Prevalence of Depression and Burden Among Informal Caregivers of People With Dementia: A Meta-Analysis

View metadata, citation and similar papers at core.ac.
uk brought to you by CORE

provided by University of East Anglia digital repository
Depression and burden in caregivers of people with dementia
Prevalence of depression and burden among informal caregivers
of people with dementia: a meta-analysis
Rebecca N. Collinsa and Naoko Kishitab
a
Department of Clinical Psychology, University of East Anglia, Norwich, NR4 7TJ, England,
UK; b School of Health Sciences, University of East Anglia, Norwich, NR4 7TJ, England,
UK
Corresponding author:
Qualifications: BSc, MSc
Mailing address: Department of Clinical Psychology, University of East Anglia,
Norwich, NR4 7TJ, England, UK
E-mail address: [email protected]
Contact telephone number: 07872578788

Abstract
This meta-analysis examined the prevalence of depression and burden among informal
caregivers of people with dementia (PwD) and compared the prevalence of depression
between male and female, and spousal and non-spousal, caregivers. The quality of studies
was evaluated and moderator variables explored. A search of six electronic databases
(PsycARTICLES, PsycINFO, MEDLINE Complete, SCOPUS, Web of Science and
ProQuest) was conducted from the first available date to the 31st October 2017. Inclusion
criteria involved observational studies that detailed the prevalence of burden or depression
among informal caregivers of PwD. Forty three studies were examined with a total of 16 911
participants. The adjusted pooled prevalence of depression was 31.24 per cent (95% CI
27.70% to 35.01%) and burden was 49.26 per cent (95% CI 37.15% to 61.46%), although
heterogeneity among prevalence estimates was high. Depression prevalence estimates
differed according to the instrument used and continent in which the study was conducted.
The odds of having depression were almost one and a half times higher in female compared
to male caregivers. No significant difference was observed between spouses and non-spouses.
Most studies had a medium risk of bias. The results indicate a great need within this
population for interventions that are effective at reducing burden and depressive symptoms. It
therefore appears imperative for dementia services that are not providing interventions
targeting these difficulties to do so.
Keywords: carers; depressive symptoms; caregiver-burden; epidemiology.

Introduction
The number of people with dementia (PwD) is rising every year. By 2050, there will
be approximately 131 million PwD worldwide (Alzheimer’s Disease International 2015). It
has therefore been perceived as one of the greatest problems facing society in the twenty-
first century (Alzheimer’s Society 2014).
The majority of PwD are community-dwelling and cared for by a spouse or an adult
child, typically of the female gender (Alzheimer’s Research UK 2015). The increasing
number of dementia cases means that the number of informal caregivers (unpaid relatives or
friends) of PwD is also increasing. Research indicates that informal caregivers of PwD can
experience positive benefits from the acquisition of the caregiving role, such as feeling as
though family members have come closer together and appraising life as more fulfilling and
meaningful (Cohen, Colantonio and Vernich 2002). However, there is an abundance of
literature that suggests that the role can lead to the presence of perceived burden (e.g. Chiao,
Wu and Hsaio 2015; Brodaty and Donkin 2009) and psychological difficulties. The strongest
evidence base is for the presence of depressive symptoms, that are more severe than those
found in older adults who are not caregivers (Vitaliano 1997) and caregivers of people
without dementia (Pinquart and Sörensen 2003).
Burden
In this review, ‘caregiver burden’ (here on referred to as ‘burden’) is conceptualised
as a multidimensional biopsychosocial reaction (Given et al. 2001) that results from the
caregiver’s perception of the degree to which the care-recipient is dependent upon them and
the caregiving role has had a negative impact upon their emotional health, physical health and
social or financial status (Zarit, Todd and Zarit 1986). Literature has frequently attempted to
distinguish ‘objective’ from ‘subjective’ burden, although this distinction still remains
unclear. The current burden definition is based on that of Zarit, Todd and Zarit (1986) which
has been suggested to include ‘objective burden’ concepts (e.g. physical, social and financial
impacts and level of dependency) and ‘subjective burden’ concepts (e.g. the caregiver’s
perceptions and the emotional impact of caregiving), and is in line with most of the well-
established and validated caregiver burden measures (Vitaliano, Young and Russo 1991).
When taking into account this burden definition and the research comparing the
experiences of caregivers of people with and without dementia, it becomes clear why
caregivers of PwD might perceive greater burden. Caregivers of PwD tend to spend more
hours per week on caregiving tasks, assist with a greater number of activities of daily living,
report more employment complications and less time for leisure and social activities due to
caregiving responsibilities (Ory et al. 1999), and spend more of their own money on
caregiving expenses (O’Brien 2016). In addition to this, many PwD display aggressive
behaviours, the presence of which increases perceived burden (Ornstein and Gaugler 2012).
Interestingly, the higher the burden experienced by caregivers of PwD, the more likely
they are to expedite nursing home placement (Gaugler et al. 2005).
Research exploring burden in caregivers of PwD has tended to focus on the
relationships between burden and psychological constructs such as depression, and predictors
of burden. This has revealed that depressive symptoms and burden are positively correlated
with one another (Epstein-Lubow et al. 2008; Medrano et al. 2014). Moreover, that there are
significant patient related predictors of burden such as the patients’ severity of dementia,
behavioural problems or psychological symptoms and extent of personality change, and
caregiver related predictors including sociodemographic variables and psychological health
(Etters, Goodall and Harrison 2008; Chiao, Wu and Hsaio 2015). These studies have
therefore been significant in uncovering the potential difficulties that may be experienced by
those with perceived burden and the types of factors that increase a caregiver’s vulnerability
to experiencing perceived burden. However, to our knowledge, there has been no meta-
analytic review of the prevalence of burden among informal caregivers of PwD. Determining
this would appear vital to further our psychological understanding of this population and help
inform the provision of services.
Depression
Depressive symptoms can include a persistent sadness/low mood, marked loss of
interest or pleasure in activities, disturbed sleep, decreased or increased appetite or weight,
loss of energy, poor concentration, feelings of worthlessness or guilt and/or suicidal ideation
or acts (American Psychiatric Association; APA 2013). To fulfil the Diagnostic and
Statistical Manual of Mental Disorders-V (DSM-V) criteria for major depression at least one
of the first two symptoms must be present alongside five of the remaining symptoms nearly
every day for at least two weeks (APA 2013). There are numerous self-report measures that
have been designed to map onto the diagnostic criteria for depression, include specified cut-
offs to determine depression, and have been validated in older adult populations. The most
frequently used measure in research on caregivers of PwD is the Centre for Epidemiological
Studies-Depression (CES-D; Radloff 1977).
Caregivers who have depression typically experience problems in daily functioning
and poorer physical health (Gallagher et al. 1989; Cucciare et al. 2010). In addition, a large
cross-sectional study of 566 informal caregivers of PwD revealed that approximately 16 per
cent had contemplated suicide more than once in the previous year (O’Dwyer et al. 2016).
Although a smaller longitudinal study found the prevalence of suicidal thoughts to be
substantially lower than this at approximately five per cent (Joling et al. 2018), both studies
reported depression to be a risk factor for suicidal ideation. Therefore, at least, depression can
compromise a caregivers’ ability to effectively maintain their role and, at worst, it can lead to
suicide; demonstrating why investigating the prevalence of depression among this population
is important.
A previous meta-analysis found a moderately significant difference in depressive
symptoms between informal caregivers of PwD and people who were not caregivers
(Pinquart and Sörensen 2003). This review however did not evaluate the prevalence of
depression among either group. A meta-analysis conducted 13 years ago estimated the pooled
prevalence of depressive disorders among informal caregivers of PwD, assessed via
interviews based on the DSM-III(-R)/IV (APA, 1980; APA, 1987; APA, 1994) or
International Classification of Diseases 10th Revision (ICD-10; WHO 1992). This was found
to be approximately five times higher than that of the general population, at 22.5 per cent
(Cuijpers 2005). A more recent meta-analysis by Sallim, et al. (2015) estimated the pooled
prevalence of depression among caregivers of people with Alzheimer’s disease (AD),
measured via self-report instruments, to be 34 per cent. However, these reviews included
relatively small numbers of studies; ten (Cuijpers 2005) and 13 (Sallim et al. 2015).
A contextual model (Fig 1) by Williams (2005) adapted from that of Dilworth-
Anderson and Anderson (1994) conceptualised the factors that may influence the likelihood
of a caregiver of someone with dementia experiencing depression. Among other factors,
gender and the relationship to the care-recipient were posited to influence this likelihood.
<insert fig 1. here>
Indeed, one meta-analysis found the prevalence of depression to be higher in female
and spousal caregivers of people with AD compared to male and non-spousal caregivers of
people with AD, respectively (Sallim et al. 2015). However, this review was limited to
caregivers of people with AD and, due to the extremely small number of included studies in
each meta-analysis (n = 3) and the lack of assessment of publication bias, findings may not be
robust. It is important to note that using meta-analytic approaches to investigate the influence
of the other contextual factors presented in the adapted model of Williams (2005) on
depression would not be appropriate, given that research often presents these factors as
summary data and conducting moderator analyses on such data would introduce aggregation
bias (Harbord 2010).
There are many psychological interventions that are being delivered to and adapted
for informal caregivers of PwD, such as Compassion-Focussed Therapy (Collins, Gillian and
Poz 2018) and Mindfulness Based Stress Reduction (e.g. Hoppes et al. 2012). Determining
the current prevalence of burden and depression is important to quantify the need for such
programmes and the requirement to develop, adapt, or change the availability of, existing
treatments to fulfil the needs of this client group, and so help delay and reduce rates of
transition into care homes (Alzheimer’s disease International 2013; Gaugler et al. 2005).
The study aimed to address the gaps in the literature on burden and depression by
conducting a current comprehensive meta-analysis with the following objectives:
(1) To quantify the prevalence of burden and depression among informal caregivers
of PwD.
(2) To compare the prevalence of depression among female and male caregivers and
spousal and non-spousal caregivers.
(3) To explore moderator variables including the methodological quality
Method
The meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews
and Meta-Analyses guidelines (PRISMA; Moher et al. 2009).
Eligibility Criteria
Articles were included if they were written in English or Japanese, as both authors are
fluent in English and the second author fluent in Japanese, and used observational study
designs (see Munn et al. 2014) including prospective and retrospective longitudinal cohort
studies, case-control studies, cross-sectional studies and studies that analysed baseline data
from other studies of randomized controlled trials (RCTs). All other study designs were
excluded, such as experimental studies, qualitative studies, and review articles.
The population studied were informal caregivers of PwD. Studies involving
caregivers of people without dementia or professional caregivers (e.g. paid support workers)
were excluded. There were no limitations on the gender or age of the caregivers, the dementia
type of the care-recipients, the setting or time spent as a caregiver. Studies were included if
they sought to recruit a representative sample of its population. Studies were therefore
excluded if they recruited only caregivers with specific mental or physical health difficulties,
or they actively excluded caregivers experiencing a depressive episode.
Similar to the meta-analyses of Krebber et al. (2014) and Wang et al. (2017), studies
were included if they reported the number or percentage of individuals with depression
assessed by semi-structured or structured diagnostic interviews based on criteria by DSM-
III(-R)/IV or ICD-10, or validated self-report measures with specified clinical cut-offs.
Studies were included if they reported the number or percentage of caregivers that scored
above a specified cut-off for burden on a burden measure that was in line with the study’s
definition, and had evidence of high internal consistency, validity, and being an effective tool
for assessing burden in caregivers of PwD. For instance, the Caregiver Burden Inventory
(CBI; Novak and Guest 1989) and the most widely referenced burden measure, The Zarit
Burden Interview (ZBI; Zarit, Reever and Bach-Peterson 1980). The cut off point for the
presence of mild to severe burden on the 22 item ZBI is >21. Studies not reporting depression
or burden prevalence data were excluded.
Initially, articles published in any year were included. However, during the screening
of full text articles the authors decided that only studies published from the year 2000
onwards were eligible for inclusion. This decision was made because a number of factors
have changed substantially from prior to the year 2000 to the present day which could have
impacted upon the accuracy of the current prevalence estimates of depression and burden. For
example, in the 1980’s, older adult services in the United Kingdom (UK) rarely diagnosed
dementia, it was common for PwD to be hospitalised, and there was a lack of psychologically
informed care (Brooker 2017). In contrast, from around the 1990’s there has been an increase
in the formal diagnosis of dementia and a shift towards community based care, with most
PwD today living in the community and receiving care from a relative or friend (Schulz and
Martire 2004). The evidence base for and provision of psychosocial and psychological
interventions (e.g. Cognitive Simulation Therapy; Spector et al. 2003) has also grown. Other
factors taken into account included life-style changes and technological advances, the
increase in the prevalence of depression in the general population (WHO 2017), and the
reduction in stigma towards depression in the last 20 years (Taylor Nelson Sofres British
Market Research Bureau Limited 2014) - potentially increasing the likelihood of caregivers
disclosing depressive symptoms.
Information sources
A comprehensive search of the literature was conducted. The databases of
PsycARTICLES, PsycINFO, MEDLINE Complete, SCOPUS and Web of Science were
searched to identify relevant published articles. Unpublished articles including dissertations
and theses were sought through the ProQuest global database. Hand searches were performed
on the reference lists of included studies and relevant prevalence reviews and meta-analyses
obtained via The Cochrane Online Library.
Search
The first author performed the search using the keywords and search strategies
outlined in Table 1. All databases were searched from their inception to 31st October 2017
and no limits were applied to language.
< insert table 1 here>
Study selection
The results of the searches were merged using EndNote software (version X8.0) and
duplicate articles removed. Eligibility assessment was conducted in a non-blinded manner.
The first author performed the initial screening of the titles and abstracts, whereby clearly
irrelevant articles were excluded. Full text articles were screened by both authors
independently using a structured checklist created by the first author (Appendix A). The
kappa coefficient was 0.68 indicating substantial agreement (Cohen 1960). Disagreements
between reviewers were resolved through discussions. When data from studies overlapped,
the report with the largest sample size or data set was included.
Data collection process
The first author developed an electronic database which was pilot tested on a
randomly-selected study by both authors collaboratively and refined accordingly. In order to
reduce errors and minimise bias, both authors independently extracted the data from 11 of the
included studies (10%) and results were compared, with no significant discrepancies
identified. Data extraction was completed on the remaining studies by the first author
independently and the data transferred to the Comprehensive Meta-Analysis software (CMA
version 3; Borenstein et al. 2005).
Data items
Information was extracted from each study based on (1) characteristics of the study
(including year of publication, country, design, recruitment process, sample size and
instruments used to assess depression and/or burden); (2) characteristics of the caregivers
(including the definition given for a caregiver, mean age, percentage female, race,
nationalities, average length of time spent caregiving in months, percentage employed,
percentage married, mean years of education and types and percentages of relationships held
with the care-recipients); (3) characteristics of the care-recipients (including procedure used
to diagnose dementia, percentages of the types of dementia diagnoses and severity of
dementia - primarily measured by a mean Mini Mental State Examination (MMSE) score);
(4) depression and burden outcome data (including the number or percentage of participants
within the sample that were diagnosed with depression or scored above the specified clinical
cut-off, and the number or percentage of females and males, and spouses and non-spouses
that were diagnosed with depression or scored above the specified cut-offs). Information was
not inputted if it was missing or unclear and not made available by study authors.
Risk of bias in individual studies
The bias risk of each study was investigated using a 13-item list (Table 2) adapted
from existing criteria lists (Krebber et al. 2014; Luppa et al. 2012). Quality rating scales for
RCT’s tend to generate an overall score of study quality or separate quality scores in key
domains. The assessment tool used in this review measured the level of risk that each study
posed to the reliability of the specific outcomes of the current review. Adaptations were to the
list were therefore made with regards to the population being studied and focused on: (i) the
description of the caregivers including information about the care-recipients’ diagnosis and
(ii) the representatives of this population. Items for the description of the caregivers included
sociodemographic characteristics (age and gender, and at least one of the following four:
marital status, education, employment or socioeconomic status), inclusion and exclusion

criteria, dementia diagnostic procedure, dementia diagnoses and severity, time spent as a
caregiver, inclusion and exclusion criteria and information about (a history) of psychiatric
problems of the caregivers. Items of the representativeness of the study population included
sample size >100, description of participation or response rate and this being at least 75 per
cent, reasons for nonresponse/nonparticipation presented or a statistical comparison of the
characteristics of responders and non-responders, description of the recruitment process and
use of a consecutive sampling method. A risk item was given a positive score if the study
provided adequate information. If the information was incomplete or unclear, a negative
score was given. If a study referred to another publication describing relevant information
about the first study (e.g. recruitment process), the additional publication was obtained to
score the item of concern. For each study, a total bias score was calculated by counting the
number of criteria scored positively; therefore the highest total score available was 13. A
study was considered of low bias risk if the score was at least 75 per cent of the total, of
medium bias risk if it was between 50–75 per cent of the total and high risk if below 50 per
cent of the total.
< insert table 2 here>
The risk assessment tool was pilot tested on a randomly selected study by both
authors collaboratively and refined accordingly. Subsequently, the authors independently
rated eleven randomly-selected studies and compared the results. There were a few
discrepancies between the ratings. If a risk item was rated positively by one author but not the
other, a discussion was held and often the conservative value was chosen. The remaining
studies were assessed by the first author independently.
Summary measures
Meta-analyses were conducted by computing the event rate of depression and burden
using CMA (Borenstein et al. 2005).

Synthesis of results
Effect sizes (event rates), their 95 per cent confidence intervals (CIs) and associated z
and p values were computed using the number of caregivers who scored above the specified
cut-offs for depression or burden and sample size. As considerable heterogeneity of event
rates was expected, the pooled prevalence estimate and its 95 per cent CI were calculated
using a random-effects model. To assess for heterogeneity among studies, the chi squared
statistic (Q; Higgins and Thompson 2002) and I squared statistic (I2; Higgins et al. 2003)
were computed. I2 provides a percentage of the total observed variability in effect estimates
due to heterogeneity rather than chance and is not affected by low statistical power. An I2 of
25 per cent is considered low, 50 per cent moderate and 75 per cent high.
Risk of bias across studies
Publication bias was assessed by constructing funnel plots, and conducting the trim
and fill method (Duval and Tweedie 2000a) and Rosenthal’s Fail Safe N (Rosenthal 1979).
The trim and fill method estimates how many studies could be missing from each meta-
analysis, corrects the funnel plot symmetry, and calculates adjusted effect size estimates.
Rosenthal’s Fail Safe N determines how many studies with a null result would be needed to
nullify the pooled prevalence estimate. If only a few studies (e.g. five or ten) are required to
cause the pooled prevalence estimate to become non-significant caution is held over the
robustness of the results (Borenstein et al. 2009).
Additional analyses
A sensitivity analysis was conducted to determine whether the burden pooled
prevalence estimate would have differed substantially if a study that measured ‘persisting’
burden (Epstein-Lubow et al. 2008) was omitted. As samples enrolled in RCTs could differ
from samples who are not, a random-effects sub-group analysis was performed to determine
whether prevalence estimates differed according to whether studies used a cross-sectional
sample or one taken from an RCT at baseline.
Odds ratio effect sizes, their 95 per cent CI and associated z and p values were
computed on the proportion of female caregivers compared to male caregivers that were
classed as depression, and the proportion of spouses compared to non-spouses that were
classed as depressed. Two meta-analyses using random effects models were conducted to
ascertain the overall odds ratio estimates and their 95 per cent confidence intervals.
A random-effects meta-regression investigated the relationship between study quality
and the prevalence estimates of depression and burden. A random-effects sub-group analysis
was also conducted to determine whether depression prevalence estimates differed according
to the type of measure used to assess depression and the continent the study was conducted
in.
Results
Study selection
The database searches produced 8568 articles and hand searching 35 articles, resulting
in a total of 8603 studies (Fig. 2). After the removal of 1905 duplicates, 6698 titles and
abstracts were reviewed, with 6584 articles deemed clearly irrelevant and excluded. The full
texts of the remaining 114 articles were screened, with 71 not fulfilling criteria and 43 studies
included in the meta-analysis.
One study used a higher cut off for the burden measure compared with other included
studies that used the same measure, as it assessed ‘persisting burden’ rather than the presence
of burden (Epstein-Lubow et al. 2008). The authors included the study and assessed its
potential impact via additional analyses.
<insert fig 2 here>
Study characteristics
The key characteristics of the 43 included studies are provided in Table 3. The total
number of participants included in the meta-analysis was 16 911. Most of the studies were
conducted in Europe (19), followed by North America (16), Asia (3), Australia (3) and South
America (2). The majority of studies used cross-sectional designs (28), with the remaining
studies using baseline RCT data (8), adopting a longitudinal prospective cohort design (4),
and using baseline data from longitudinal prospective cohort studies (3). The recruitment
procedures varied greatly across studies. Sixteen recruited from multiple different platforms.
For example, Cheng, Lam and Kwok (2013) recruited caregivers from memory clinics,
outpatient clinics, day hospitals, day care centres and social services. Seventeen recruited
from one database or service, and 10 recruited from two or more of the same types of service,
such as memory clinics (e.g. Brodaty et al. 2014).
Of the 40 studies that reported the proportionality of genders, all were predominantly
female. Thirty-three studies reported the mean age of the sample (ranging from 51.8 to 83.5
years old). Of the 40 studies that reported the percentages of relationships between the
caregivers and care-recipients, 20 had a majority of spouses and 20 a majority of non-spouses
(typically adult children). Twenty-four studies reported the tools used to diagnose dementia
or a form of dementia in all care-recipients; seven used the National Institute of Neurological
and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders
Association (NINCDS-ADRDA; McKhann et al. 1984) alone or in conjunction with other
diagnostic tools or procedures. Twenty one studies reported the percentages of the care-
recipients’ dementia diagnoses. Ten studies were 100 per cent Alzheimer’s Disease (AD),
five were
<insert table 3 here>
primarily AD followed by Vascular Dementia (VD) then other dementias, one was
primarily AD followed by other dementias then VD, one was 75 per cent AD and 25 per cent
Lewy Body Dementia (LBD), one was a majority of Frontotemporal Dementia (FTD)
followed by AD then other dementias, and one was 100 per cent FTD.
Structured diagnostic interviews were used in two of the 38 studies that reported the
prevalence of depression; leaving 36 studies that used self-report depression measures (Table
3D). The 20-item CES D (Radloff 1977) with cut-off ≥16 was used the most times (11) to
measure depression. Of the nine studies that reported the prevalence of burden, eight used a
version of the 22-item Zarit Burden Interview (ZBI; Zarit, Reever and Bach-Peterson 1980).
Risk of bias within studies
The mean bias score was seven (SD = 1.65), and scores ranged from four (highest risk
bias) to 11 (lowest risk) (Fig. 3a). Of the 43 studies assessed, 18 had a high risk, 22 had a
medium risk and three a low risk. As can be seen in Fig. 3b, over 80 per cent of the studies
reported the percentages of the types of relationships between caregivers and care-recipients,
and inclusion and exclusion criteria. More than half had a sample size ≥100 and reported
sufficient socio-demographic information, the dementia diagnostic procedure, percentages of
dementia diagnoses, dementia severity, and provided an adequate description of the
recruitment method. The most underreported risk items were ‘(history of) psychiatric
problems’ (14%) and ‘participation and response rates are described and are more than 75 per
cent (27%). See Figures 3a and 3b for a full description of the risk bias assessment results.
< insert fig 3a and fig3b here>

Results of individual studies
Fig. 4 and Fig.5 show forest plots of prevalence estimates for burden and depression,
including their CI and associated z and p values.
Synthesis of results
Prevalence of depression
Thirty-eight studies included prevalence estimates of depression. These ranged from
three per cent to 57 per cent; although it must be noted that the study with a three per cent
prevalence estimate (Lowery et al. 2000) had the highest standard error and could be
considered an outlier (Copas and Shi 2000). Overall, prevalence estimates of depression
yielded a pooled prevalence of 33.6 per cent (CI 29.9% to 37.5% p <.001). However, the
heterogeneity of the prevalence estimates was significantly high (I2 = 93.96%, Q = 612.31, p
<.001).
<insert fig 4 here>
Prevalence of burden
Nine studies reported prevalence estimates of burden. These estimates ranged from
35.8 per cent to 88.5 per cent, with a pooled prevalence of 62.5 per cent (CI 51.2% to 72%, p
=.031). However, heterogeneity of the prevalence estimates was significantly high (I2 =
94.90%, Q = 157, p <.001).
<insert fig 5 here>
Risk of bias across studies
Studies on depression
The depression pooled prevalence estimate corresponded to a z value of -28.77
(p<.00001) indicating that 8149 studies with a null effect size would be needed before the
combined two-tailed p-value would exceed 0.05, suggesting that the observed effect
estimates may be extremely robust. The trim and fill method indicated four potentially
missing studies that would need to fall on the left side of the pooled prevalence estimate to
make the plot symmetrical (Fig 6). Assuming a random-effects model, the new pooled
prevalence estimate reduced to 31.24 per cent (CI 27.70% to 35.01%).
<insert fig 6 here>
Studies on burden
The burden pooled prevalence estimate corresponded to a z value of 5.914 (p<.00001)
indicating that 73 studies with a null effect would be needed before the combined two-tailed
p-value would exceed 0.05, suggesting that the observed prevalence estimates may be robust.
The trim and fill method indicated three potentially missing studies that would need to fall on
the left side of the pooled prevalence estimate to make the plot symmetrical (Fig 7).
Assuming a random-effects model, the new pooled prevalence estimate reduced to 49.26 per
cent (CI 37.15% to 61.46%).
<insert fig 7 here>
Additional analyses
Sensitivity analysis
Following the omission of Epstein-Lubow et al. (2008) the prevalence of burden
increased by a minimal percentage (1.4%). The analysis found no deviations from the main
analysis in terms of heterogeneity or significance of results.
Subgroup analysis
Random-effects sub group analysis comparing RCT data to non-RCT data was not
appropriate for burden outcomes, given that only one of the nine studies used baseline RCT
data (Epstein-Lubow et al. 2008). The depression pooled prevalence estimate of studies that
used baseline RCT data did not significantly differ to that of studies where samples were
obtained via cross-sectional or longitudinal prospective cohort designs (p = .734). The second
random-effects sub-group analysis included thirty-two studies and revealed that depression
prevalence estimates differed according to the type of measure used (p = .003); two studies
that used diagnostic criteria reported the lowest prevalence rate (8.9%, CI 3.4% to 21.4%, I2 =
88.01%), although one of these studies may be considered an outlier, followed by studies that
used a form of the Hospital Anxiety and Depression Scale (HADS; 26%, CI 15.6% to 40.1%,
I2 = 95.89%). Five studies that used a form of the Beck Depression Inventory (BDI) reported
the highest prevalence estimate (49.2%, CI 34.3% to 64.2%, I2 = 59.66%). As there were no
studies conducted in Africa and only one study based in South America reporting depression
prevalence data, the random-effects sub-group analysis for continent compared the pooled
prevalence estimates of Asia, Europe, Australia and North America. There was a significant
difference between the depression pooled prevalence estimates of the continents entered into
the analysis (p<.0007), with Asia reporting the lowest estimate of 26.8 per cent (CI 17.2% to
39.2%), followed by North America 29.1 per cent (CI 24.3% to 34.6%), Europe 36.8 per cent
(CI 31.1% to 42.8%) and Australia yielding the highest estimate of 58.1 per cent (CI 40.0%
to 74.3%).
Meta-regression results
Study quality was not a significant moderator of depression prevalence estimates
(0.0254, CI -0.0816 to 0.1324, p = .641) or burden prevalence estimates (-0.18, CI 0.144 to -
0.461, p = .215).
Odds-ratio meta-analyses
The first meta-analysis included eight studies (Fig 8) and revealed that the odds of a
female caregiver having depression was one point four five times higher than a male
caregiver (CI 1.125 to 1.874, p = .004). There was no significant heterogeneity of the odds
ratio estimates. The pooled odds ratio estimate corresponded to a z value of three point eight
five four (p = .001) indicating that 23 studies with a null effect would be needed to reduce the
p-value to below the significance level, suggesting that the odd ratios may not be robust.
However, the trim and fill method indicated no missing studies from the analysis. The second
meta-analysis included seven studies and the odds of a spouse compared to a non-spouse
having depression was found to be one point one five, however this was not significant (CI
0.737 to 1.779, I2 = 84.42, p = .547). The trim and fill method suggested there were no
missing studies from this analysis.
Discussion
Forty-three studies set across five of the seven continents, predominantly comprising
of cross-sectional designs, were examined with a combined total of 16 911 participants. To
our knowledge, this was the first meta-analysis to quantify the prevalence of perceived
burden among informal caregivers of PwD. Overall the trim-and-fill adjusted prevalence
estimate of burden was 49.26 per cent. In other words, approximately half of all the informal
caregivers of PwD perceive their caregiving role to be mildly to severely burdening. This
result was indicated to be robust in the context of publication bias. There may be numerous
reasons for why the remaining half of the population perceives their role to have little or no
burden, including that these caregivers perceive more positive benefits from the acquisition
of the role. For example, if a caregiver perceives that their family has become closer together,
this could impact upon their response to questions regarding the social impact of the role - a
construct of burden. Importantly, the finding highlights a great need within this population
for interventions effective at reducing burden. Such interventions could increase the
wellbeing of caregivers during their role, which could prolong the transition of care-
recipients to care homes, and prevent post-death psychiatric morbidity (Gaugler et al. 2005).
The trim-and-fill adjusted prevalence estimate of depression was 31.24 per cent,
suggesting that almost a third of all caregivers of PwD are experiencing depression.
Rosenthal’s fail safe N indicated that this finding was extremely robust, with over 8000 extra
studies with a null effect required to nullify the result. The depression prevalence estimate is
substantially higher than that of the prevalence of depression among adult primary care
patients, assessed via structured diagnostic interviews (Mitchell, Vaze and Rao 2009) and the
prevalence of depression in older adult populations, assessed via self-report measures (Li et
al. 2014; Luppa et al. 2012). Given that depression has been found to be a risk factor for
suicidal ideation among family caregivers of PwD, the high prevalence of depression
supports the finding of higher prevalence rates of suicidal ideation in this population
compared to the general population (O’Dwyer et al 2013; O’Dwyer et al. 2016). Overall, the
finding demonstrates that more informal caregivers of PwD are in need of interventions to
reduce depressive symptoms than the adult/older adult general population.
Interestingly the depression prevalence estimate is higher than that found in the study
of Cuijpers (2005). This could be attributed to the fact that all of the studies within Cuijpers
(2005) were conducted at least 12 years ago and therefore its estimate may not reflect the
current prevalence in today’s population. The difference could also be due to the fact that all
studies in Cuijpers (2005) were based in either the UK or the United States, unlike the current
review which included depression prevalence estimates from studies conducted in numerous
countries across Europe, multiple states in North America, and several places in Asia and
Australia. In addition to this, the current review included almost four times as many studies
and so may have provided a more accurate prevalence estimate. Finally, the review of
Cuijpers (2005) only included studies that assessed depression via semi-structured or
structured diagnostic interviews, whereas the current meta-analysis also included studies that
assessed depression via self-report measures. It has been reported that, compared with self-
report measures, interview methods commonly underestimate the prevalence of psychiatric
disorders (Mitchell et al. 2011). In line with this and the findings of other meta-analytic
reviews (e.g. Krebber et al. 2014), the current review discovered that the depression
prevalence estimates differed according to the instrument used to assess depression, with
interviews based on diagnostic criteria yielding the lowest pooled prevalence estimate. This
could also explain why the overall depression prevalence estimate was similar to that found
in Sallim et al. (2015), where studies assessing depression via self-report measures were
included.
The review also found that female caregivers are 1.45 times more likely to experience
depression than male caregivers. Although, this finding may not be robust in the context of
publication bias, and further observational studies comparing the prevalence of depression
between male and female caregivers of PwD are warranted. No significant difference in terms
of depression prevalence was observed between spousal and non-spousal caregivers;
indicating that caregivers who are adult children, friends or other relatives of the care-
recipient may be just as much at risk of developing depression as caregivers who are spouses
of the care-recipient. This outcome did not support the finding of Sallim et al. (2015), where
spousal caregivers of patients with AD were significantly more likely than non-spousal
caregivers of patients with AD to experience depression. It is not thought that this is
attributed to the fact that the current review included caregivers of people with all forms of
dementia, but because it included over twice as many studies - three of which reported a
higher prevalence of depression in non-spousal compared to spousal caregivers. Some
research has indicated that it may not be the type of relationship that poses a risk for
depression but the caregiver’s perception of the quality of the relationship. For example,
Kramer (1993), Williamson and Schulz (1993) and Fauth et al. (2012) found closer
relationships prior to the onset of dementia predicted lower levels of depressive symptoms.
Furthermore, Morris, Morris and Britton (1998) found caregivers with lower levels of
intimacy prior to and following the onset of dementia had higher levels of depressive
symptoms.
Limitations
Although study quality was not found to be a significant moderator of the burden or
depression prevalence estimates, 18 studies were rated as having a high risk of bias and only
three studies rated as having a low risk of bias. The majority of studies failed to report any
details of the history of psychiatric problems for the informal caregivers. Most did not report
details of the participation and response rates or when these were reported they were less than
75 per cent, and most studies did not compare those that did respond/participate to those that
did not (either qualitatively or quantitatively). This could mean that within these studies a
large proportion of caregivers did not respond/participate. If this were true, this could have
affected the accuracy of the burden prevalence estimate particularly given that one of the
reasons some informal caregivers of PwD do not engage with services is due to a high level
of burden (Brodaty, Thompson and Fine, 2005).
Another limitation of the review, and a major limitation of this field of research, is
that most studies used convenience based samples rather than population based samples.
Pruchno et al. (2008) discovered that caregivers recruited via convenience sampling methods
reported higher levels of burden and increased depressive symptomatology relative to those
identified using a population based sampling method. This is therefore a serious
methodological concern in that convenience samples are likely to exaggerate the prevalence
of depression and burden considerably and therefore the findings may not be reliably
generalizable (Pruchno et al. 2008). Future research should endeavour to recruit a
consecutive sample of the population.
Another limitation is the findings of significantly high heterogeneity of depression
and burden prevalence estimates. This suggests that these are not similar across studies and
conclusions drawn are limited by this fact. Interestingly, the purpose of recruitment did not
appear to impact the prevalence estimates as the pooled prevalence of studies that used
baseline RCT data did not significantly differ to that obtained for studies using cross-
sectional designs and longitudinal prospective cohort designs. The heterogeneity among
depression prevalence estimates was however partially explained by the type of instruments
used to measure depression, with studies using diagnostic criteria yielding the lowest pooled
prevalence estimate. In terms of self-report measures, studies that used a form of the HADS
yielded the lowest pooled prevalence estimate and studies using a form of the BDI had the
highest pooled prevalence estimate. These findings reflect those of a recent meta-analysis of
the prevalence of depression among medical outpatients (Wang et al. 2017). The self-report
measures are designed to assess clinically significant depressive symptoms but they are not
tools for diagnosing different types of mood disorders; for example, the HADS does not
include all of the diagnostic criteria for depression based on DSM (Laidlaw 2015). It is
therefore perhaps unsurprising that the two studies that used diagnostic criteria reported the
lowest prevalence rate. Moreover, the HADS was designed to detect depression and anxiety
in people with medical conditions, and thus it is useful for older people with chronic physical
illnesses. Although the BDI is a well-established measure, it can be criticised for having
somatic scale items as this may inflate scores when used with older people (Laidlaw 2015).
Considering that many informal caregivers of PwD are older people, this may account for the
significantly large difference observed between the pooled prevalence estimates of studies
that used the HADS and the BDI. It is also acknowledged that different cut-offs may have
affected the diagnostic sensitivity and specificity.
The study also revealed that prevalence estimates differed by continent. Asia appeared
to have the lowest prevalence of depression, followed by North America, Europe and
Australia, respectively. Unfortunately, the review could not include South America within the
sub-group analysis as only one study conducted in this region reported the prevalence of
depression, and overall no included study was conducted in Africa. This leaves a question as
to whether the prevalence of depression among informal caregivers of PwD differs greatly in
these continents.
Conclusion
In summary, this review revealed that almost one third of informal caregivers of PwD
experience depression and approximately one half appraise their caregiving role to be
burdensome. Unfortunately, significant heterogeneity of depression and burden prevalence
estimates was observed. As reported in other reviews, different screening instruments were
found to produce different estimates of depression. The heterogeneity of depression
prevalence estimates was also partially explained by the continent the studies were conducted
in, with Asia reporting the lowest prevalence and Australia the highest. Female caregivers
were found to be more at risk of experiencing depression than male caregivers. However,
further observational studies investigating this finding are warranted. No significant
difference in terms of depression prevalence was observed between spousal and non-spousal
caregivers. Based on previous literature, it is suggested that a caregiver’s vulnerability to
developing depression may be more related to the quality of the relationship with the care-
recipient as opposed to the relationship type. The review demonstrates that within this
population there is a great need for the provision of interventions that are effective at
reducing burden and depressive symptoms. Given that these difficulties can negatively
impact upon a caregiver’s health, ability to perform their role (Gallagher et al. 1989; Cucciare
et al. 2010), and increase the likelihood of the care-recipient being transitioned to a nursing
home placement (Gaugler et al. 2005), economically, it would appear vital for dementia
services to promptly establish or tailor existing interventions to treat these difficulties.

Role of funding sources: This work was supported by the Clinical Psychology Course at the
University of East Anglia, who provided financial support for the CMA software. The
authors work within the institution, however no other UEA personal were involved in the
study design, collection, analysis or interpretation of the data, writing the manuscript, or the
decision to submit the paper for publication.
Contributors: The first author designed the study, collected the data, analysed the title,
abstracts and all full texts, conducted statistical analyses, interpreted the data and wrote the
manuscript. The second author contributed to analysing full text articles and assessing
eligibility, extracting the data and analysing the quality of selected studies, and editing and
proof reading the manuscript. Both authors have approved the final manuscript.
Conflicts of interest: All authors disclose no conflicts of interest.

References
Adams B, Aranda MP, Kemp B. and Takagi K (2002) Ethnic and gender differences in
distress among Anglo American, African American, Japanese American, and Mexican
American spousal caregivers of persons with dementia. Journal of Clinical
Geropsychology, 8, 279-301.
Alzheimer’s Disease International (2013) World Alzheimer’s Report 2013: An analysis of
long-term care for dementia. Retrieved from,
https://www.alz.co.uk/research/WorldAlzheimerReport2013.pdf
Alzhiemer’s Disease International (2015) World Alzheimer’s Report 2015. The global
impact of dementia: The Global Impact of Dementia: An analysis of prevalence,
incidence, cost and trends. Retrieved from, https://www.alz.co.uk/research/world-
report-2015
Alzheimer’s Research UK (2015) Women and dementia: a marginalised majority. Retrieved
from, https://www.alzheimersresearchuk.org/wp-content/uploads/2015/03/Women
-and-Dementia-A-Marginalised-Majority1.pdf
Alzheimer’s Society (2014) Care for today and cure for tomorrow: Alzhimer’s Society
strategy for funding research 2014-2017. Retrieved from,
https://www.alzheimers.org.uk/download/downloads/id/2254/care_for_today_and
_cure_for_tomorrow.pdf
American Psychiatric Association (1980) Diagnostic and statistical manual of mental
disorders (3rd ed.). Washington, DC: American Psychiatric Publishing.
disorders (3rd ed. Text Revision.). Washington, DC: American Psychiatric Publishing.
disorders (4th ed.). Washington, DC: American Psychiatric Publishing.
disorders (4th ed. Text revision.). Washington, DC: American Psychiatric Publishing
disorders (5th ed.). Washington, DC: American Psychiatric Publishing.
Andresen EM, Malmgren JA, Carter WB and Patrick DL (1994) Screening for depression in
well older adults: Evaluation of a short form of the CES-D (Center for Epidemiologic
Studies Depression Scale). American Journal of Preventive Medicine, 10, 77–84.
Arango Lasprilla JC, Moreno A, Rogers H and Francis K (2009) The effect of dementia
patient’s physical, cognitive, and emotional/behavioral problems on caregiver well-
being: findings from a Spanish-speaking sample from Colombia, South
America. American Journal of Alzheimer's Disease & Other Dementias®, 24, 384-
395
Beck AT and Beck RW (1972) Screening depressed patients in family practice: a rapid
technique. Postgraduate Medicine, 52, 81-85.
Beck AT, Steer,RA, Ball R and Ranieri W (1996) Comparison of Beck Depression
Inventories – IA and –II in psychiatric outpatients. Journal of Personality
Assessment, 67, 588-597.
Beck AT, Ward CH, Mendelson M, Mock J and Erbaugh J (1961) An inventory for
measuring depression. Archives of General Psychiatry, 4, 561-571.
Bednarek A, Mojs E, Krawczyk-Wasielewska A, Głodowska K, Samborski W, Lisiński P,
Kopczyński P, Gregersen R and Millán-Calenti JC (2016) Correlation between

depression and burden observed in informal caregivers of people suffering from
dementia with time spent on caregiving and dementia severity. Eur Rev Med
Pharmacological Science, 20, 59-63.
Bejjani C, Snow AL, Judge KS, Bass DM, Morgan RO, Wilson N, Walder A, Looman WJ,
McCarthy C and Kunik ME (2015) Characteristics of depressed caregivers of veterans
with dementia. American Journal of Alzheimer's Disease & Other Dementias, 30,
672-678.
Berger G, Bernhardt T, Weimer E, Peters J, Kratzsch T and Frolich L (2005)
Longitudinal study on the relationship between symptomatology of dementia and
levels of subjective burden and depression among family caregivers in memory clinic
patients. Journal of Geriatric Psychiatry and Neurology, 18, 119-128.
Borenstein M, Hedges L, Higgins, J and Rothstein H (2005) Comprehensive meta-
analysis (version 2). Englewood, NJ: Biostat.
Borenstein M, Hedges LV, Higgins J and Rothstein HR (2009) Introduction to meta -
analysis. Chichester, West Sussex: John Wiley & Sons, Ltd.
Borsje P, Hems MA, Lucassen PL, Bor H, Koopmans RT and Pot AM (2016)
Psychological distress in informal caregivers of patients with dementia in primary
care: course and determinants. Family Practice, 33, 374-381.
Breinbauer H, Vásquez H, Mayanz S, Guerra C and Millán T (2009) Validación en
Chile de la Escala de Sobrecarga del Cuidador de Zarit en sus versiones original y
abreviada. Revista Médica de Chile, 137, 657-665.
Brodaty H and Donkin M (2009) Family caregivers of people with dementia. Dialogues in
Clinical Neuroscience, 11, 217.

Brodaty H, Thomson C, Thompson C and Fine M (2005) Why caregivers of people with
dementia and memory loss don't use services. International Journal of Geriatric
Psychiatry, 20, 537-546.
Brodaty H, Woodward M, Boundy K, Ames D, Balshaw R and PRIME Study Group
(2014) Prevalence and predictors of burden in caregivers of people with
dementia. The American Journal of Geriatric Psychiatry, 22, 756-765.
Brooker D (2017) How has the field of dementia care changed in the past 30 years?
Retrieved from, http://www.jkp.com/jkpblog/2017/05/dementia-field-changes-30-
years.
Caspar S and O’Rourke N (2009) The composition and structure of depressive
symptomatology among young and older caregivers of persons with dementia. Ageing
International, 34, 33-41.
Chan YF, Leung DY, Fong DY, Leung CM and Lee AM (2010) Psychometric evaluation of
the Hospital Anxiety and Depression Scale in a large community sample of
adolescents in Hong Kong. Qual Life Res, 19, 865-873.
Collins RN, Gilligan LJ and Poz R (2018) The Evaluation of a Compassion-Focused
Therapy Group for Couples Experiencing a Dementia Diagnosis. Clinical
Gerontologist, 41, 474-486.
Conde C and Useros E (1975) Adaptación castellana de la escala de evaluación conductual
para la depresión de Beck. Revista de Psiquiatría y Psicología Médica de Europa y
América Latinas, 12, 217–236.
Cheng ST, Lam LC and Kwok T (2013) Neuropsychiatric symptom clusters of Alzheimer
disease in Hong Kong Chinese: correlates with caregiver burden and depression. The
American Journal of Geriatric Psychiatry, 21, 1029-1037.

Chiao CY, Wu HS and Hsiao CY (2015) Caregiver burden for informal caregivers of
patients with dementia: a systematic review. International Nursing Review, 62, 340
-350.
Cohen J (1960) A coefficient of agreement for nominal scales. Educational and
Psychological Measurement, 20, 37-46.
Cohen CA, Colantonio A and Vernich L (2002) Positive aspects of caregiving: rounding out
the caregiver experience. International Journal of Geriatric Psychiatry, 17, 184-188.
Contador I, Fernández-Calvo B, Palenzuela DL, Miguéis S and Ramos F (2012)
Prediction of burden in family caregivers of patients with dementia: A perspective of
optimism based on generalized expectancies of control. Aging & Mental Health, 16,
675-682.
Copas J and Shi JQ (2000) Meta-analysis, funnel plots and sensitivity
analysis. Biostatistics, 1, 247-262.
Covinsky KE, Newcomer R, Fox P, Wood J, Sands L, Dane K and Yaffe K (2003)
Patient and caregiver characteristics associated with depression in caregivers of
patients with dementia. Journal of General Internal Medicine, 18, 1006-1014.
Cucciare MA, Gray H, Azar A, Jimenez D and Gallagher-Thompson D (2010) Exploring the
relationship between physical health, depressive symptoms, and depression diagnoses
in Hispanic dementia caregivers. Aging & Mental Health, 14, 274-282.
Cuijpers P (2005) Depressive disorders in caregivers of dementia patients: a systematic
review. Aging & Mental Health, 9, 325-330.
Duval S and Tweedie R (2000a) Trim and fill: a simple funnel‐plot–based method of
testing and adjusting for publication bias in meta‐analysis. Biometrics, 56, 455-
463.
Dilworth-Anderson P and Anderson N (1994) Dementia caregiving in Blacks: A contextual
approach to research. In E Light, G Niederehe and B Lebowitz (Eds.) Stress effects
on family caregivers of Alzheimer’s patients (pp. 385–409). New York: Springer.
Epstein-Lubow G, Davis JD, Miller IW and Tremont G (2008) Persisting burden predicts
depressive symptoms in dementia caregivers. Journal of Geriatric Psychiatry and
Neurology, 21, 198-203.
Etters L, Goodall D and Harrison BE (2008) Caregiver burden among dementia patient
caregivers: a review of the literature. Journal of the American Association of Nurse
Practitioners, 20, 423-428.
Fauth E, Hess K, Piercy K, Norton M, Corcoran C, Rabins P, Lyketsos C and Tschanz J
(2012) Caregivers’ relationship closeness with the person with dementia predicts both
positive and negative outcomes for caregivers’ physical health and psychological
well-being. Aging & Mental Health, 16, 699-711.
First MB, Spitzer RL, Gibbon M and Williams JBW (2008) Structured clinical interview for
DSM-IV axis I disorders-clinician version (SCID-I). In AJ Rush, MB First and D
Blacker (Eds.) Handbook of psychiatric measures (2nd ed., pp. 40–43). Washington,
DC: American Psychiatric Publishing.
Folstein MF, Folstein SE and McHugh PR (1975) “Mini-mental state”: a practical
method for grading the cognitive state of patients for the clinician. Journal of
Psychiatric Research, 12, 189-198.
Gallagher D, Ni Mhaolain A, Crosby L, Ryan D, Lacey L, Coen RF, Walsh C, Coakley D,
Walsh JB, Cunningham C and Lawlor BA (2011) Self-efficacy for managing
dementia may protect against burden and depression in Alzheimer's caregivers. Aging
& Mental Health, 15, 663-670

Gallagher D, Rose J, Rivera P, Lovett S and Thompson LW (1989) Prevalence of depression
in family caregivers. The Gerontologist, 29, 449-456.
García-Alberca JM, Cruz B, Lara JP, Garrido V, Gris E, Lara A and Castilla C (2012)
Disengagement coping partially mediates the relationship between caregiver burden
and anxiety and depression in caregivers of people with Alzheimer's disease. Results
from the MÁLAGA-AD study. Journal of Affective Disorders, 136, 848-856.
Gaugler JE, Kane RL, Kane RA and Newcomer R (2005) Unmet care needs and key
outcomes in dementia. Journal of the American Geriatrics Society, 53, 2098-
2105.
Germain S, Adam S, Olivier C, Cash H, Ousset PJ, Andrieu S, Vellas B, Meulemans T,
Reynish E and Salmon E (2009) Does cognitive impairment influence burden in
caregivers of patients with Alzheimer's disease? Journal of Alzheimer's Disease, 17,
105-114.
Given CW, Given B, Azzouz F, Kozachik S and Stommel M (2001) Predictors of pain
and fatigue in the year following diagnosis among elderly cancer patients. Journal of
Pain Symptom Management, 21, 456-66.
Givens JL, Mezzacappa C, Heeren T, Yaffe K and Fredman L (2014) Depressive symptoms
among dementia caregivers: Role of mediating factors. The American Journal of
Geriatric Psychiatry, 22, 481-488.
Goldberg D, Bridges K, Duncan-Jones P and Grayson D (1988) Detecting anxiety and
depression in general medical settings. British Medical Journal, 97, 897–899.
Hamilton M (1980) Rating depressive patients. Journal of Clinical Psychiatry. 41, 21-24.
Harbord R (2010) Investigating heterogeneity: sub-group analysis and meta-regression
[PowerPoint slides]. Retrieved from

http://methods.cochrane.org/sites/methods.cochrane.org.statistics/files/public/upl
oads/
SMG_training_course_cardiff/2010_SMG_training_cardiff_day1_session4_harbord.p
df
Hasegawa N, Hashimoto M, Koyama A, Ishikawa T, Yatabe Y, Honda K, Yuuki S.,
Araki K and Ikeda M (2014). Patient-related factors associated with depressive
state in caregivers of patients with dementia at home. Journal of the American
Medical Directors Association, 15, 371.e15-317.e18.
Hays JC, Blazer DG and Gold DT (1993) CES‐D: Cutpoint or Change Score? Journal of the
American Geriatrics Society, 41, 344-345.
Higgins JPT and Thompson SG (2002) Quantifying heterogeneity in a meta-
analysis. Statistics in Medicine, 21, 1539-1558.
Higgins JPT, Thompson SG, Deeks JJ and Altman DG (2003) Measuring inconsistency in
meta-analyses. BMJ: British Medical Journal, 327, 557.
Holland JM, Thompson LW, Tzuang M and Gallagher-Thompson D (2010)
Psychosocial factors among Chinese American women dementia caregivers and their
association with salivary cortisol: Results of an exploratory study. Ageing
International, 35, 109-127.
Hoppes S, Bryce H, Hellman C and Finlay E (2012) The effects of brief mindfulness training
on caregivers' well-being. Activities, Adaptation & Aging, 36, 147-166.
Hughes CP, Berg L, Danziger WL, Coben LA and Martin R (1982) A new clinical
scale for the staging of dementia. The British Journal of Psychiatry, 140, 566-572.
Jang Y, Clay OJ, Roth DL, Haley WE and Mittelman MS (2004) Neuroticism and
longitudinal change in caregiver depression: Impact of a spouse-caregiver
intervention program. The Gerontologist, 44, 311-317.
Joling KJ, O'Dwyer ST, Hertogh CM and Hout HP (2018) The occurrence and persistence of
thoughts of suicide, self‐harm and death in family caregivers of people with dementia:
a longitudinal data analysis over 2 years. International Journal of Geriatric
Jorm AF (2004) The Informant Questionnaire on cognitive decline in the elderly
(IQCODE): a review. International Psychogeriatrics, 16, 275-293.
Kaiser S and Panegyres PK (2007) The psychosocial impact of young onset dementia on
spouses. American Journal of Alzheimer's Disease & Other Dementias®, 21, 398-
402.
Kemp BJ and Adams B (1995) The Older Adult Health and Mood Questionnaire: A
measure of geriatric depressive disorder. Journal of Geriatric Psychiatry and
Neurology. 8, 162-167.
Kramer BJ (1993) Marital history and the prior relationship as predictors of positive and
negative outcomes among wife caregivers. Family Relations, 42, 367–375
Krebber AMH, Buffart LM, Kleijn G, Riepma IC, Bree R, Leemans CR, Becker A, Brug J,
Straten A, Cuijpers P and Verdonck‐de Leeuw IM (2014) Prevalence of depression in
cancer patients: a meta‐analysis of diagnostic interviews and self report
instruments. Psycho‐Oncology, 23, 121-130.
Kroenke K, Spitzer RL and Williams JB (2001) The PHQ9: validity of a brief depression
severity measure. Journal of General Internal Medicine, 16, 606-613.
Kurz X, Scuvée-Moreau J, Vernooij-Dassen M and Dresse A (2003) Cognitive
impairment, dementia and quality of life. Acta Neurologica Belgica, 103, 24-34.
Laidlaw K (2015) Cognitive behaviour therapy for older people: An introduction. London:
SAGE Publications.
Li D, Zhang DJ, Shao JJ, Qi XD and Tian L (2014) A meta-analysis of the prevalence of
depressive symptoms in Chinese older adults. Archives of Gerontology and
Geriatrics, 58, 1-9.
Liang X, Guo Q, Luo J, Li F, Ding D, Zhao Q and Hong Z (2016) Anxiety and depression
symptoms among caregivers of care-recipients with subjective cognitive decline and
cognitive impairment. BMC Neurology, 16, 191.
Lowery K, Mynt P, Aisbett J, Dixon T, O’Brien J and Ballard C (2000) Depression in the
carers of dementia sufferers: a comparison of the carers of patients suffering from
dementia with Lewy bodies and the carers of patients with Alzheimer’s
disease. Journal of Affective Disorders, 59, 61-65.
Lu YYF and Austrom MG (2005) Distress responses and self-care behaviors in dementia
family caregivers with high and low depressed mood. Journal of the American
Psychiatric Nurses Association, 11, 231-240.
Luchsinger JA, Tipiani D, Torres-Patiño G, Silver S, Eimicke JP, Ramirez M, Teresi J and
Mittelman M (2015) Characteristics and mental health of hispanic dementia
caregivers in New York City. American Journal of Alzheimer's Disease & Other
Dementias®, 30, 584-590.
Luppa M, Sikorski C, Luck T, Ehreke L, Konnopka A, Wiese B, Weyerer S, König, HH and
Riedel-Heller SG (2012) Age-and gender-specific prevalence of depression in latest-
life–systematic review and meta-analysis. Journal of Affective Disorders, 136, 212-
221.
Mahoney R, Regan C, Katona C and Livingston G (2005) Anxiety and depression in
family caregivers of people with Alzheimer disease: the LASER-AD study. The
American Journal of Geriatric Psychiatry, 13, 795-801.
Martín M, Salvadór I, Nadal S, Miji L, Rico J, Lanz P and Taussig M (1996) Adaptación para
nuestro medio de la Escala de Sobrecarga del Cuidador (Caregiver Burden Interview)
de Zarit. Revista Gerontolo´gica, 6, 338–346.
Medrano M, Rosario RL, Payano AN and Capellán NR (2014) Burden, anxiety and
depression in caregivers of Alzheimer patients in the Dominican Republic. Dementia
& Neuropsychologia, 8, 384-388.
McConaghy R and Caltabiano ML (2005) Caring for a person with dementia: Exploring
relationships between perceived burden, depression, coping and well‐being. Nursing
& Health Sciences, 7, 81-91.
McKhann G, Drachman D, Folstein M, Katzman R, Price D and Stadlan EM (1984) Clinical
diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under
the auspices of Department of Health and Human Services Task Force on Alzheimer's
Disease. Neurology, 34, 939–44
Medrano M, Rosario RL, Payano AN and Capellán NR (2014) Burden, anxiety and
depression in caregivers of Alzheimer patients in the Dominican Republic. Dementia
& Neuropsychologia, 8, 384-388.
Mitchell AJ, Chan M, Bhatti H, Halton M, Grassi L, Johansen C and Meader N (2011)
Prevalence of depression, anxiety, and adjustment disorder in oncological,
haematological, and palliative-care settings: a meta-analysis of 94 interview-based
studies. The Lancet Oncology, 12, 160-174.

Mitchell AJ, Vaze A and Rao S (2009) Clinical diagnosis of depression in primary care: a
meta-analysis. The Lancet, 374, 609-619.
Moher D, Liberati A, Tetzlaff J, Altman DG and Prisma Group (2009) Preferred reporting
items for systematic reviews and meta-analyses: the PRISMA statement. PLoS
Medicine, 6, e1000097.
Montgomery SA and Asberg M (1979) A new depression scale designed to be sensitive to
change. Br. J. Psychiatry 134, 382-389.
Morris RG, Morris LW and Britton PG (1998) Factors affecting the emotional well-being of
caregivers of dementia sufferers. British Journal of Psychiatry, 153, 147-156.
Munn Z, Moola S, Lisy K and Riitano D (2014) The Joanna Briggs institute reviewer’s
manual 2014: The systematic review of prevalence and incidence data. South
Australia: The Joanna Briggs Institute.
Novak M. and Guest C (1989) Application of a multidimensional caregiver burden
inventory. The Gerontologist, 29, 798-803.
O’Brien E (2016) Alzheimer's diagnosis can mean steep costs for family caregivers.
Retrieved from, http://time.com/money/4501444/alzheimers-family-caregivers
-steep-costs/
O'Dwyer ST, Moyle W, Zimmer‐Gembeck M and De Leo D (2013) Suicidal ideation in
family carers of people with dementia: a pilot study. International Journal of
Geriatric Psychiatry, 28, 1182-1188.
O'Dwyer ST, Moyle W, Zimmer-Gembeck M and De Leo D (2016) Suicidal ideation in
family carers of people with dementia. Aging & Mental Health, 20, 222-230.
Orgeta V and Lo Sterzo E (2013) Sense of coherence, burden, and affective symptoms in
family carers of people with dementia. International Psychogeriatrics, 25, 973-
980.
Ornstein K and Gaugler JE (2012) The problem with “problem behaviors”: a systematic
review of the association between individual patient behavioral and psychological
symptoms and caregiver depression and burden within the dementia patient–caregiver
dyad. International Psychogeriatrics, 24, 1536-1552.
Ory MG, Hoffman III RR, Yee JL, Tennstedt S and Schulz R (1999) Prevalence and impact
of caregiving: A detailed comparison between dementia and non dementia
caregivers. The Gerontologist, 39, 177-186.
Ostojić D, Vidović D, Baceković A, Brecić P and Jukić V (2014) Prevalence of anxiety and
depression in caregivers of Alzheimer's dementia patients. Acta Clin Croat, 53, 17-21.
Pinquart M and Sörensen S (2003) Differences between caregivers and non caregivers in
psychological health and physical health: a meta-analysis. Psychology and Aging, 18,
250-267.
Piercy KW, Fauth EB, Norton MC, Pfister R, Corcoran CD, Rabins PV, Lyketsos C and
Tschanz JT (2012) Predictors of dementia caregiver depressive symptoms in a
population: The Cache County Dementia Progression Study. Journals of Gerontology
Series B: Psychological Sciences and Social Sciences, 68, 921-926.
Pruchno RA, Brill JE, Shands Y, Gordon J R, Genderson MW, Rose M and Cartwright F
(2008) Convenience samples and caregiving research: how generalizable are the
findings? The Gerontologist, 48, 820-827.
Radloff LS (1977) The CES-D scale: A self-report depression scale for research in the
general population. Applied Psychological Measurement, 1, 385-401.
Raggi A, Tasca D, Panerai S, Neri W and Ferri R (2015) The burden of distress and
related coping processes in family caregivers of patients with Alzheimer's disease
living in the community. Journal of the Neurological Sciences, 358, 77-81.

Ramos-Brieva JC (1986) Validacion de la versión castellana de la escala de Hamilton para
la depresión. Actas Luso-Esp Neurol Psquiatr, 6, 324-334.
Reisberg B, Ferris SH, de Leon MJ and Crook T (1982) The Global Deterioration Scale for
assessment of primary degenerative dementia. American Journal of Psychiatry, 139,
1136- 1139.
Riedel O, Klotsche J and Wittchen HU (2016) Overlooking informal dementia caregivers'
burden. Research in Gerontological Nursing, 9, 167-174.
Roche L, Croot K, MacCann C, Cramer B and Diehl-Schmid J (2015) The role of coping
strategies in psychological outcomes for frontotemporal dementia
caregivers. Journal of Geriatric Psychiatry and Neurology, 28, 218-228.
Rosenthal R (1979) The file drawer problem and tolerance for null results. Psychological
Bulletin, 86, 638-341.
Rosness TA, Mjørud M and Engedal K (2011) Quality of life and depression in carers of
patients with early onset dementia. Aging & Mental Health, 15, 299-306.
Roth DL, Ackerman ML, Okonkwo OC and Burgio LD (2008) The four-factor model of
depressive symptoms in dementia caregivers: A structural equation model of ethnic
differences. Psychology and Aging, 23, 567-576.
Roth M, Tym E, Mountjoy CQ, Huppert FA, Hendrie H, Verma S, and Goddard R (1986)
CAMDEX. A standardised instrument for the diagnosis of mental disorder in the
elderly with special reference to the early detection of dementia. The British Journal
of Psychiatry, 149, 698-709.

Sallim AB, Sayampanathan AA, Cuttilan A and Ho RCM (2015) Prevalence of mental health
disorders among caregivers of patients with Alzheimer disease. Journal of the
American Medical Directors Association, 16, 1034-1041.
Sansoni J, Vellone E and Piras G (2004) Anxiety and depression in community‐dwelling,
Italian Alzheimer's disease caregivers. International Journal of Nursing
Practice, 10, 93-100.
Schulz R and Martire LM (2004) Family caregiving of persons with dementia: prevalence,
health effects, and support strategies. The American Journal of Geriatric
Simpson CE (2010) Sleep: effect on dementia caregiver mastery, perceived stress and
depression (unpublished doctoral dissertation). The University of Texas at Austin,
Austin, TX.
Slachevsky A, Budinich M, Miranda-Castillo C, Nunez-Huasaf J, Silva JR, Munoz-Neira C,
Gloger S, Jimenez O, Martorell B and Delgado C (2013) The CUIDEME Study:
determinants of burden in chilean primary caregivers of patients with
dementia. Journal of Alzheimer's Disease, 35, 297-306.
Sleath B, Thorpe J, Landerman LR, Doyle M and Clipp E (2005). African‐American and
White Caregivers of Older Adults with Dementia: Differences in Depressive
Symptomatology and Psychotropic Drug Use. Journal of the American Geriatrics
Society, 53, 397-404.
Spector A, Thorgrimsen L, Woods BOB, Royan L, Davies S, Butterworth M and Orrell M
(2003) Efficacy of an evidence-based cognitive stimulation therapy programme for
people with dementia. The British Journal of Psychiatry, 183, 248-254.

Spitzer RL and Robins E (1978) Research diagnostic criteria: rationale and reliability.
Archives of General Psychiatry, 35, 773-782
Taylor Nelson Sofres British Market Research Bureau Limited (2014) Attitudes to mental
illness: 2013 research report. Retrieved from, https://www.mind.org.uk/news-
campaigns/news/survey-shows-greatest-improvement-in-public-attitudes-to-mental-
health-in-20-years/#.WmUMqKhl-00
Välimäki TH, Martikainen JA, Hallikainen IT, Väätäinen ST and Koivisto AM (2015)
Depressed spousal caregivers have psychological stress unrelated to the progression
of Alzheimer disease: a 3-year follow-up report, Kuopio ALSOVA study. Journal of
Geriatric Psychiatry and Neurology, 28, 272-280.
Vitaliano PP, Young HM and Russo J (1991) Burden: A review of measures used among
caregivers of individuals with dementia. The Gerontologist, 31, 67-75.
Vitaliano PP (1997) Physiological and physical concomitants of caregiving: Introduction to
special issue. Annals of Behavioral Medicine, 19, 75-77.
Waite A, Bebbington P, Skelton‐Robinson M and Orrell M (2004) Social factors and
depression in carers of people with dementia. International Journal of Geriatric
Wang J, Wu X, Lai W, Long E, Zhang X, Li W, Zhu Y, Chen C, Zhong X, Liu Z and Wang
D (2017) Prevalence of depression and depressive symptoms among outpatients: a
systematic review and meta-analysis. BMJ Open, 7, 017173.
Williams IC (2005) Emotional health of black and white dementia caregivers: A contextual
examination. The Journals of Gerontology Series B: Psychological Sciences and
Social Sciences, 60, 287-295.

Williamson GM and Schulz R (1993) Coping with specific stressors in Alzheimer's disease
caregiving. The Gerontologist, 33, 747-755.
Wittchen HU, Höfler M and Meister W (2001) Prevalence and recognition of depressive
syndromes in German primary care settings: Poorly recognized and treated?
International Clinical Psychopharmacology, 16, 121-135.
World Health Organization (1978) The ICD-9 classification of mental and behavioural
disorders: Clinical descriptions and diagnostic guidelines. Geneva: World Health
Organization.
World Health Organization (1992) The ICD-10 classification of mental and behavioural
disorders: Clinical descriptions and diagnostic guidelines. Geneva: World Health
Organization.
World Health Organization (2003) International Classification of Primary Care, Second
edition (ICPC-2). Retrieved from
http://www.who.int/classifications/icd/adaptations/icpc2/en/index.
World Health Organization (2017) Depression and other common mental disorders: global
health estimates. Geneva, Switzerland: WHO Document Production Services.
Yesavage JA and Sheikh JI (1986) Geriatric depression scale (GDS) recent evidence and
development of a shorter version. Clinical Gerontologist, 5, 165-173.
Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M and Leirer VO (1983)
Development and validation of a geriatric depression screening scale: a preliminary
report. Journal of Psychiatric Research, 17, 37-49.

Zarit SH, Reever KE and Back-Peterson J (1980) Relatives of the impaired elderly: correlates
of feelings of burden. The Gerontologist, 20, 649-655.
Zarit SH, Todd PA and Zarit JM (1986) Subjective burden of husbands and wives as
caregivers: A longitudinal study. The Gerontologist, 26, 260-266.
Zigmond AS and Snaith RP (1983) The hospital anxiety and depression scale. Acta
Psychiatrica Scandinavica, 67, 361-370.

Sociocultural
context
Education
Gender
Employment
Income
Interpersonal context Marital status Personal context
Formal support Social activities

Social network Religiosity and spirituality
Received support Emotional health Financial difficulty
Stressfulness appraisal
Depressive symptoms
Positive aspects of caregiving
Situational context (e.g. ≥16 CES-D)
No. of health problems
Cognitive status of CR
CG helps with ADLs Temporal context
CG helps with IADLs
Memory and problem Age
behaviours Relationship to CR
Vigilance
Fig 1. The conceptual model for understanding the effects of context on emotional health outcomes
among caregivers of people with dementia, adapted from the model of Dilworth-Anderson and
Anderson (1994). CES-D = Center for Epidemiological Studies–Depression Scale; CR = care recipient;
CG = caregiver; ADLs = activities of daily living; IADLs = instrumental ADLs.
For Review only Page 46 of 78
Table 1. Search strategy and key terms
Concepts Search terms
Epidemiology1 ‘epidemiologic’ OR ‘epidemiological’ OR ‘epidemiol*’ OR ‘prev*’
OR ‘inciden*’
Burden/depression2 ‘depress*’ OR ‘depression emotion’ OR ‘distress’ OR ‘depressive
disorder’ OR ‘major depression’ OR ‘burden’
Type of participants3 ‘Dementia’ OR ‘Alzheimer’s’ OR ‘cognitively impaired’ OR
‘caregiver’ OR ‘carer’ OR ‘care’ OR ‘caring’ OR ‘caregiving’ OR
‘family caregiver’ OR ‘family carer’ OR ‘informal caregiver’ OR
‘informal carer’
Combined 1 2 AND 3
Note: For the databases PsycARTICLES, PsycINFO and MEDLINE Complete the key words
in the ‘epidemiology concept’ were searched for in the abstracts of texts and the
‘burden/depression’ and ‘participants’ concepts in the title of texts. The SCOPUS search was
limited to articles, reviews and conference papers, and all key words were searched for in the
titles and abstracts of articles. The key words were searched for in the titles of texts within the
Web of Science database and abstracts of texts within the Proquest database.
Page 47 of 78 For Review only
Table 2. 13-item adapted bias risk assessment tool
Bias Risk Item Positive Score Negative Score

Study: (1): Adequate (0): incomplete,
information is unclear or lack
provided of description
A. Patient population
1. Socio-demographic descriptions are described
including age, gender, marital status or
educational/employment/socioeconomic
status
2. Inclusion and exclusion criteria are
formulated
3. Type and percentage of the different forms of
dementia
4. Diagnostic procedure used to identify
probable dementia
5. Range or mean and SD of dementia severity
in sample is described
6. Mean or median and range or standard
deviation of time as carer given
7. (History of) psychiatric problems are detailed
8. Percentages of the types of relationship to
person with dementia is given
B. Sample recruitment
1. Sample size ≥ 100
2. Participation and response rates are described
and are more than 75%
3. Reasons for non-response or non-
participation are described or there is a
comparison between responders and non-
†dThe
Adequate
period
and
(setting
Adequateresponders
escription
geographical
Adequate
of recruitment,
of sampling
location)
and
frame,
placerecruitment
of recruitment
methods,
4. Description of recruitment method, period of
recruitment and place of recruitment (setting
and geographical location
5. Consecutive sample (sought to include all
accessible subjects)
Total: 5 (high risk)
Divide total by 13:
Risk: low risk, medium risk or high risk
Note: low risk = ≥ 9.75 medium risk = ≥ 6.5 – 9.75 high risk = < 6.5
Table 3: Characteristics of included studies (N = 43)

Study Location Design Recruitment Dementia diagnostic Forms of dementia n x̅ %♀
source procedure age
Adams et al. (2002) USA; CA Cross- Senior centre, N/A N/A 202 74.9 67.3%
and Hawaii sectional rehab centre and
agencies of the
USC
Arango et al. (2009) South Cross- A memory clinic N/A N/A 73 57.7 82.2%
America; sectional
Columbia
Bednarek et al. (2016) Poland; Baseline A project aimed to Medically N/A 41 61.7 73.2%
Greater RCT data understand and diagnosed; unknown
Poland support caregivers procedure
of PwD
Bejjani et al. (2016) USA; MA, Baseline Veterans’ admin N/A N/A 486 68.4 94%
TX, RI and RCT data health care system
OK
Berger et al. (2005) Germany; Longitudinal A memory clinic Neurological and AD 72% 45 60.7 62%
Frankfurt prospective neuropsy-chological VD 9%
cohort assessment in line FTD 9%
with ICD-10 Mixed dementia
4%
Unknown 4%
LBD 2%
Borsje et al. (2016) Netherlands Longitudinal General ICPC-2 N/A 117 67.3 68.4%
; Southern prospective Practitioner
regions cohort surgeries
Brodaty et al. (2014) Australia; Longitudinal Three memory DSM-IV criteria for AD 71.2% 524 N/A 64.9%
multiple prospective clinics dementia VD 7%
locations cohort FTD 4.2%
Mixed dementia
17.6%
Caspar and O’Rourke Canada; all Cross- Government MMSE, N/A 1426 N/A N/A
(2009) provinces sectional health records neurological and
except neuropsyc-hological
Ontario assessment
Cheng et al. (2013) China, Cross- Memory clinics, NINCDS-ADRDA AD 100% 142 58.9 73%
Hong Kong sectional outpatient clinics, criteria for AD
day hospitals, day
care centres and
social services
Contador et al. (2012) Spain; Cross- Referrals to the DSM-IV-R for AD 40.8% 130 58.6 72%
Salamanca sectional Association of dementia VD 28.4%
Family Members Mixed dementia
of Patients with 30.8%
Alzheimer's
Covinsky et al. (2003) USA; MN, Baseline Physician referrals N/A N/A 5627 64 71.7%
FL, OR, RCT data and self-referrals
NY, TN,
OH and IL
Cucciare et al (2010) USA; CA Baseline Health and social Physician diagnosis N/A 89 51.8 100%
RCT data services or MMSE below 23
professionals,
media, and word
of mouth
Epstein-Lubow et al. USA, New Baseline Memory clinics, DSM-IV for N/A 33 N/A 79%
(2008) England RCT data support groups dementia and
and media Clinical Dementia
Rating of mild or
moderate
Gallagher et al (2011) Ireland, Cross- A memory clinic DSM-IV-R, AD 100% 84 63.3 57.1%
Dublin sectional NINCDS-ADRDA,
neurological and
neuropsychological
assessment
García-Alberca et al Spain, Cross- Local health NINCDS-ADRDA AD 100% 80 62.2 77.5%

(2012) Malaga sectional services and the criteria for AD
voluntary sector
Germain et al. (2009) Belgium, Cross 29 specialist NINCDS-ADRDA AD 100% 1091 62.3 63.5%
Denmark, sectional outpatient clinics criteria for AD
France, UK, using data
Germany, from a
Greece, longitudinal
Italy, The prospective
Netherlands cohort study
, Romania,
Spain,
Sweden
Switzerland
Givens et al. (2014) USA; MN, Cross “Population based None caregiver self- N/A 206 82.4 100%
OR, PA sectional listings” report
using a
longitudinal
prospective
cohort study
Hasegawa et al. (2014) Japan; Cross- Two memory Neurological and AD 62.2% 135 N/A 68.2%
Kumamoto sectional clinics neuropsychological VD 16.3%
city on the assessments and LBD 14.1%
island of DSM-III-R for Other dementia
Kyushu dementia. 7.4%
Holland et al. (2010) USA; San Cross- Media, and MMSE ≤ 23 or N/A 47 59.5 100%
Fransisco sectional professional and documented
Bay CA non-professional diagnosis
referrals
Jang et al. (2004) USA; NY Baseline Alzheimer's N/A N/A 160 NS 61.5%
RCT data Disease centre,
adult day care
services, social
services, and
media
Kaiser and Panegyres Australia; Cross- Neuroscience NINCDS-ADRDA FTD 42% 100 62.3 54%
(2007) Perth sectional assessment and criteria for AD, AD 36%
care clinic consensus criteria Primary
for FTD and PPA Progressive
Aphasia 6%
Other dementia
16%
Kurz et al. (2003) Belgium, Cross- General CAMDEX and N/A 188 N/A 66.7%
multiple sectional practitioners, diagnosed in line
locations specialists and with DSM-III-R
psychologists.
Liang et al. (2016) China; Cross- A memory clinic Neuropsychological N/A 139 N/A N/A
Shanghai sectional assessments and
DSM-IV criteria for
dementia
Lowery et al. (2000) UK, Cross- Two dementia Consensus criteria 25% LBD 100 83.5 68%
Tyneside sectional case register for DLB and 75% AD
and cohorts NINCDS-ADRDA
Birmming- for probable AD
ham
Lu and Austrom (2005) USA; OH Cross- University N/A N/A 97 N/A 73.2%
sectional Alzheimer
Disease Center
Caregiver
Registry
Luchsinger et al. USA, NY Baseline Memory clinics, Documented N/A 139 59.3 N/A
(2015) RCT data physicians, health diagnosis; unknown
fairs and talks, procedure
support groups
and media
Mahoney et al. (2005) UK; Cross- Local psychiatric NINCDS-ADRDA AD 100% 153 64 69.9%
London and sectional services, the criteria for AD and
South-East voluntary sector, DSM-IV
regions nursing and
residential homes
McConaghy and Australia, Cross- Homecare MMSE and other N/A 42 62 76.2%
Caltabiano (2005) North sectional dementia services methods N/A
Queensland
Medrano et al. (2014) Dominican Cross- A health database NINCDS-ADRDA AD 100% 67 61 84%
Republic; sectional of over 1500 criteria for AD
multiple patients
locations
Orgeta and Lo Sterzo UK; Cross- Local voluntary N/A. N/A 170 62.4 81.2%
(2013) multiple sectional sectors supporting
locations caregivers of PwD
Ostojic et al. (2014) Croatia, Cross- Psychiatric DSM-IV criteria for AD 100% 30 57.7 73.3%
Zagreb sectional hospital AD
Piercy et al. (2013) USA; UT Cross N/A N/A AD 60% 256 67.5 76%
sectional VD 14%
using data Other dementia
from a 16%
longitudinal
prospective
cohort study
Raggi et al. (2015) Italy; Sicily Cross- Outpatients in DSM-5 criteria for AD 100% 73 N/A N/A
sectional community AD, medical history,
neurological and
neuropsychological
assessments
Riedel et al. (2016) Germany; Cross- Referrals from MMSE AD 100% 403 62.1 69%
multiple sectional office-based
locations neurologists
Roche et al. (2015) Germany; Cross- Caregiver support Medical diagnosis; FTD 100% 94 59.1 72.3%
multiple sectional groups, German procedure unknown
locations Alzheimer's
Association, and
German FTD
consortium
Rosness et al. (2011) Norway; Cross- A memory clinic ICD-10 criteria for AD 77.6% 49 60.3 69.4%
Oslo sectional early onset FTD 14.3%
dementia, physical VD 6.1%
and neurological LBD 2%
assessments
Roth et al. (2008) USA; AL Baseline Multiple Medical diagnosis of N/A 1183 62.2 81.5%
MA, TN, RCT data community sites probable AD or
FL, CA and and health social related dementia
PA. agency settings (unknown
procedure) or
MMSE < 24
Sansoni et al. (2014) Italy; Rome, Cross- Three ambulatory N/A N/A 34 59.2 100%
Florence sectional care clinics
and Genoa
Simpson (2010) USA; TX Cross- Flyers distributed N/A AD 71.3% 80 63.3 88.8%
sectional by a geriatric VD 11.3%
psychiatric LBD/FTD 7.5%
service, support Mixed 1.3%
groups, respite Alcohol induced
care and outreach 1.3%
educational Unknown 7.5%
programmes
Slachevsky et al. Chile; Cross- Referrals from N/A N/A 291 60.1 75.3%
(2013) primarily sectional primary care
Santiago centres,
neurological
consultations and
support groups.
Sleath et al. (2005) USA; Cross- A national ICD-9 criteria for N/A 2032 68.1 100%
multiple sectional database AD or VD
locations
and Puerto
Rico
Valimaki et al. (2015)

Finland; Longitudinal Three hospitals NINCDS-ADRDA AD 100% 170 65.7 66.5%
three prospective criteria for AD and
locations cohort DSM-IV
unnamed
Waite et al. (2004)

UK; Cross- Referrals from DSM-IV criteria for AD 100% 72 80 80%
London sectional two old age dementia and a
psychiatry MMSE score of <
services and two 24
dementia care
centres
Table 3: Characteristics of included studies (N = 43)
Study Relationship Depression Cut-off Depression Burden Cut-off Burden Quality score
measure prevalence % measure prevalence % (risk)
Adams et al. (2002) Spouse 100% OAHMQ >11 30.2% N/A N/A N/A 5 (high)
Arango et al. (2009) Spouse 54.8% PHQ-9 >5 39.7% ZBI-22 ≥21 68.5% 5 (high)
Child 41.1% item
Other relative
4.1%
Bednarek et al. (2016) Spouse 43.9% CES-D ≥16 39% N/A N/A N/A 6 (high)
Child 29.3%
Other relative
26.8%
Bejjani et al. (2016) Unknown CES-D ≥16 13.6% N/A N/A N/A 5 (high)
Berger et al. (2005) Spouse 69% BDI & >10 26.3% N/A N/A N/A 7 (medium)
Child 27 % GDS 15-item
Other 4% ≥5
Borsje et al. (2016) Spouse 65 % CES-D ≥16 23.1% N/A N/A N/A 8 (medium)
Child 29.1%
Other 5.9%
Brodaty et al. (2014) Spouse 71.2% N/A N/A N/A ZBI-22 ≥21 50% 8 (medium)
Child 21.7 % item
Other 7.1%
Caspar and O’Rourke N/A CES-D ≥16 14.7% N/A N/A N/A 5 (high)
(2009)
Cheng et al. (2013) Spouse 32% HRSD >6 27.5% N/A N/A N/A 9 (medium)
Child 59%
Other relative
8%
Contador et al. (2012) Spouse 28.5% GADS >2 36.2% N/A N/A N/A 11 (low)
Child 51.5%
Other relative
20%
Covinsky et al. (2003) Spouse 50.5% GDS 15-item ≥6 32% N/A N/A N/A 7 (medium)
Child 36.7%
Other 12.9%
Cucciare et al (2010) Spouse 23.5% Interview, N/A 16.9% N/A N/A N/A 7 (medium)
Daughter, SCID-I for
daughter-in-law DSM-IV
and
granddaughter
76.5%
Epstein-Lubow et al. Spouse 61% N/A N/A N/A ZBI-22 ≥29 45.5% 6 (high)
(2008) Child 39% item
Gallagher et al (2011) Spouse 64.3% CES-D-10 ≥10 33.3% N/A N/A N/A 6 (high)
No other details
specified
García-Alberca et al Spouse 38.8% BDI Spanish >20 53.7% N/A N/A N/A 8 (medium)
(2012) Child 43.8% version
Sibling 7.4%
Other relative
10%
Germain et al. (2009) Spouse 52.2% N/A N/A N/A ZBI-22item ≥21 45% 9 (medium)
Child 36.7%
Friend 2.00%
Other 9.1%
Givens et al. (2014) Spouse 63.6% CES-D ≥16 22.8% N/A N/A N/A 5 (high)
No other details
specified
Hasegawa et al. (2014) Spouse 37% CES-D ≥16 32.6% N/A N/A N/A 7 (medium)
Child 48%
Other 15%
Holland et al. (2010) Spouse 39% CES-D ≥16 46.8% N/A N/A N/A 9 (medium)
Daughters 54%
Daughter-in-
law 7%
Jang et al. (2004) Spouse 100% GDS >11 41.9% N/A N/A N/A 9 (medium)
Kaiser and Panegyres Spouse 100% BDI >10 57% N/A N/A N/A 7 (medium)
(2007)
Kurz et al. (2003) Spouse 53.6% BDI-short ≥5 42.6% N/A N/A N/A 6 (high)
Child 26.6% form
Sibling 3.9%
Other 15.8%
Liang et al. (2016) N/A HADS ≥8 20.9% N/A N/A N/A 6 (high)
Chinese
version
Lowery et al. (2000) Spouse 44% MADRS, N/A 3% N/A N/A N/A 5 (high)
Child 40% interview
Other relative and RDC
16% criteria
Lu and Austrom (2005) Spouse 75.3% CES-D ≥16 28.9% N/A N/A N/A 6 (high)
Child 19.6%
Daughter-in-
law 3.1%
Other 2%
Luchsinger et al. (2015) Spouse 38.8% GDS ≥10 51.1% ZBI-22 ≥21 88.5% 8 (medium)
Child 56.8% item
Other 4.3%
Mahoney et al. (2005) Spouse 44.4% HADS ≥11 10.5% N/A N/A N/A 9 (medium)
Child 44.4%
Friends 4.6%
Other relative
6.6%
McConaghy and Spouse 54.8% CES-D ≥16 59.5% ZBI-22 ≥21 78.6% 7 (medium)
Caltabiano (2005) Child 35.7 item
Friends 2.4%
Other relative
7.1%
Medrano et al. (2014) Spouse 15% HRSD >8 43.3% ZBI-22 ≥46 35.8% 7 (medium)
Child 55% Spanish item
Grandchild12% Version Spanish
Brother 9% version
Other relative
9%
Orgeta and Lo Sterzo Spouse 52.6% HADS ≥8 54.7% N/A N/A N/A 6 (high)
(2013) Child 29.3%
Other relative
18.1%
Ostojic et al. (2014) Spouse 26.7% HADS ≥11 26.7% N/A N/A N/A 5 (high)
Child 63.3% Croatian
translation
Piercy et al. (2013) Spouse 45% BDI-II ≥14 16.4% N/A N/A N/A 8 (medium)
Child 50%
Other 5%
Raggi et al. (2015) Spouse 57.5% N/A N/A N/A CBI >24 60.3% 5 (high)
Child 38.4%
Sibling 2.7%
Nephew 1.4%
Riedel et al. (2016) Spouse 48.5% DSQ ≥10 43.7% N/A N/A N/A 10 (low)
Child 36.3%
Son-in-
law/daughter-
in-law 5.5%
Other 9.4%
Roche et al. (2015) Spouse 79.8% BDI-II ≥13 48.9% N/A N/A N/A 4 (high)
no other details
specified
Rosness et al. (2011) Spouse 100% GDS-15 item ≥5 53.1% N/A N/A N/A 7 (medium)
Roth et al. (2008) Spouse 48.2% CES-D ≥16 41% N/A N/A N/A 7 (medium)
Child 41.8%
Other 10.1%
Sansoni et al. (2014) Spouse 73.53% GDS >15 52.9% N/A N/A N/A 6 (high)
Sister 1.94%
Daughter
11.76% Friend
2.94% Other
relative 8.82%
Simpson (2010) Spouse 50.1% CES-D ≥16 31.3% N/A N/A N/A 10 (low)
Child 41.3%
Sibling 2.5%
Other 6.3%
Slachevsky et al. (2013) Spouse 40% N/A N/A N/A ZBI-22 >46 74.2% 8 (medium)
Child 43% item
Sibling 5% Chilean
Relative in law version
4% Friends 1%
Sleath et al. (2005) Spouse 92% CES-D ≥9 31% N/A N/A N/A 9 (medium)
Sister and modified
daughter 8% version
Spouse 70.24% BDI >10 44.1% N/A N/A N/A 7 (medium)

Valimaki et al. (2015) Non-spouse
29.76%
Spouse 45.8% GDS-15 item ≥5 43.1% N/A N/A N/A 6 (high)

Waite et al. (2004) Daughters
31.9% Friends
or other
relatives 22.5%
Note: N/A = not available. Location: AL = Alabama; CA = California; FL = Florida; IL = Illinois; MA = Massachusetts; MN = Minnesota; NY
= New York; OH = Ohio; OK = Oklahoma; OR = Oregon; PA = Pennsylvania; RI = Rhode Island; TN = Tennessee; TX = Texas; UK = United
Kingdom; USA = United States of America; UT = Utah. Design: RCT = Randomized Controlled Trial. Recruitment source: USC = University
of Southern California. Dementia diagnostic tools: CAMDEX = Cambridge Mental Disorders of the Elderly Examination (Roth et al., 1986);
CDRS = Clinical Dementia Rating Scale (Hughes, Berg, Danziger, Coben & Martin, 1982); DSM-III-R/IV/IV-R/5 = Diagnostic and Statistical
Manual of Mental Disorders third edition revised (APA, 1987)/fourth edition (APA, 1994)/ fourth edition revised (APA, 2000)/fifth edition
revised (APA, 2013); ICD-9/10 = International Statistical Classification of Diseases and Related Health Problems-9th Revision (WHO,
1978)/10th Revision (WHO, 1992); ICPC-2 = International Classification of Primary Care-Second Edition (WHO, 2003); MMSE = Mini Mental
State Exam (Folstein, Folstein & McHugh, 1975); NINCDS-ADRDA = National Institute of Neurological and Communicative Disorders and
Stroke and the Alzheimer's Disease and Related Disorders Association (McKhann et al., 1984). Dementia terms: AD = Alzheimer’s Disease;
FTD = Frontotemporal Lobe Dementia; DLB = Dementia with Lewy Bodies; PPA = Primary Progressive Aphasia; PwD = people with
dementia; VD = Vascular Dementia. Depression measures: BDI-I/short form/II/Spanish Version/Chilean Version = Beck Depression Inventory
(Beck et al., 1961)/short-form (Beck & Beck,1972)/second edition (Beck, Steer, Ball & Ranieri, 1996)/Spanish Version (Conde & Useros,
1975)/Chilean Version; CES-D/-10/modified version = Centre for Epidemiological Studies-Depression (Radloff 1977)/10-item (Andresen,
Malmgren, Carter, & Patrick, 1994)/modified version (Hays, Blazer & Gold, 1993); DSQ = Depression Screening Questionnaire (Wittchen,
Höfler, & Meister, 2001); GADS = Goldberg Anxiety and Depression Scale (Goldberg, Bridges, Duncan-Jones, & Grayson, 1988); GDS/-15-
item = Geriatric Depression Scale (Yesavage et al., 1983)/15-item (Yesavage & Sheikh, 1986); HADS original/Chinese Version = Hospital
Anxiety and Depression Scale (Zigmond & Snaith, 1983)/-Chinese Version (Chan, Leung, Fong, Leung, & Lee, 2010); HRSD/-Spanish version
= Hamilton Rating Scale for Depression (Hamilton, 1980)/-Spanish version (Ramos-Brieva, 1986); MADRS = Montgomery Asberg Depression
Rating Scale (Montgomery & Asberg, 1979); PHQ-9 = Patient Health Questionnaire-9 (Kroenke, Spitzer & Williams, 2001); OAMHQ = The
Older Adult Health and Mood Questionnaire (Kemp and Adams, 1995); SCID-I = Structured Clinical Interview for DSM-IV Axis I Disorders
(First, Spitzer, Gibbon, & Williams, 2008); RDC = Research Diagnostic Criteria (Spitzer & Robins, 1978). Burden measures: ZBI 22-
item/Spanish version/Chilean version = Zarit Burden Interview (Zarit et al., 1980)/Spanish version (Martín et al., 1996)/Chilean Version
(Breinbauer et al., 2009); CBI = Caregiver Burden Inventory (Novak & Guest, 1989).

Prevalence of Depression and Burden Among Informal Caregivers of People With Dementia: A Meta-Analysis

Uploaded by

Copyright:

Available Formats

Prevalence of Depression and Burden Among Informal Caregivers of People With Dementia: A Meta-Analysis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Prevalence of Depression and Burden Among Informal Caregivers of People With Dementia: A Meta-Analysis

Uploaded by

Copyright:

Available Formats

View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

Depression and burden in caregivers of people with dementia

Prevalence of depression and burden among informal caregivers

of people with dementia: a meta-analysis

Rebecca N. Collinsa and Naoko Kishitab

Qualifications: BSc, MSc

Mailing address: Department of Clinical Psychology, University of East Anglia,

Norwich, NR4 7TJ, England, UK

E-mail address: [email protected]

Contact telephone number: 07872578788

(PsycARTICLES, PsycINFO, MEDLINE Complete, SCOPUS, Web of Science and

heterogeneity among prevalence estimates was high. Depression prevalence estimates

targeting these difficulties to do so.

Keywords: carers; depressive symptoms; caregiver-burden; epidemiology.

be approximately 131 million PwD worldwide (Alzheimer’s Disease International 2015). It

first century (Alzheimer’s Society 2014).

meaningful (Cohen, Colantonio and Vernich 2002). However, there is an abundance of

without dementia (Pinquart and Sörensen 2003).

In this review, ‘caregiver burden’ (here on referred to as ‘burden’) is conceptualised

they are to expedite nursing home placement (Gaugler et al. 2005).

Research exploring burden in caregivers of PwD has tended to focus on the

behavioural problems or psychological symptoms and extent of personality change, and

caregiver related predictors including sociodemographic variables and psychological health

inform the provision of services.

Depressive symptoms can include a persistent sadness/low mood, marked loss of

interest or pleasure in activities, disturbed sleep, decreased or increased appetite or weight,

Studies-Depression (CES-D; Radloff 1977).

Caregivers who have depression typically experience problems in daily functioning

Although a smaller longitudinal study found the prevalence of suicidal thoughts to be

A previous meta-analysis found a moderately significant difference in depressive

prevalence of depressive disorders among informal caregivers of PwD, assessed via

prevalence of depression among caregivers of people with Alzheimer’s disease (AD),

A contextual model (Fig 1) by Williams (2005) adapted from that of Dilworth-

of a caregiver of someone with dementia experiencing depression. Among other factors,

<insert fig 1. here>

Indeed, one meta-analysis found the prevalence of depression to be higher in female

bias (Harbord 2010).

conducting a current comprehensive meta-analysis with the following objectives:

spousal and non-spousal caregivers.

(3) To explore moderator variables including the methodological quality

and Meta-Analyses guidelines (PRISMA; Moher et al. 2009).

excluded, such as experimental studies, qualitative studies, and review articles.

The population studied were informal caregivers of PwD. Studies involving

or they actively excluded caregivers experiencing a depressive episode.

assessed by semi-structured or structured diagnostic interviews based on criteria by DSM-

III(-R)/IV or ICD-10, or validated self-report measures with specified clinical cut-offs.

or burden prevalence data were excluded.

disclosing depressive symptoms.

A comprehensive search of the literature was conducted. The databases of

PsycARTICLES, PsycINFO, MEDLINE Complete, SCOPUS and Web of Science were

searched to identify relevant published articles. Unpublished articles including dissertations

obtained via The Cochrane Online Library.

and no limits were applied to language.

< insert table 1 here>

duplicate articles removed. Eligibility assessment was conducted in a non-blinded manner.

Data collection process

randomly-selected study by both authors collaboratively and refined accordingly. In order to

version 3; Borenstein et al. 2005).