score; and between resilience and the problem-focused sub-scale of planful problem-solving (r = 0.30, p < .01)

The Brief Resilience Scale

  Canevelli et al. [36]

(To compare biological age and functional status assessed through the frailty index (FI) in caregivers and matched controls, and (ii) within caregivers, to test the association of FI with measures of perceived psychological stress and resilience)

N = 64 caregivers of people with dementia (mean age = 67.62, SD = 11.59; 38 females/26 males) and N = 64 non-caregiver controls (mean age = 67.70, SD = 11.63; 38 females/25 males) matched for age and gender

Relationship of caregivers: spouses/partners (n = 42, 65.6%) or children (n = 20, 31.2%)

Type of dementia not reported

Rome, Italy (Italian)

Cross-sectional matched control design

(Mode of data collection unclear)

The authors state: “It can be hypothesized that caregiving, intended as a condition of chronic psychological stress exposure, is associated with accelerated senescence and higher accrual of health deficits, and that, among caregivers, frailty levels are directly related to the intensity of perceived psychological stress and inversely related to the individual’s capacity of psychological resilience, i.e. the capacity of maintaining positive emotional responses in the presence of psychosocial stressors.”

Ideal internal consistency (α = 0.89)

Within the caregivers, FI was negatively associated with BRS through a large correlation (r = -0.637, p < 0.001). This association remained statistically significant (p ≤ 0.001) when age, gender, education, BMI, years of caregiving, and type of relationship with the care receiver (i.e. being spouses/partners, children, siblings, or parents of care receivers) were included as covariates

Not hypothesised: Resilient caregivers (n = 17), i.e. those with high resilience according to the BRS cut-off, had mean FI similar (non-significantly lower) to controls (0.11 ± 0.06 vs 0.16 ± 0.11, F = 2.247; p = 0.138). Effect sizes not reported

  Chan et al. [38]

(1 to explore the caregiver strains and resilience level of caregivers of patients with AD in Malaysia; 2: to determine the factors associated with caregiver strains in caregivers

of AD patients, and 3: to determine the effect of resilience on the relationship between caregiver strains and caregivers or patient's

factors.)

N = 230 carers of people living with Alzheimer’s Disease. 79% (N = 165) female. Mean age = 50.4 (SD = 14.5). 59.9% married. Relationship: 16.4% were spouse, 61.4% children, and 17.4% were identified as “others.”

Chinese was the majority (57.2%), followed by Malay (13.5%), Indian (8.7%) and 10.6% were Bidayuh, Kadazan and others

Cross-sectional design

(Mode of data collection unclear, but possibly self-report/self-completion. Three languages appear to be used-Bahasa Malaysia, English or Mandarin)

No hypotheses specified. The study does not appear to test the theoretical model specified and explores a diverse range of variables. They do not appear to have used the mean scores for the BRS as described in their methods, but used the range

Reliability not reported in this study sample

Not hypothesised:

The authors applied t-tests and present a range of mean scores for resilience that are not significant according to ethnicity, marital status, education level, kinship, hours of care

They note differences in gender with males having higher resilience (M = 20.3, SD = 3.8) than females (M = 18.9, SD = 3.2) p = 0.03; and employment status with those unemployed/homemaker/retiree having higher resilience (M = 19.7, SD = 2.9) than those employed full or part time (M = 18.8, SD = 3.5) p = 0.04

There were no differences for carers resilience according to levels of functional impairment of the care recipient. There were no differences in carers resilience according to whether or not they lived in the same house, the number of years providing care, whether they received help from family members, emotional support or hired help. Age was not associated with resilience

The ‘path’ analysis shows a medium correlation between resilience and caregiver strain (r = -0.37, p < 0.001)

  Kalaitzaki et al. [39]

(To identify the perceived symptoms of PwD (i.e. functional impairment, cognitive deterioration and behavioural-psychological symptoms) associated with poorer CGs’ QoL and examine whether CGs’ resilience reduces the effect of dementia symptoms on their QoL)

N = 118 caregivers. 78.8% female; mean age = 58.9 (SD = 11.5), 85.6% were married. 90% were the children of the person with dementia

Crete, Greece (Greek)

Cross-sectional design

(Face to face interview)

None stated

Low internal consistency α = 0.56. The authors suggest this may be due to the brevity of the scale and cite another study suggesting a reliability coefficient

of .50–.70 is considered moderately reliable

Not hypothesis driven: There was no difference in resilience between those who care few days per week (≤ 3) and those who care many days per week (≥ 4) (18.6 vs. 19.2, respectively; t = .391, p = .697)

A mediation model found the person with dementia’s behavioural symptoms significantly predicted CGs’ resilience (B = .04, SE = .02, p < .05) and CGs’ resilience significantly predicted cares QoL (B = .25, SE = .06, p < .001). The direct effect of PwD’s BP on CGs’ QoL (path c) was not statistically significant (B = .02, SE = .01, n.s.), but the indirect effect of PwD’s BP on CGs’ QoL through the mediating role of CGs’ resilience (path c΄) was statistically significant (B = .01, SE = .01, p < .05)

  McManus et al. [40]

(1. To determine the feasibility and acceptability of a performing arts intervention (MPAI) for caregivers of people with mild to moderately severe dementia; 2. To examine how MPAI might change caregiver burden, caregiver resiliency, and perceived quality of life (QoL) for care recipients

N = 32 carers. Age ranged from 41- > 80 mean age not reported. The majority (84%) were female, white (84%) and the spouses of the person with dementia (78%),

Mixed-methods feasibility intervention study with data collected at 5 time-points

(Online survey, self completion)

None specified

Internal consistency ranged across the five timepoints from α = 0.79 to α = 0.86

There were no significant differences in resilience scores across the five time-points

  Prins et al. [41]

(To examine the relationship between involvement

of family caregivers (FCs) of people with dementia (PwD) living in LTCFs and FCs mental health during the

visitor-ban, and whether this relationship was moderated by the frequency of alternative contact with PwD

during the visitor-ban and FC resilience)

N = 958 family carers, mean age = 60.30 (SD = 8.95, range

between 16 and 89). 71.7% were female. Three quarters (75.7%) indicated that the person with dementia was their father or mother. Spouses or partners represented 10.3% of the sample, while 14.0% had another type of relationship with the person with dementia (for example, other family members, friends, neighbours

or legal guardians)

Cross-sectional design

(Online survey, self-completion, part of a larger study looking at social isolation during the pandemic)

Hypothesis 1: More family involvement before the visitor banleads to more worries of family caregivers during the visitor ban. This

relationship is moderated by the frequency of contact during the visitor

ban and the resilience of the family caregiver

Hypothesis 2: More family involvement before the visitor ban

leads to more experienced loneliness in family caregivers during the

visitor ban. This relationship is moderated by the frequency of contact during the visitor ban and the resilience of the family caregiver

Reliability not reported in this study sample. The authors do not use the full scale and reduce it to two items (“I have a hard time making it through stressful events” and “It does not take me long to recover from a stressful event”)

No univariate correlations are presented. The role of resilience in hypothesis 1 was not supported

The only interaction of resilience was found for carers who undertook task related and social activities before the visitor ban, and also had higher resilience, which predicted lower loneliness (β = -0.32, p < 0.05)

  Sutter et al. [42]

(To examine the relationships between personal strengths (optimism, sense of coherence, resilience) and the mental health of dementia caregivers from Latin America)

N = 127 family caregivers ( n = 107 from Argentina and n = 20 from Mexico).72% (n = 98) female; 82% (n = 100) married; mean age = 57.14 (SD = 13.01). Relationship: 60% siblings; 22% child; 15.6% spouse; 2.1% partner

Type of dementia: Not reported

Instituto de Neurociencias de San Lucas,Argentina

Baja California, Mexico

(Not reported)

Cross-sectional design

(Self-report and face to face interview)

None specified

Reliability not reported in this study sample

The study reports medium correlations between resilience and satisfaction with life (r = 0.32, p < 0.01; depression (r = -0.35, p < 0.01) and large correlations between resilience and optimism (r = 0.50, p < 0.01). Most of the correlations between resilience and the five RSA subscales were small: personal competence (r = 0.48, p < 0.01); social competence (r = 0.23, p < 0.01); family coherence (r = 0.27, p < 0.01); social support (r = 0.26, p < 0.01); personal structure (r = 0.25, p < 0.01); and the three Sense of Coherence subscales meaningfulness (r = 0.25, p < 0.01); manageability (r = 0.22, p < 0.05) and comprehensibility (r = 0.26, p < 0.01)

In a linear regression (including the RSA subscales and the SoC subscales) resilience (BRS) predicted depression (β = -0.24, p < 0.05) but did not predict burden or satisfaction with life. Effect sizes not reported

  Vatter et al. [43]

(To explore the factor structure of the Zarit Burden Scales (ZBI) in life partners of people with Parkinson related dementia; to examine the relationships among the emerging ZBI factors and demographic and clinical features)

N = 136 spouse/life partner caregivers. 85% 9n = 116) female; 95% (n = 129) married; 89% (n = 122) white British; mean age = 69.4 (SD = 7.62)

England (English)

Cross sectional design derived from a pilot feasibility study

(Face to face interviews and self-completion)

None specified

Reliability not reported in this study sample

Not hypothesis driven: The study reports a large negative correlation between carer resilience and the Zarit Burden scale (r = -0.53, p < 0.001), and medium and large correlations between carer resilience and five dimensions derived from a factor analysis; social and psychological constraints (r = -0.40, p < 0.001); personal strain (r = -0.50, p < 0.001); interference with personal life (r = -0.38, p < 0.001); concerns about the future (r = -0.34, p < 0.001) and guilt (r = -0.31, p < 0.001)

  Vatter et al. [44]

(To explore and compare levels of mental health, care burden, and relationship satisfaction among caregiving spouses

of people with mild cognitive impairment or dementia in Parkinson disease (PD-MCI or PDD) or dementia with Lewy bodies

(DLB)

N = 136 spouse/life partner caregivers (same participants as reported in Vatter et al., 2018)

England (English)

Cross sectional design derived from a pilot feasibility study

(Face to face interviews and self-completion)

None specified

Reliability not reported in this study sample

Not hypothesis driven: No differences were found for carer resilience across the three types of dementia

Medium negative and large correlations between carer resilience and ZBI (r = -0.47, p < 0.01); anxiety (r = -0.59, p < 0.01); depression (r = -0.54, p < 0.01); relatives stress (r = -0.50, p < 0.01); relationship strain (r = -0.33, p < 0.01); role resentment (r = -0.42, p < 0.01); role anger (r = -0.32, p < 0.01). Positive correlations between carer resilience and mental health (r = 0.59, p < 0.01); health related quality of life (r = 0.35, p < 0.01); self-rated health (r = 0.34, p < .01)

  Vatter and Leroi [45]

(To explore resilience in people with Parkinson’s disease mild cognitive disorder or dementia, or dementia with Lewy bodies, and their care partners, and its relation to outcomes related to their mental well-being and quality of life)

N = 76 dyads (n = carers and n = 76 people with dementia)

Of the participants with Lewy body cognitive disorders, 19.8% (n = 15) had a diagnosis of PD-MCI, 52.6% (n = 40) had PDD, and 27.6% (n = 21) had DLB., 78.9% (n = 60) were male, and 93.4% (n = 71) were white with a mean age of 74.5 years (SD = 6.74)

Of the care partners, 85.6% (n = 65) lived with their study partner, 77.6% (n = 59) were spouses or partners, 17.1% (n = 13) were relatives, and the remainder 5.3% (n = 4) included a live-in care partner, a live-in divorcee, a friend, and a grandchild. Of the care partners, 89.5% (n = 68) were female, and 92.1% (n = 70) were white with a mean age of 65.0 years (SD = 11.81)

Cross-sectional study as part of a pilot feasibility study of an adapted CST for people with Lewy-body related cognitive disorders and their study partners/carers

(Face to face interview)

Lower resilience predicts lower mental well-being, quality of life, and relationship satisfaction in both members of the care dyad

In care partners, lower resilience predicts higher stress and burden

Reliability not reported in this study sample

Carer hypotheses were all met but not all hypotheses for people with dementia were met

Care partners self-reported higher resilience scores (m = 3.79, SD = 0.82) than people with Lewy body cognitive disorders (m = 3.23, SD = 0.71, p < 0.001)

Most participants with Lewy body-related cognitive disorders (74%; n = 56) and care partners (83%; n = 63) reported high resilience (i.e., above a mean score of 3.00). People with Lewy body-related cognitive disorders with lower levels of resilience had higher levels of anxiety (HADS, p < 0.001), higher frequency and severity of neuropsychiatric symptoms (NPI, p = 0.047), lower levels of quality of life related to Parkinson’s (PDQ-39, p = 0.006), and overall quality of life (EQ-5D, p = 0.004) compared to those with higher resilience scores

Care partners with lower levels of resilience reported lower relationship satisfaction (RSS, p = 0.002), lower quality of life (EQ-5D, p = 0.001), lower scores on mental health (SF-12-MCS, p < 0.001) and physical health (SF-12-PCS, p = 0.037), and higher levels of anxiety (HADS, p < 0.001), depression (HADS, p < 0.001), burden (ZBI, p < 0.001), and stress (Rel.SS, p < 0.001)

Higher resilience in people with Lewy body-related cognitive disorders was associated with lower anxiety (HADS-A, r = -0.52 p < 0.001) and higher overall quality of life (EQ5D-VAS, r = 0.39, p < 0.001) and PD-related quality of life (PDQ-39, r = -0.36 p = 0.001 – lower scores = better HRQoL)

In care partners, medium to high correlations show higher resilience was related to higher relationship satisfaction (RSS, r = 0.35,p = 0.002), better mental health (SF-12-MCS, r = 0.55, p < 0.001), and higher quality of life (EQ5D, r = 0.38, p ≤ 0.002), as well as lower burden (ZBI r = -0.44), stress (Rel.SS r = -0.44), anxiety (HADS r = -0.65), and depression (HADS r = -0.54) (all at p < 0.001)

Multiple regression (note: the methods are unclear as to how this was undertaken). In people with Lewy body-related cognitive disorders, resilience was predicted by anxiety level (F(1,74) = 19.97, p < 0.001, adjusted R2 = 0.20), relationship satisfaction (F(1,74) = 4.21, p < 0.05, adjusted R2 = 0.04), quality of life (EQ5D-VAS: F(1,74) = 8.51, p < 0.01, adjusted R2 = 0.09), and Parkinson’s-related quality of life (PDQ-39: F(1,74) = 11.39, p < 0.01, adjusted R2 = 0.12)

In care partners, resilience was predicted by: anxiety (F(1,74) = 64.859, p < 0.001, adjusted R2 = 0.460), depression (F(1,74) = 31.849, p < 0.001, adjusted R2 = 0.291), overall mental health (SF12-MCS: F(1,74) = 31.009, p < 0.001, adjusted R2 = 0.286), stress (Rel.SS: F(1,74) = 27.290, p < 0.001, adjusted R2 = 0.260), and care burden (ZBI: F(1,74) = 24.749, p < 0.001, adjusted R2 = 0.240)

  Wuttke-Linnemann et al. [46]

(To examine associations among depressive symptoms, partnership quality, and individual resilience in PWD and their caregivers from an intrapersonal, interpersonal, and dyadic perspective and examine the incremental variance explained by interpersonal and dyadic parameters concerning each dyad member’s resilience)

Study 1: N = 13 spousal carers, 12 females, mean age = 70.31 (SD = 7.57. N = 13 people with dementia, 12 males, mean age = 75.85, (SD = 4.69)

Study 2 N = 16 spousal carers, 11 females, mean age = 74.75 (SD = 6.79). N = 16 people with dementia, 5 females, mean age = 76.94, (SD = 6.75)

Type of dementia: Alzheimer’s Disease

Germany (German)

Cross-sectional secondary analysis of data from two intervention studies

H1: The prediction of each person’s resilience score can be incrementally increased by adding the scores of the respective partner

H2: Dyadic interdependencies among the dyad in self-rated depression and self-rated partnership quality predict individual perceptions of resilience

Reliability not reported in this sample

Carers: there were no significant correlations between resilience, depression and the marital quality questionnaire

People with dementia: medium and large correlations between resilience and depression (r = -0.44, p < 0.05) and marital quality (r = 0.52, p < 0.01)

H1: not supported in the carers or the person with dementia – their individual resilience was not enhanced by their spouses data

H2: The authors indicate the interdependencies were calculated by creating a ‘similarity score’ by calculating the negative squared difference between two measures. These were then standardised as Z scores. They report a similarity in depression scores is associated with lower individual resilience in caregivers and with higher individual resilience in patients

Resilience Scale 14 (RS-14)

  D’Onofrio et al. [47]

(To illustrate the key results and evidence obtained in the final evaluation of the Mario robot)

N = 38 people living with Alzheimer’s Disease (M = 14; F = 24). Age range 55–93 years (m = 77.08 ± 9.91 years). Ethnicity not reported

Galway, Ireland (English)

Rotondo, Italy (Italian)

Stockport, England (English)

Feasibility study with pre-post evaluation of a robot intervention for people living with Alzheimer’s Disease

(Face to face interview)

None specified

Reliability not reported in this study sample

No correlations reported between resilience and other measures

Not hypothesised: The MARIO robot significantly increased resilience scores between pre (M = 31.33, SD = 21.45) and post (M = 36.96, SD = 15.35) intervention (p = 0.02)

  McGee et al. [48]

(To adapt and evaluate the psychometric properties of positive psychology measures for people living with dementia)

N = 36 people living with early-stage dementia (most frequent diagnosis being Alzheimer’s Disease). Age range 56 to 93 (M = 74.39,SD = 10.70). 61% were female and 67% were married. 33% (n = 12) were educated to Bachelor level or above

Southern USA (English)

Cross sectional design

(Self-completion)

The authors do not define specific hypotheses, but state they test the convergent validity of their positive psychology measures (resilience, optimism, meaning in life, gratitude, life satisfaction) and discriminant validity between the positive psychology measures, depression and anxiety (p.311)

Ideal internal consistency (α = 0.81)

Medium correlations between resilience and the ‘presence’ subscale of the meaning in life (r = 0.48, p < 0.01); optimism (r = 0.38, p < 0.05); and gratitude (r = 0.39, p < 0.05), and large correlations between resilience and depression r = -0.54, p < 0.01) and anxiety (r = -0.72, p < 0.01)

Resilience did not correlate with life satisfaction or the ‘search’ sub-scale of the meaning in life scale

  Orgeta et al. [49]

(To evaluate the clinical effectiveness and cost-effectiveness of carer-delivered individual cognitive stimulation therapy for people with dementia and their family carers, compared with treatment as usual)

N = 356 caregivers. 73% (n = 261) female; 92% (n = 321) white ethnicity; 66% (n = 236) living with spouse/partner with dementia; mean age = 65.73 (SD = 12.92)

Type of dementia: 64% (n = 227) Alzheimer’s Disease; 11% (n = 40) vascular dementia; 3% (n = 11) Lewy Body; 12% (n = 41) unknown

England and Wales, UK

(English)

A multicentre, single-blind, randomised controlled trial assessing clinical effectiveness and cost-effectiveness. Assessments were at baseline, 13 weeks and 26 weeks (primary end point)

( Face to face interview)

iCST will improve the primary and secondary outcomes (including resilience) compared to TAU

Reliability not reported in this study sample

There were no significant differences in resilience over time between carers accessing iCST and carers receiving treatment as usual

  Sánchez-Teruel et al. [50]

(To identify the variables that predict a high degree of well-being in family caregivers of people with dementia during the Covid-19 lockdown)

N = 320 carers; N = 266 females; age range 20–73, mean age = 46.45, SD = 15.97). N = 205 were educated to degree level/vocational training or higher

Type of dementia as described by authors: AD = 82, ‘senile’ = 88, Parkinson = 67, vascular = 63, other = 10

Spain (Spanish)

Cross-sectional design

(Self-completion/online survey)

None stated

Ideal internal consistency (α = 0.88)

Not hypothesised:

Large correlations between resilience and well-being (r = 0.92; p < 0.01), self-efficacy (r = 0.78; p < 0.01), coping strategies (r = 0.65, p < 0.01), self-compassion (r = 0.59, p < 0.01) and difficulties in emotion regulation (r = -0.88, P < 0.01)

Multiple regression found resilience predicted well being along with other variables but the lack of hypotheses make the interpretation difficult

  Stansfeld et al. [51]

(To undertake a psychometric evaluation of the Sense of Coherence Scale)

N = 583 caregivers. 80.3% were female

Age range between 18–89, M = 59.5 years SD = 12.3. 94% were white British or Irish. 69% were married, 59% were the son or daughter and 30% the spouse

Type of dementia:

Alzheimer’s disease (50.5%), vascular (18.9%), DLB (3.3%), FTD (23.3%)

UK (English)

Cross-sectional mixed methods design

( Self-completion)

Sense of coherence will be positively correlated with

caregiver’s resilience

No hypotheses regarding expected differences in resilience between groups

Reliability not reported in this study sample

As hypothesised, carers resilience positively correlated with sense of coherence (r = 0.56, p < 0.001)

Not hypothesised: small correlation between resilience and sense of competence (r = 0.25, p < 0.001); medium correlation between resilience and self-efficacy (r = 0.45, p < 0.001) and high correlation between resilience and health related quality of life (r = 0.56, p < 0.001)

  Wilks et al. [52]

(To explore the moderating effects of spiritual support on the relationship between caregiver burden and resilience)

N = 684 caregivers, 80% were female. Mean age = 61, range not reported. 62% (n = 426) were married. 62% (n = 424) reported their ethnicity as Caucasian 36% (n = 246)

reported as African American. 51% (n = 350) were the child of the care recipient

Cross sectional secondary data analysis

(Self-report)

There will be no significant ethnic difference in the

relationship between care giving burden and the resilience outcome

There will be no significant ethnic difference regarding the relationship

between spiritual support and the resilience outcome

Spiritual support will not serve as a significant moderating factor among the risk- resilience relationship among African American caregivers. A similar, non significant result for

Caucasian caregivers will be found

Reliability not reported in this study sample

Spiritual support significantly predicted resilience in African American carers (β = 0.36, p < 0.01) and Caucasian carers (β = 0.01, p < 0.05)

Caregiver burden did not predict resilience in either ethnic group

[The methods for the analyses in this paper are unclear]

  Wilks et al. [75]

(To examine whether caregiver coping strategy moderates the relationship between aggression in Alzheimer’s and caregiver resilience; whether aggression is associated with specific, caregiver coping strategies; whether aggression is associated with diminished caregiver resilience)

N = 419 (330 female, 86 male)

Mean age = 61

Caucasian/White = 57%, n = 236, African American/Black = 41%, n = 171children of care recipients = 52%, n = 215, spouse/partners = 15%, n = 64 Stage of AD = early (n = 72, 18.4%), middle (n = 156, 39.9%), late (n = 161, 41.2%)

Dementia type: Alzheimer’s disease

US (English)

Cross sectional design

(data obtained via self-report and Likert scales)

None specified

Ideal internal consistency (α = .94)

All strategies of coping correlated significantly with RS14 scores. Specifically, RS-14 scores positively correlated with CITS-task (r = 0.39, p < 0.01) and negatively correlated with CITS-emotion (r = -0.19, p < 0.01), CITS-avoidance (r = -0.16, p < 0.01) and RMPBC aggression (r = -0.11, p < 0.05) scores

Reported significant interaction of RMBPC aggression × CITS-emotion (β =  − .15). and RMPBC aggression × CITS-avoidance (β =  − .12) on predicting RS14 scores

Instead of moderation, the negative effect of RMBPC aggression (β =  − .11) increased upon the interaction of CITS-emotion and CITS-avoidance, respectively

No interaction of RMBPC aggression × CITS-task on predicting RS14 scores (β = .01)

Task-focused coping accounted for the most variability in RS-14 scores (23%) which more than doubled the effect sizes of emotion-focused (11%) and avoidance-focused (10%) strategies

Connor-Davidson Resilience Scale (CD-RISC)

  Duran-Gomez et al. [53]

To assess the resilience of caregivers of people with AD

N = 140 carers; 104/86.7% female; mean age = 50.5, SD = 4.2. 74% married, 68 lived with the person they provided care for

Badajoz, Spain (Spannish)

Cross sectional design

(Face to face interview)

High scores in resilience will be associated with exposure to a lower number of stressors derived from care, perceived stressors caregiver assessment, and adequate physical and psychological state of health and better quality of life

Resilience will be related to other mediator variables, such as perceived social support and the intrapsychic resources of the caregiver (coping styles, self-esteem or sense of competence)

Reliability not reported in this study sample

Carers mean score = 69.24 ± 14.07. A cut-off score of 70 identify highly resilient caregivers (= / > 51.66% of the sample)

Medium correlations between resilience and lower levels of dependency (r =  − 0.417, p < 0.01), care recipient’s cognitive decline (r =  − 0.393, p < 0.01) and large between resilience and carer burden (r =  − 0.623, p < 0.01), but not with NPI scores

Resilience correlated with anxiety (r = -0.33, p < 0.01); depression (r = -0.51, p < 0.01), HRQoL (r = 0.58, p < 0.01); role-physical (r = -0.19, p < 0.05) but not any other of the SF-36 domains

Resilience correlated with self esteem (r = 0.04, p < 0.01); social support (r = 0.22, p < 0.05); emotional support (r = 0.31, p < 0.01); positive social interaction (r = 0.20 p < 0.02), problem focussed coping (r = 0.36, p < 0.01) but not emotion focussed or avoidance focussed coping. All significant variables were entered in a linear regression, and burden, anxiety, coping, social support, cohabitation, help provided and HRQoL all predicted resilience. Effect sizes not reported for regression

  Gómez-Trinidad et al. [54]

(To analyze the relationship of resilience and emotional intelligence with functional performance in the main caregivers of people with dementia in Spain according to the severity of the disease)

N = 144 carers; 70% female, 79% married, mean age not presented

67% of the care recipients had Alzheimer’s Disease, with 61.1%(88) in the moderate stages and 25.7%(37) severe

Spain (Spanish)

Cross-sectional design

(Self-completion)

None stated

Reliability data not reported in this study sample

Not hypothesised; Small correlations between resilience and longer time spent on self-care (r = 0.196; p = 0.019) and leisure (r = 0.172; p = 0.040). The time dedicated to productivity was not related to the level of resilience (r = 0.091; p = 0.278). These variables did not correlate when looking specifically at carers of people with mild phase dementia; for carers of people in the moderate phase: small correlations between resilience and the time dedicated to self-care (r = 0.227; p = 0.033) and leisure (r = 0.262; p = 0.014). For carers of people with severe dementia, a medium correlation between resilience and the time dedicated to productivity (r = 0.355; p = 0.034)

  Lavretsky et al. [55]

(A double-blinded placebo-controlled trial to investigate the

efficacy of an antidepressant drug, Escitalopram, to

improve depression, resilience to stress, and quality

of life in depressed family dementia caregivers)

N = 40 caregivers (age range 45–91, 25 adult children and 15 spouses; 26 women)

Type of dementia: Alzheimer’s Disease

California, USA

(English)

A double-blinded placebo-controlled trial

(Mode of data collection unclear)

Escitalopram would improve resilience in those receiving the drug compared to those in the control arm

Reliability data not reported in this study sample

The authors note participants who took Escitalopram showed an improvement in resilience, but the data is not presented/unclear

  Rivera-Navarro et al. [56]

(To validate the Caregiver Abuse Screen (CASE) as an instrument for detecting the maltreatment of people with dementia in Spain)

N = 326 carers. Most were women (67.2%) and offspring or spouses (93.9%), mean age = 60.1 years (SD = 14.5)

Northwest Spain (Spanish)

Cross sectional design

(Face to face interview)

The CASE may be a reliable instrument to measure different components of maltreatment (i.e., interpersonal abuse and neglect) in Spain

A positive and statistically significant relationship between risk factors of maltreatment (e.g., burden, anxiety and depression in caregivers, functional dependence and cognitive impairment in people with dementia) and CASE scores, will be found

Caregiver resilience and caregiver social support, as protective factors, will show a negative association

Reliability data not reported in the study sample

As hypothesised, medium correlations were found between resilience and the CASE (r = -0.350, p < 0.01) and the CASE subscales interpersonal abuse (r = -0.30 p < 0.01) and a small correlation with neglect/dependency (r = -0.28, p < 0.01)

  Ruisoto et al. [57]

(To examine different predictive factors of burden in a sample of family caregivers of patients with dementia (PWD))

N = 283 carers. N = 186 females; mean age = 59.93 (SD = 65.72). N = 157 were children/children in law and N = 115 were spouses

Spain (Spanish)

Cross sectional design

(Face to face interview)

None stated

Reliability not reported in this study sample

Not hypothesised: a small negative correlation between resilience and burden (r =  − 0.218, p < .001) and resilience predicted burden in a regression model. The authors test a model where social support mediates the relationship between resilience and burden, but it isn’t clear why this model was tested

  Serra et al. [58]

(To understand the associations between abuse related behaviour and key characteristics of dementia carers and the person they care for)

N = 283 caregivers who lived with the person with dementia (Son/daughter = 115/40.6; husband/wife = 157/55.5%)67% were female. Mean age = 59.9 ± 14.6

Type of dementia: Alzheimer’s Disease (85%); ‘other’ 14.5%

Castilla and León, Northwest Spain (Spanish)

Cross-sectional design

(Some data collected through an interview, but unclear how the resilience scale was administered)

Resilience is

expected to decrease the possibility of abuse

No hypotheses regarding expected differences in resilience between groups

Reliability data not reported in the study sample

Abuse scores measured the CASE correlated negatively with resilience (r =  − .35, P < .001), and in a linear regression, carer resilience predicted lower levels of abuse (β = -0.28, p < 0.001)

  Wilks & Vonk [76]

(To determine whether private prayer acts as a mediator for caregiver burden and perception of resiliency in Alzheimer’s caregivers)

N = 304 (233 female)

Mean age = 63 (SD = 13.5)

86% White (n = 261), 13% African American (n = 40), 1% Hispanic (n = 2)

married (79%, n = 240),10% divorced (n = 31), 5% single (n = 14), 5% widowed (n = 14

spouses (43%, n = 131), 39% children(n = 118), 4% friends (n = 13), 14% “other” (n = 42)

Type of dementia: Alzheimer’s Disease

US (English)

Cross sectional design

(Self report: Questionnaires)

None specified

Reliability not reported in the study sample

Participants reported a moderate-to-high level of perceived resiliency. Burden correlated negatively with perceived resiliency (r = -.53, p < .01), with higher burden associated with lower resiliency. Burden accounted for approx. 16% of the variation in perceived resiliency. Further, as prayer coping increased, so did perceived resiliency (r = 0.23, p ≤ 0.05)

With the inclusion of private prayer, the indirect effect of caregiving burden on perceived resiliency decreased (-.49), with private prayer accounting for 7% of variation in perceived resiliency

Burden and private prayer accounted for 23% of the variation in perceived resiliency

Connor Davidson Resilience Scale 10 (CD-RISC 10)

  Bravo-Benitez et al. [59]

(The objectives were to adapt a grief intervention

program to family caregivers of patients with dementia and

assess its effectiveness in improving their symptoms of grief and

other health-related variables)

N = 52 family caregivers

Mean age = 63.88 years

(SD = 17.55; range: 21–89), 21.15% were male and 78.85% were female. Relationship: 57.69% spouses, 34.62% children, and 7.69% were other relatives. 7.69% no education, 26.92% had primary education, 19.23% had secondary education, and 46.15% had higher education

Grenada, Spain (Spanish)

A repeated measures quasi-experimental

randomized design with allocation of participants

to either the intervention group (IG) or to the control group

(CG) (on a waiting list). Randomisation process not described

(Interviews undertaken at the Association of Relatives of patients with AD Centre in Grenada)

No hypotheses stated in relation to the resilience measure

The authors state “It was expected that caregivers who participated in this intervention program would exhibit significant improvements in their overall perceived health,

quality of life, as well as a significant decrease in maladaptive manifestations associated with grief.”

Reliability not reported in this study sample (the authors cite the original development paper)

Not hypothesised: significant differences were found in the Time × Group interaction [F(1, 50) = 16.961; p < 0.001]. A decrease in resilience was observed in the CG between the pre (M = 29.60) and post (M = 26.04) assessment, and increased in the IG between pre (M = 23.74) and post (M = 27.89) assessment. The groups are not balanced at baseline and there is no main effect of time

  Carbone et al. [37]

(1: To explore

changes due to the COVID-19 lockdown in the BPSD of community-dwelling PwD and the distress experienced by their

family caregivers; 2: to explore the associations between caregivers’ ratings of the frequency and severity of their relative’s BPSD and of their own related distress; 3: to explore caregivers’ perceived social and emotional loneliness, and resilience, i.e., the ability to cope with adversity,

and adapt to the physical and psychological challenges of

caregiving)

N = 35 family caregivers of community dwelling people with dementia. Relationship: N = 34 were family members (spouses, children, or siblings), while one was

a paid living-in carer. N = 26 female (74.28%)

Type of dementia: Alzheimer’s disease 17.1%; vascular dementia

37.1%, and mixed or other types of dementia 60%

Trevisio, Italy (Italian)

Cross-sectional design

(Telephone interview)

The authors state:

“We expected a greater degree of resilience to be associated with lower caregiver ratings, and fewer reported lockdown-related changes

in the frequency and severity of the BPSD in their RwD, and in

their own related distress.”

Reliability not reported in this study sample

Not all hypothesised relationships confirmed. The authors found a medium correlation between resilience and changes in total NPI caregiver distress scores (r = -0.32, p < 0.01). The authors suggest this indicates that greater resilience was associated with a more limited worsening under lockdown of the distress experienced by caregivers regarding the BPSD of their RwD

Not hypothesised: female gender was associated (medium) with higher resilience scores (r = 0.32, p < 0.05)

  Sarabia-Cobo and Sarria [60]

(To examine Sense of Coherence, Resilience and Emotional Regulation as predictors of satisfaction with care in caregivers of older adults people with dementia)

N = 63 caregivers; 85.7% were women (n = 54), 82.5% were unemployed (n = 52). The mean age was 63.40 (SD: 14.92)

Cross-sectional design

(Telephone or Zoom interview)

Higher levels of Sense of Coherence, Resilience and Emotional Regulation in caregivers will be associated with greater satisfaction with caregiving

Adequate to ideal internal consistency (α = 0.77)

Satisfaction with care was negatively associated with Resilience (r =  − 0.65, p < .001). The authors state the negative correlation is due to the lower values of the satisfaction scale which relates to greater satisfaction (higher scores = lower satisfaction)

Resilience was associated with Sense of Coherence (r = 0.97, P < 0.05); and Emotional Regulation (r = 0.25, p < 0.05)

Dispositional Resilience Scale

  O’Rourke et al. [61]

(To determine if three aspects of psychological resilience (commitment, control, challenge) of dementia carers predicts lower levels of depression one year later; to ascertain if change in resilience occurs simultaneously with change in depression between measurement points)

N = 105 cohabiting spouses (n = 58 wives/n = 47 husbands). Mean age = 69.59 years (SD = 8.66; range 46–89). M = 14.36 years of formal education (SD = 3.27)

Type of dementia: Probable or possible Alzheimer’s disease

Vancouver, Canada (English)

Longitudinal cohort design

(Mode of data collection unclear)

Psychological resilience would (inversely)

predict depressive symptoms one year later; in addition, a reported increase in resilience between points of measurement would correspond to a further reduction

in depressive symptoms over this interval

No hypotheses regarding expected differences in resilience between groups

Reliability not reported in this study sample

The authors hypotheses were partially supported. Higher control scores at baseline predicted lower levels of depression a year later (ƴ = -1.17, SE = 0.28, p < 0.005). Higher challenge scores at baseline predicted lower levels of depression a year later (ƴ = -1.77, SE = 0.46, p < 0.005). The increase in the challenge score between the two time points predicted a reduction in depressive symptoms over time

( ƴ = -7.74, SE = 3.38, p < 0.05). No significant findings for commitment as a predictor of depression at time 2, or as a predictor of change in depression scores over time

Resilience Scale for Adults (RSA)

  Altieri & Santangelo [62]

(To explore changes in depression and anxiety in caregivers of people with dementia during the Italian Covid-19 lockdown, and the extent to which these differed by low and high resilience)

N = 84 caregivers (71 females; 13 males). Relationship: 72.6% children, 11.9% spouses, 8.3% grandchildren, 6% other. 75% lived in same house. Mean age 48.7 years (SD = 11.7). Type of dementia 56% AD, 31% VD, 10.7% FtD, 2.4% LBD

Italy (Italian)

Cross-sectional design

(Online survey)

None stated

Reliability not reported in this study sample

Repeated measures ANOVA found a significant: 1) main effect of resilience level between subjects (L = 0.795, F(2,81) = 10.445; p < 0.001, partial h2 = 0.969), 2) main effect of time (L = 0.865, F (2,81) = 6.297, p = 0.003, partial h2 = 0.135), and 3) interaction between time and resilience (L = 0.910, F (2,81) = 4.013, p = 0.022, partial h2 = 0.090)

Univariate analysis found and a significant interaction between time and resilience on HADS-A scores (F(1,82) = 6.955, p = 0.010, partial h2 = 0.078) but not on HADS-D scores (F (1,82) = 2.987, p = 0.088, partial h2 = 0.035). There was a significant between subjects effect of resilience levels on HADS-D (F(1,82) = 19.644, p < 0.001, partial h2 = 0.193) and HADS-A (F(1,82) = 7.811, p = 0.006, partial h2 = 0.087) scores

Higher Caregiver Burden Inventory scores were negatively associated with RSA scores (b = -.398, t(81) = 3.644, p = 0.001)

  Elnasseh et al. [63]

(To examine whether healthier family dynamics were associated with higher personal strengths of resilience, sense of coherence, and optimism among dementia caregivers in Argentina)

N = 105 family caregivers (relationship not reported, status is ‘primary’ caregiver). 74% were female, mean age 57.71 years (SD = 13.35). 75% female, 32% completed college

Type of dementia not reported

Rosaria, Argentina (Spanish)

Cross-sectional design

(Face to face interviews)

Healthier family dynamics will be

related to a higher sense of coherence, greater resilience,

and more optimism

Greater family communication is hypothesized to be asssociated

with resilience

Empathy is hypothesized to be associated with resilience

No hypotheses regarding expected differences in resilience between groups

Ideal internal consistency

(α = 0.96). The authors also note that ‘Adequate reliability has been shown for each subscale with α’s ranging from 0.67 to 0.90’. It is not clear if this relates to the current study, but in the absence of a reference to another study, we assumed it relates to the current study

Medium correlations between caregiver resilience and empathy (r = 0.33, p < 0.01); cohesion (r = 0.38, p < 0.01); communication (r = 0.40, p < 0.01); family satisfaction (r = 0.42 p < 0.01) and sense of coherence (r = 0.423 p < 0.01), and a small correlation between resilience and supporting the hypotheses

flexibility (r = 0.22, p < 0.05); Caregiver resilience also correlated with income (r = 0.33, p < 0.001)

  Gulin et al. [64]

Aim: To examine the influence of resilience, coping and optimism on the quality of care provided by dementia carers in Argentina and Mexico

N = 110 caregivers from Argentina and N = 20 from Mexico. Mean age = 56.84 years (SD = 13.18). 77.7% female. Relationship: 43.8% spouses, 43.1% children, 7.7% uncles/aunts, 2.3% ‘other’, 1.5% professional caregivers, 0.8% friends, 0.8% parents

Type of dementia: Not reported

Rosario, Argentina and Baja California, Mexico (Spanish)

Cross-sectional design

(Face to face interviews)

Greater personal strengths will be associated with higher quality of informal care

No hypotheses regarding expected differences in resilience between groups

Ideal internal consistency (α = 0.95). The analysis does not use the full scale but examines each of the 5 sub-scales and no reliability statistics are presented for the sub-scales

Of the five resilience sub-scales, family coherence predicted quality care (β = 0.33, SE = 0.04, p < 0.001) and respectful quality of care (β = 0.34, SE = 0.06, p < 0.001), suggesting minimal support for their hypothesis

  Senturk et al. [65]

(To investigate

the relationship between caregiver burden and

psychological resilience of caregivers individuals

with dementia)

N = 103 caregivers. The mean age of caregivers is 56.5 ± 9.91 85.4% of them are female, 42.7% of them have an undergraduate degree, 42.7% of them provide

care for their mother,

Turkey (language not stated)

Cross-sectional design

(Mode of data collection not stated)

None specified

Ideal internal consistency (α = 0.82)

There was a large negative correlation between caregiver burden and psychological resilience (r = -0.869, p < 0.001)

  Pandya [66]

(To examine the impact of a long-term meditation program in enhancing self-efficacy and resilience of home-based caregivers of older adults with Alzheimer’s in two South Asian Cities)

185 carers at the pre-test phase (96 in the intervention group and 89 in the control group) and 145 participants at the post-test phase (78 intervention group and 67 control group)

Mean age pre test control group 52.5(10.67) intervention group 52.68(11.03), post test control group 57.6(11.32) intervention group 58.02(10.38)

Females pre test control group N = 77/86.52%; pre-test intervention group N = 78/81.25%; post test control group N = 59/88.06%; N = 67/85.90%

Male 12 13.48 18 18.75 8 11.94 11 14.10 Female 77 86.52 78 81.25 59 88.06 67 85.90

Mumbai, India and Kathmandu, Nepal (English)

Intervention study with post-intervention data collected five years later

(Mode of data collection unclear)

Hypothesis 1: Meditation program would reduce the perceived caregiving burden overload and enhance self-efficacy and resilience of home-based caregivers of older adults with Alzheimer’s. Hypothesis 2: It is hypothesized that the intervention group participants’ characteristics as well as program-related characteristics would predict different levels of self-efficacy and burden. Hypothesis 3: Home practice by caregivers would be the strongest predictor of perceived caregiving burden reduction, caregiving self-efficacy enhancement and increased resilience

Cronbach’s α = .92; item-scale intercorrelation = .89; P

The authors state there was no significant difference between intervention and control groups on socio-demographic and caregiving characteristics

Post-test RSA scores were higher in the intervention than the control group (mean difference = 87.92, p = .001, d = 7.55) and their own pre-test scores (mean difference = 87.62, p = .001, d = 7.36). The authors state “Post-hoc analyses indicated that the post-test perceived caregiver burden overload scores were lower and self-efficacy and resilience scores were higher for women caregivers, caregivers who were spouses of older adult patients, Hindus, middle class, with college and higher education, homemakers, who attended at least 187 (i.e. at least 75%) meditation lessons and regularly practiced at home (i.e. once weekly for 187 weeks or at least 75% weeks)

  Sutter et al. [42]

Aim: To examine the relationships between personal strengths (optimism, sense of coherence, resilience) and the mental health of dementia caregivers from Latin America

N = 127 family caregivers (n = 107 from Argentina and n = 20 from Mexico).72% (n = 98) female; 82% (n = 100) married; mean age = 57.14 (SD = 13.01). Relationship: 60% siblings; 22% child; 15.6% spouse; 2.1% partner

Type of dementia: Not reported

Instituto de Neurociencias de San Lucas,Argentina

Baja California, Mexico

(Not reported)

Cross-sectional design

(Self-report and face to face interview)

None specified

Reliability not reported in this study sample

The authors do not present data for the total score of the RSA, but the five sub-scales. They also included another measure of resilience (the BRS)

the Brief Resilience Scale correlated with the five RSA sub-scales: personal competence (r = 0.48, p < 0.01); and small correlations with social competence (r = 0.23, p < 0.01); family coherence (r = 0.27, p < 0.01); social support (r = 0.26, p < 0.01); personal structure (r = 0.25, p < 0.01);

  Trapp et al. [67]

Aim: To examine whether personal strengths (optimism, sense of coherence, resilience) were associated with mental and physical health related quality of life in dementia caregivers

N = 130 dementia caregivers (n = 20 in Mexico n = 110 in Argentina). 77.7% were female, 76.9% were married. The mean age was 56.84 years (SD = 13.18)

Instituto de Neurociencias de San Lucas,

Argentina

Baja California, Mexico

(Not reported)

Cross-sectional design

(Telephone interviews in Mexico, interviews in Argentina)

Personal strengths will be positively associated with both mental and physical health related HRQOL

No hypotheses regarding expected differences in resilience between groups

Reliability not reported in this study sample

A small correlation between resilience and mental health (r = 0.24, p < 0.01); medium correlations between resilience and sense of coherence (r = 0.41, p < 0.001); optimism (r = 0.48, p < 0.001) but not physical health. Linear regression of the personal strength measures found resilience predicted mental health (β = -0.25, p < 0.05), but the negative direction of the association is in contrast to the authors hypothesis. The results partially confirm the hypotheses

  Trujillo et al. [68]

(To use structural equation modelling (SEM) to investigate the role of family dynamics and personal strengths in the mental health of dementia caregivers from Latin America)

N = 110 carers; n = 83 female, mean age = 57.20 (SD = 13.47). Main relationship to person with dementia – spouse (n = 54), children (n = 50). N = 87 married

Rosario, Argentina (Spanish)

Cross sectional design

(Face to face interview)

It is hypothesized that personal strengths will mediate the relationship between family dynamics and caregiver mental health in a sample of dementia caregivers from Argentina

Ideal internal consistency (α = 0.97)

Large correlation between resilience and satisfaction with life (r = 0.52, p < 0.01)

Medium correlations between resilience and empathy (r = 0.37, p < 0.01), pathology (r = -0.41, p < 0.01), cohesion (r = 0.31, p < 0.01), communication (r = 0.35, p < 0.01), family satisfaction (r = 0.35, p < 0.01), sense of coherence (r = 0.44, p < 0.01), optimism (r = 0.49, p < 0.01), depression (r = -0.42, p < 0.01), anxiety (r = -0.40, p < 0.01)

Small correlation between resilience and flexibility (r = 0.18, p < 0.01)

The latent variable personal strengths, which included resilience, optimism and coherence was associated with mental health (β1.48, p < 0.001). Bootstrapping analyses with 2,000 calculated samples found that the indirect effect of family dynamics on caregiver mental health through personal strengths was also statistically significant (β = .99, p < .001)

Brief Resilient Coping Scale (BRCS)

  Jones et al. [69]

Aim: To describe the demographic and psychosocial characteristics of caregivers, and which of these aspects may influence dementia café attendance)

N = 80 family caregivers of people living with a dementia. n = 21 male, n = 59 female. N = 52 were spouses, n = 28 ‘other’. Age range 30–80 + (no mean age reported)

Type of dementia: N = 26 Alzheimer’s Disease; n = 12 vascular; n = 20 mixed (AD and vascular); n = 4 DLB/FTD; n = 6 ‘other’

Norfolk, England (English)

Cross-sectional between group design

( Self-completion)

Cafe attendees will ́

have greater wellbeing, resilience, and social support

than non-attendees

Reliability not reported in this study sample

Caregivers attending a dementia café reported higher resilience than non-attendees as hypothesised (mean difference –3.54, 95% CI –5.34 to 1.73; p < 0.001)

  Jones et al. [70]

(To compare socio-demographic characteristics and the availability of social support for carers with low and high resilient coping, and identify if any domain of social support predicted high resilient coping in informal carers of people with dementia)

N = 108 carers. The majority (69%) were women (69%). 61% of carers were aged 70 years or above (mean age not available for total sample). Spousal relationship was most common (61%) as was carer co-residence with the person with dementia (78%)

Norfolk, England (English)

Cross-sectional design/postal survey

(Self-completion)

1. Carers who report high resilient coping would have greater perceived social support

2. High resilient coping would be associated with emotional/informational support and tangible support in line with qualitative studies

Reliability not reported in this study sample

Low resilient carers reported significantly less availability of emotional/informational support than high resilient carers (Mean rank difference = 20.17, U = 913.00, z = 3.35, p = 0.001) although the effect size was small (h = 0.10). The perceived availability of tangible support was lower for carers who report low resilient coping (Mean rank difference 14.77, U = 1059.00, z = 2.47, p = 0.014) with a small effect size (h = 0.06). Low resilient coping carers perceived they had less availability of positive social interaction than carers who had high resilient coping scores (Mean rank difference = 18.89, U = 947.5, z = 3.175, p = 0.001) and the effect was small (h = 0.09)

Emotional/informational support had greatest influence on high resilient coping (OR = 1.92, 95%CI = 1.29 to 2.88, p = 0.001). Carers with greater access to tangible support were also more likely to be highly resilient copers (OR = 1.43, 95%CI = 1.07 to 1.91, p = 0.017). greater availability of affectionate support (OR = 1.49, 95%CI 1.10 to 2.00, p = 0.010) and positive social interaction (OR = 1.76, 95%CI = 1.24 to 2.49, p = 0.002) predicted high resilient coping. Gender was a significant predictor, with females being more likely to be high resilient copers (OR = 3.45, 95%CI = 0.448.27, p = 0.01)

(The authors categorised high and low resilience groups through a mean score split, which is not the intention of the measure)

  Meléndez et al. [71]

(To ascertain whether patients with MCI and Alzheimer’s Disease experience changes in psychological wellbeing, resilience and coping compared to older adults without cognitive impairment

N = 32 healthy elderly people mean age = 73.9 years, SD = 5.05, range 65–87 years old, 31 amnestic mild cognitively impaired (aMCI) patients mean age = 75.93 years, SD = 6.23, range 64–88 years old, and 32 Alzheimer's disease (AD) diagnosed patients mean age = 76.84 years, SD = 4.57, range 65–83 years old

Valencia, Spain

(Spanish)

Cross-sectional between group design

(Face to face interview)

None specified

Adequate to ideal internal consistency (α = 0.78)

Significant differences in resilience (F(2,94) = 12.67; p = 0.001) between people living with Alzheimer’s Disease (M = 13.19), mild cognitive impairment (M = 15.69) and older people with no impaired cognition (M = 17.53). It is unclear as to why these differences might be expected

Caregiver Resilience Scale (CRS)

  Maneewat et al. (2016) [72]

(1) To develop the caregiver resilience scale for Thai caregivers of older people with dementia. (2) To examine the validity and reliability of the caregiver resilience scale

Pre-test N = 30 carers; field test

N = 150 carers

No demographic data is presented. The inclusion criteria suggests they were between age 20–60 and able to speak Thai

Upper Southern Thailand (Thai)

Study design unclear (two cross-sectional studies?)

Mode of data collection unclear, semi-structured interview is suggested

None stated

Pre-test: The first draft of the CRS was composed of 36 items within six domains of (physical competence, relationship competence, emotional competence, cognitive competence, moral competence, and spiritual competence). This was derived from a scoping review and “semi-structured interview among 10 caregivers in order to confirm the pre-specified structure of caregivers’ resilience and congruence with Thai caregivers’ context” (the data is not presented for either of these)

The draft was examined by three experts who recommended the deletion of six items, which were considered redundant (decision not explained). The authors developed a content validity form for the experts who rated each item (− 1 = not relevant, 0 = somewhat relevant, and 1 = quite relevant). They state that for the 30 items, the content validity index was 0.84, although the range and interpretation is not provided. Internal consistency from the 30 carers α = 0.87

Item and item, item and subscale, and item and total correlations were between 0.56 to 0.88 (data not presented)

Field test: Internal consistency of the CRS was 0.87. The internal consistency of the subscales of the physical competence domain, relationship competence domain, emotional competence domain, moral competence domain, cognitive competence domain, and spiritual competence domain ranged from 0.52 to 0.87 (data not presented). The correlations of items and their subscale ranged from 0.32 to 0.83 (data not presented). The results of Bartlett’s test of sphericity showed a significant high inter—item correlation (χ2 = 17,124.13, p < 0.01). The communalities of the 30 items of factor extraction ranged from 0.40 to 0.78. The six components with initial eigenvalues greater than 1 ranged from 1.43 to 21.34, and the total variances at 63.67 is reported. (Some evidence of structural validity?)

The rationale for the caregiver resilience scale is theoretically under-developed. The authors note there are only two measures for assessing resilience, stating that one (the CD-RISC) has only limited evidence in the general population (which is incorrect, as it is widely used) and another the Responses to Stressful Experiences Scale, has been mainly applied in the military. There are other measures of resilience (as identified in this review)

They correctly note most of the resilience measures have been developed with Western population, and there is a gap for a resilience measure developed with and for Thai caregivers. The authors note the domains of the measure were identified through a concept analysis, although this work is not presented or the article cited. They state qualitative interviews were undertaken with 10 carers to ‘confirm’ the domains suggested by the concept analysis, however this data is not presented

  Pandya et al. [66]

185 carers at the pre-test phase (96 in the intervention group and 89 in the control group) and 145 participants at the post-test phase (78 intervention group and 67 control group)

Mean age pre test control group 52.5(10.67) intervention group 52.68(11.03), post test control group 57.6(11.32) intervention group 58.02(10.38)

Females pre test control group N = 77/86.52%; pre-test intervention group N = 78/81.25%; post test control group N = 59/88.06%; N = 67/85.90%

Male 12 13.48 18 18.75 8 11.94 11 14.10 Female 77 86.52 78 81.25 59 88.06 67 85.90

Mumbai, India and Kathmandu, Nepal (English)

Intervention study with post-intervention data collected five years later

(Mode of data collection unclear)

Hypothesis 1: Meditation program would reduce the perceived caregiving burden overload and enhance self-efficacy and resilience of home-based caregivers of older adults with Alzheimer’s. Hypothesis 2: It is hypothesized that the intervention group participants’ characteristics as well as program-related characteristics would predict different levels of self-efficacy and burden. Hypothesis 3: Home practice by caregivers would be the strongest predictor of perceived caregiving burden reduction, caregiving self-efficacy enhancement and increased resilience

Cronbach’s α = .87; item-scale inter-correlation = .82

Post-test CRS scores of the intervention group were higher than the control group (mean difference = 27.68, p = .001, d = 4.54) and their own pre-test scores (mean difference = 27.5, p = .001, d = 4.65)