DOI: 10.1111/ggi.

14250

ORIGINAL ARTICLE
SOCIAL RESEARCH, PLANNING AND PRACTICE

Factors associated with psychological distress in family

caregivers: Findings from nationwide data in Japan
Yu Sun,1 Masao Iwagami,2,3 Taeko Watanabe,3 Nobuo Sakata,2,3 Takehiro Sugiyama,2,3,4,5
Atsushi Miyawaki3,6 and Nanako Tamiya2,3
1
Department of Health Services Research, Graduate School of Comprehensive Human Sciences, University of Tsukuba,
Ibaraki, Japan
2
Department of Health Services Research, Faculty of Medicine, University of Tsukuba, Ibaraki, Japan
3
Health Services Research and Development Center, University of Tsukuba, Ibaraki, Japan
4
Diabetes and Metabolism Information Center, Research Institute, National Center for Global Health and Medicine,
Tokyo, Japan
5
Institute for Global Health Policy Research, Bureau of International Health Cooperation, National Center for Global
Health and Medicine, Tokyo, Japan
6
Department of Public Health, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan

Correspondence
Dr Masao Iwagami MD MPH MSc
PhD, Department of Health
Services Research, University of
Tsukuba, Ibaraki, Japan. Institutes
of Medicine building #861, 1-1-1
Tenno-dai, Tsukuba 305-8575,
Ibaraki, Japan.
Email: iwagami-tky@umin.ac.jp
Received: 13 January 2021
Revised: 26 May 2021
Accepted: 7 July 2021
Aim: Family caregivers can experience psychological distress, resulting in physical and men-
tal health problems, and discontinuation of caregiving. This study’s objective was to examine
factors associated with caregiver psychological distress.
Methods: We analyzed data from the Comprehensive Survey of Living Conditions in 2007,
2010, 2013 and 2016, which included 12 504 pairs of caregivers and care recipients sharing a
household in Japan. Kessler’s Psychological Distress Scale (K6) was used to measure caregiver
psychological distress. Multivariable logistic regression analyses identified factors associated
with caregiver psychological distress (K6 score ≥5).
Results: Caregivers’ median K6 score was 3 (interquartile range 0–7), and 38.6% had a K6
score ≥5. K6 scores ≥5 were positively associated with female sex (adjusted odds ratio 1.35,
95% CI 1.21–1.51), poor caregiver health status (compared with “very good,” 9.48, 95% CI
7.91–11.37 for “not very good/poor”), longer care time (compared with “help only when
needed,” 1.40, 95% CI 1.25–1.58 for “almost all day”) and dementia (1.16, 95% CI 1.05–
1.28), lower respiratory tract disease (1.25, 95% CI 1.06–1.49) and diabetes (1.16, 95% CI
1.00–1.33) in care recipients. K6 scores ≥5 were negatively associated with older caregiver age
(compared with 20–54 years, 0.65, 95%CI 0.58–0.74 for 55–64 years, 0.54, 95%CI 0.46–0.63
for 65–74 years and 0.50, 95% CI 0.40–0.62 for ≥75 years), employment (0.88, 95% CI 0.80–
0.97) and being a care recipient’s child-in-law (compared with spouse 0.75, 95% CI
0.61–0.92).

Conclusions: The findings identified several factors associated with caregiver psychological
distress, showing that particular attention might need to be paid to caregivers with these risk
factors. Geriatr Gerontol Int 2021; ••: ••–••.

Keywords: caregiver burden, caregivers, observational study, psychological distress, risk

factors.

Introduction medical insurance system, in 2000.2 Although at-home long-term

The global population is rapidly aging, resulting in an increase of
older people in need of care, and a subsequent increase in family
caregivers.1 To meet the increasing caregiving demand and miti-
gate care burden among family members, several countries, such
as Germany, the Netherlands and Korea, have introduced a long-
term care system. Japan, which has the largest proportional
population of older adults worldwide, also started a mandatory
long-term care insurance system, distinct from the national
care services have mitigated the care burden for family caregivers
(sometimes called informal caregivers), they still play an important
role in providing care to their family members.3
Since the concept of caregiver burden was first defined in
1980,4 a number of studies have suggested that many caregivers
experience depression when providing care.5 Depression in care-
givers could affect their physical health,5 and could lead to mis-
treatment of care recipients and discontinuation of caregiving.5 To
mitigate depression risk among family caregivers and the related

© 2021 Japan Geriatrics Society. | 1

 Y Sun et al.
consequences, healthcare professionals and social workers need to
assess caregivers and care recipients for characteristics associated
with caregiver psychological distress in clinical and long-term care
settings.
Previous studies suggested that caregivers with poor health
status,6 fewer financial resources7 and longer hours of caregiving,8
as well as women9 and spouses6 who provide care, are at a higher
risk for depression. However, most previous studies have been lim-
ited to a specific disease of care recipients (e.g. Alzheimer’s disease,
cancer, stroke), which limits generalizability and the ability to iden-
tify factors that are common across diagnoses.6 To our knowledge,
no national-level studies have been previously published that
include a wide range of caregiver- and care recipient-related factors
as potential risk factors for psychological distress in caregivers.
To help close this knowledge gap, we aimed to investigate the
association between a wide range of caregiver- and care recipient-
related factors, including care recipient’s disease and psychological
distress in caregivers.
Methods
Data source
The Comprehensive Survey of Living Conditions is a nationwide
cross-sectional survey carried out annually by the Ministry of
Health, Labor and Welfare in Japan.10 Since 1986, large-scale sur-
veys (including household, income and savings, and health and
long-term care questionnaires) have been carried out every
3 years, whereas small-scale surveys (including only household
and income questionnaires) have been carried out during the
other years. Questionnaires that utilize the Kessler Psychological
Distress Scale (K6) have been included in the large-scale surveys
since 2007. The present study used the large-scale survey datasets
from 2007, 2010, 2013 and 2016.
The sampling method of the Comprehensive Survey of Living
Conditions is based on census enumeration districts in Japan. In
2007, 2010, 2013 and 2016, approximately 5500 enumeration dis-
tricts covering nearly 300 000 households were randomly selected,
and all households within each enumeration district were
Figure 1 Flow chart of study participant selection.
2 |
surveyed with household and health questionnaires. Respondents
included all household members, except for individuals who were
temporarily staying outside the house for some reason
(e.g. hospitalized or institutionalized) during the survey period. In
nearly 2500 randomly selected districts, those who required long-
term care (around 6000–8000 people) were asked to answer a
long-term care questionnaire on care recipients. The response
rates for household and health questionnaires were 79.9%,
79.1%, 79.4% and 77.5% for 2007, 2010, 2013 and 2016, respec-
tively. As for the long-term care questionnaire, the response rates
were 89.7%, 87.2%, 82.2% and 89.1% for 2007, 2010, 2013 and
2016.11
De-identified individual-level data from the Comprehensive
Survey of Living Conditions are available for scientific research on
approval by the Ministry of Health, Labor and Welfare, through
application procedures under Article 33 of the Statistics Act.12
The study was also approved by the ethical committee of the insti-
tution that the first author belongs to. The need for informed con-
sent was waived, because the data were anonymized before being
obtained from the Ministry of Health, Labor and Welfare.
Study population
Using the information obtained from the long-term care question-
naire, we first identified 15 239 care recipients and 14 803
corresponding primary caregivers in the same household. In
households with one caregiver and multiple care recipients, we
selected the care recipient with the highest level of care need, or
the older person if there were more than two care recipients who
required the same level of care. We excluded pairs without infor-
mation on the caregiver’s K6 score (n = 2280), without informa-
tion on either person’s age, or that included caregivers aged
<20 years (n = 19). Thus, a final sample of 12 504 pairs was ana-
lyzed (Fig. 1).
Outcome variable
The outcome variable was caregivers’ self-reported psychological
distress, as measured by the Japanese version of the K6.13 The K6
© 2021 Japan Geriatrics Society.
 Psychological distress in caregivers

Table 1 Characteristics of the study sample by each survey year and overall

2007 2010 2013 2016 Total

n = 2357 n = 2801 n = 3582 n = 3764 n = 12 504
Variable n (%) n (%) n (%) n (%) n (%)
Caregiver characteristics
Mean age, years (SD) 61.9 (12.2) 62.5 (12.5) 64.8 (12.0) 65.3 (12.1) 63.9 (12.3)
Age category (years)
20–54 645 (27.4) 685 (24.5) 659 (18.4) 664 (17.6) 2653 (21.2)
55–64 780 (33.1) 1013 (36.2) 1219 (34.0) 1202 (31.9) 4214 (33.7)
65–74 508 (21.6) 558 (19.9) 815 (22.8) 967 (25.7) 2848 (22.8)
≥75 424 (18.0) 545 (19.5) 889 (24.8) 931 (24.7) 2789 (22.3)
Sex: male 598 (25.4) 834 (29.8) 1082 (30.2) 1262 (33.5) 3776 (30.2)
Relationship to care recipient
Spouse 805 (34.2) 950 (33.9) 1394 (38.9) 1478 (39.3) 4627 (37.0)
Child 720 (30.6) 946 (33.8) 1261 (35.2) 1390 (36.9) 4317 (34.5)
Child-in-law 731 (31.0) 793 (28.3) 797 (22.3) 776 (20.6) 3097 (24.8)
Other 101 (4.3) 112 (4.0) 130 (3.6) 120 (3.2) 463 (3.7)
Employment status
Unemployed 1292 (54.8) 1533 (54.7) 2125 (59.3) 2128 (56.5) 7078 (56.6)
Employed 1055 (44.8) 1246 (44.5) 1414 (39.5) 1600 (42.5) 5315 (42.5)
Unknown 10 (0.4) 22 (0.8) 43 (1.2) 36 (1.0) 111 (0.9)
Household expenditure/month
<150 000 JPY 590 (25.0) 716 (25.6) 1041 (29.1) 973 (25.9) 3320 (26.6)
150 000–250 000 JPY 730 (31.0) 951 (34.0) 1251 (34.9) 1334 (35.4) 4266 (34.1)
>250 000 JPY 875 (37.1) 1005 (35.9) 1164 (32.5) 1350 (35.9) 4394 (35.1)
Unknown 162 (6.9) 129 (4.6) 126 (3.5) 107 (2.8) 524 (4.2)
Self-rated health status
Very good 271 (11.5) 283 (10.1) 380 (10.6) 386 (10.3) 1320 (10.6)
Good 327 (13.9) 391 (14.0) 494 (13.8) 601 (16.0) 1813 (14.5)
Fair 1120 (47.5) 1353 (48.3) 1922 (53.7) 1929 (51.3) 6324 (50.6)
Not very good/poor 552 (23.4) 649 (23.2) 771 (21.5) 833 (22.1) 2805 (22.4)
Unknown 87 (3.7) 124 (4.5) 15 (0.4) 15 (0.4) 242 (1.9)
Use of social services (yes) 1861 (79.0) 2195 (78.4) 2872 (80.2) 2940 (78.1) 9868 (78.9)
Someone to help with care (yes) 1094 (46.4) 1392 (49.7) 1766 (49.3) 2051 (54.5) 6303 (50.4)
Care time spent per day
Help only when needed 904 (38.4) 1175 (42.0) 1558 (43.5) 1759 (46.7) 5396 (43.2)
2–3 h 284 (12.1) 309 (11.0) 413 (11.5) 428 (11.4) 1434 (11.5)
About half the day 257 (10.9) 307 (11.0) 382 (10.7) 425 (11.3) 1371 (11.0)
Almost all day 500 (21.2) 590 (21.1) 831 (23.2) 725 (19.3) 2646 (21.2)
Other 262 (11.1) 275 (9.8) 308 (8.6) 323 (8.6) 1168 (9.3)
Unknown 150 (6.4) 145 (5.2) 90 (2.5) 104 (2.8) 489 (3.9)
Care recipient characteristics
Mean age, years (SD) 81.5 (9.3) 82.1 (9.1) 82.7 (9.0) 82.8 (9.2) 82.4 (9.1)
Age category (years)
40–74 453 (19.2) 486 (17.4) 604 (16.9) 591 (15.7) 2134 (17.1)
75–84 938 (39.8) 1073 (38.3) 1287 (35.9) 1357 (36.1) 4655 (37.2)
85–95 849 (36.0) 1092 (39.0) 1472 (41.1) 1569 (41.7) 4982 (39.8)
≥95 117 (5.0) 150 (5.4) 219 (6.1) 247 (6.6) 733 (5.9)
Sex: male 849 (36.0) 962 (34.3) 1318 (36.8) 1403 (37.3) 4531 (36.2)
Care need level
Support levels 1–2 511 (21.7) 627 (22.4) 794 (22.2) 965 (25.6) 2897 (23.2)
Care need levels 1–3 1313 (55.7) 1556 (55.6) 2058 (57.5) 2135 (56.7) 7062 (56.5)
Care need levels 4–5 476 (20.2) 547 (19.5) 673 (18.8) 596 (15.8) 2292 (18.3)
Unknown 57 (2.4) 71 (2.5) 57 (1.6) 68 (1.8) 253 (2.0)
Disease
Stroke 799 (33.9) 877 (31.3) 1045 (29.2) 958 (25.5) 3679 (29.4)
Dementia 571 (24.2) 721 (25.7) 961 (26.8) 1028 (27.3) 3281 (26.2)
Cardiac disease 289 (12.3) 355 (12.7) 457 (12.8) 465 (12.4) 1566 (12.5)
Cancer 78 (3.3) 100 (3.6) 155 (4.3) 158 (4.2) 491 (3.9)

(Continues)

© 2021 Japan Geriatrics Society. | 3

 Y Sun et al.

Table 1 Continued

2007 2010 2013 2016 Total

n = 2357 n = 2801 n = 3582 n = 3764 n = 12 504
Variable n (%) n (%) n (%) n (%) n (%)
Lower respiratory tract disease 125 (5.3) 166 (5.9) 215 (6.0) 220 (5.8) 726 (5.8)
Joint disease 415 (17.6) 544 (19.4) 636 (17.8) 648 (17.2) 2243 (17.9)
Diabetes 202 (8.6) 274 (9.8) 352 (9.8) 340 (9.0) 1168 (9.3)
Parkinson’s disease 118 (5.0) 138 (4.9) 173 (4.8) 169 (4.5) 598 (4.8)
Vision or hearing impairment 219 (9.3) 230 (8.2) 318 (8.9) 280 (7.4) 1047 (8.4)
Fractures 420 (17.8) 533 (19.0) 735 (20.5) 782 (20.8) 2470 (19.8)
Spinal cord injury 73 (3.1) 72 (2.6) 114 (3.2) 125 (3.3) 384 (3.1)
Senility 636 (27.0) 769 (27.5) 985 (27.5) 998 (26.5) 3388 (27.1)
Other 239 (10.1) 293 (10.5) 328 (9.2) 414 (11.0) 1274 (10.2)
Outcome
K6 score ≥5 934 (39.6) 1068 (38.1) 1391 (38.8) 1428 (37.9) 4821 (38.6)
K6 score ≥13 159 (6.8) 160 (5.7) 213 (6.0) 215 (5.7) 747 (6.0)

is a well validated screening tool frequently used to detect symp-
toms of major depression and dysthymia. The K6 includes six
items: “During the past 30 days, about how often did you feel
(a) nervous, (b) hopeless, (c) restless or fidgety, (d) so depressed
that nothing could cheer you up, (e) that everything was an effort
or (f) worthless?” Responses to each question were rated on a
5-point scale ranging from 0 to 4 points, and a summed total
score (range 0–24 points) was calculated. A higher total score cor-
responds to a higher level of psychological distress. Following pre-
vious studies, we defined a K6 score ≥5 as indicating moderate
psychological distress (in the main analysis), and ≥13 as indicating
severe psychological distress (in a sensitivity analysis).14
Exposure variables
We identified demographic information for caregivers, including
age (categorized as 20–54, 55–64, 65–74 or ≥75 years), sex, rela-
tionship to the care recipient (spouse, child, child-in-law and
other), employment status (employed, unemployed), household
expenditure per month (<150 000, 150 000 250 000 and
>250 000 Japanese yen), self-rated health status (“very good,”
“good,” “fair” and “not good” or “poor”), use of social services
(including long-term care insurance services or other private ser-
vices, such as food delivery and outing support services; dichoto-
mized as yes or no), whether there was another person to help
with caregiving (yes or no) and time spent providing care per day
(only when needed, 2 3 h, about half the day, almost all day and
other). We used household expenditure as a surrogate indicator of
household economy.15 Self-rated health status, which is contained
in health-related quality of life core module, has been shown to be
a powerful predictor of all-cause mortality in general
populations.16
We also identified care recipient factors, including age (catego-
rized as 40–74, 75–84, 85–95 or ≥95 years), sex, care need level in
the Japanese long-term care insurance system (support levels 1–2,
care need levels 1–3 and care need levels 4–5) and diseases poten-
tially associated with the initiation of long-term care (yes or no for
each disease: stroke, dementia, cardiac disease, cancer, lower
respiratory tract disease, joint disease, diabetes, Parkinson’s dis-
ease, vision or hearing impairment, fractures, spinal cord injury,
senility and other). Care need level is a nationally standardized
certification that incorporates the individual’s physical and cogni-
tive functions, with support level 1 representing the lowest level,
and care need level 5 representing the highest level of requirement
for long-term care.17
Statistical analysis
First, we examined the summary statistics of variables and K6
score distribution of caregivers overall. We also estimated the pro-
portion of caregivers with K6 scores ≥5 and ≥13, overall and by
each survey year to identify temporal trends, using the
Cochran Armitage test (i.e. a test for trends among binominal
proportions against ordinal explanatory variables). Next, we com-
pared caregivers with K6 scores <5 and ≥5, as well as those with
K6 scores <13 and ≥13, for each variable, using χ2-tests. We car-
ried out a multivariable logistic regression analysis to identify fac-
tors associated with K6 scores ≥5 (in the main analysis) and ≥13
(in a sensitivity analysis), where the exposures were survey year,
caregiver factors (age, sex, relationship to the care recipient,
employment status, household expenditure per month, self-rated
health status, use of social services, whether there was another
person to help with caregiving, time spent providing care per day)
and care recipient factors (age, sex, care need level, diseases poten-
tially associated with the initiation of long-term care). Additionally,
as we suspected an interaction with caregiver gender during the
main analysis, we carried out separate multivariable logistic regres-
sion analyses for K6 scores ≥5 in male and female caregivers.9,18
As a post-hoc analysis, we compared the characteristics between
the groups of caregiver’s relationship to care recipient to investi-
gate between-group differences in characteristics. All analyses
were carried out using Stata version 15 (Stata Corp., College Sta-
tion, TX, USA). P-values <0.05 were considered statistically
significant.
Results
The characteristics of caregivers and their care recipients, overall
and by each survey year, are presented in Table 1. The mean care-
giver age increased from 61.9 in 2007 to 65.3 in 2016, and the
proportion of male caregivers showed an increasing trend. With
respect to the caregiver’s relationship with care recipient, the pro-
portions of spouses and children increased, whereas that of
children-in-law decreased. Stroke was the most frequent disease
until 2013, at which point dementia became the most frequent
disease cited as the primary reason for long-term care need.

4 | © 2021 Japan Geriatrics Society.

 Psychological distress in caregivers
Figure 2 Distribution of Kessler’s
psychological distress scale
(K6) scores.
K6 score distribution is shown in Figure 2. The distribution
was right-skewed, with 0 being the most frequent K6 score. Over-
all, 38.6% of caregivers (4821/12 504) had a K6 score ≥5. By sur-
vey year, the proportion of caregivers with a K6 score ≥5
decreased slightly from 39.6% in 2007 to 37.9% in 2016; how-
ever, this was not statistically significant (P for trend = 0.30).
Table 2 shows the characteristics caregivers and their care
recipients in caregivers with K6 scores <5 and those with scores
≥5. Caregivers and their care recipients in the higher K6 group
were younger than those in the lower K6 group. In the higher K6
group, the proportion of spouses was higher, whereas that of
children-in-law was lower. In the lower K6 group, employed care-
givers were more common, and caregiver self-rated health status,
time spent providing care per day and care need level were signifi-
cantly better. Regarding care recipients’ diseases, stroke, dementia,
cardiac disease, lower respiratory tract disease and diabetes were
significantly more common in the higher K6 group. Table 3 shows
the results of a multivariable logistic regression analysis for care-
givers with a K6 score ≥5. Female sex, longer time spent providing
care, caregiver’s own poor health status, and dementia, lower
respiratory tract disease and diabetes in care recipients showed
positive associations with a K6 score ≥5, whereas caregiver age
(older), employment and being the child-in-law of a care recipient
(compared with the spouse of care recipient) showed negative
associations.
In a sensitivity analysis, 6% (747/12 504) of caregivers overall
had a K6 score ≥13, decreasing from 6.8% in 2007 to 5.7% in
2016, with no statistical significance (P for trend = 0.18). The
results of bivariate and multivariable analyses for K6 scores ≥13
were generally similar to those for K6 scores ≥5, whereas associa-
tions with household expenditure and senility in care recipients
were specific to K6 scores ≥13 (Appendix S1). In an additional
analysis by sex, psychological distress was negatively associated
with being employed and having higher household expenditure
only among male caregivers, and with being the child-in-law of a
care recipient only among female caregivers (Appendices S2 and
S3). In our post-hoc analysis, most of the child-in-law caregivers
were female (98.3%), with higher household expenditure per
month, higher rates of use of social services and having someone
© 2021 Japan Geriatrics Society.
to help with caregiving, and shorter time spent providing care per
day compared with spouse and child. Meanwhile, their care recipi-
ents were relatively old, with relatively high rates of dementia and
senility (Appendix S4).
Discussion
In the present repeated cross-sectional study using a nationally
representative sample of caregivers and their care recipients in
Japan, we found that higher levels of caregiver distress were asso-
ciated with a wide range of caregiver characteristics, including
female sex, poor self-reported health status, providing care for
longer periods of time, younger age, being unemployed and being
the spouse or child of a care recipient (compared with the child-
in-law). We also found higher distress levels among caregivers
whose care recipients had dementia, lower respiratory tract disease
or diabetes.
In the present study, 38.6% and 6.0% of caregivers had K6
scores ≥5 and ≥13, respectively. Notably, we found a high preva-
lence of psychological distress in caregivers, compared with a pre-
vious study that used the same data source and included both
caregivers and non-caregivers (~29% and 4% had K6 scores ≥5
and ≥13, respectively).19 This finding is in line with a previous
study in Japan, showing the increased risk for depressive symp-
toms in caregivers compared with non-caregivers.7 Given this
higher prevalence of psychological distress, health professionals
and social workers might need to screen caregivers for depressive
symptoms.
The present findings regarding the negative association
between employment and caregiver distress are in line with a pre-
vious cross-sectional study using nationally representative Japa-
nese data that reported that being employed increased the
probability of experiencing satisfaction with life, especially in care-
givers.18 Although some studies suggested employment increases
stress and role overload for caregivers, other studies showed that
employment provides caregivers with benefits, such as income
and enhanced interpersonal relationships.18 Notably, we found
substantial sex differences, which is in line with a previous
| 5
 Y Sun et al.

Table 2 Characteristics of caregivers with Kessler’s Psychological Distress Scale (K6) scores of <5 and ≥5 and their care recipients

K6 < 5 K6 ≥ 5 P-value
n = 7683 n = 4821
Variable n (%) n (%)
Survey year
2007 1423 (18.5) 934 (19.4) 0.556
2010 1733 (22.6) 1068 (22.2)
2013 2191 (28.5) 1391 (28.9)
2016 2336 (30.4) 1428 (29.6)
Caregiver characteristics
Age category (years) <0.001
20–54 1517 (19.6) 1146 (23.8)
55–64 2703 (35.2) 1511 (31.3)
65–74 1804 (23.5) 1044 (21.7)
≥75 1669 (21.7) 1120 (23.2)
Sex: male 2478 (32.3) 1298 (26.9) <0.001
Relationship to care recipient <0.001
Spouse 2740 (35.7) 1887 (39.1)
Child 2653 (34.5) 1664 (34.5)
Child-in-law 2008 (26.1) 1089 (22.6)
Other 282 (3.7) 181 (3.8)
Employment status <0.001
Unemployed 4173 (54.3) 2905 (60.3)
Employed 3435 (44.7) 1880 (39.0)
Unknown 75 (1.0) 36 (0.8)
Household expenditure/month 0.903
<150 000 JPY 2042 (26.6) 1278 (26.5)
150 000–250 000 JPY 2609 (34.0) 1657 (34.4)
>250 000 JPY 2715 (35.3) 1679 (34.8)
Unknown 317 (4.1) 207 (4.3)
Self-rated health status <0.001
Very good 1113 (14.5) 207 (4.3)
Good 1339 (17.4) 474 (9.8)
Fair 4022 (52.4) 2302 (47.8)
Not very good/poor 1052 (13.7) 1753 (36.4)
Unknown 157 (2.0) 85 (1.8)
Use of social services (yes) 6039 (78.6) 3829 (79.4) 0.273
Someone to help with care (yes) 3864 (50.3) 2439 (50.6) 0.745
Care time spent per day <0.001
Help only when needed 3537 (46.0) 1859 (38.6)
2–3 h 884 (11.5) 550 (11.4)
About half the day 781 (10.2) 590 (12.2)
Almost all day 1450 (18.9) 1196 (24.8)
Other 728 (9.5) 440 (9.1)
Unknown 303 (3.9) 186 (3.9)
Care recipient characteristics
Age category (years) <0.001
40–74 1265 (16.5) 869 (18.0)
75–84 2762 (36.0) 1893 (39.3)
85–95 3161 (41.1) 1821 (37.8)
≥95 495 (6.4) 238 (4.9)
Sex: male 2603 (33.9) 1928 (40.0) <0.001
Care need level <0.001
Support levels 1–2 1905 (24.8) 992 (20.6)
Care need levels 1–3 4283 (55.8) 2779 (57.6)
Care need levels 4–5 1350 (17.6) 942 (19.5)
Unknown 145 (1.9) 108 (2.2)
Disease
Stroke 2151 (28.0) 1528 (31.7) <0.001

(Continues)

6 | © 2021 Japan Geriatrics Society.

 Psychological distress in caregivers
Table 2 Continued
K6 < 5
n = 7683
Variable n (%)
Dementia 1919 (25.0)
Cardiac disease 919 (12.0)
Cancer 281 (3.7)
Lower respiratory tract disease 390 (5.1)
Joint disease 1374 (17.9)
Diabetes 650 (8.5)
Parkinson’s disease 369 (4.8)
Vision or hearing impairment 624 (8.1)
Fractures 1511 (19.7)
Spinal cord injury 234 (3.1)
Senility 2136 (27.8)
Other 763 (9.9)
longitudinal study carried out in Japan.20 Due to differences in
traditional sex roles, Japanese men tend to view their role as
obtaining paid work and providing their family’s primary eco-
nomic support; thus, a loss of employment and the related finan-
cial stress might lead to psychological distress. In 2018 alone, it
was reported in Japan that 100 000 people left their jobs and
260 000 people did not look for work, despite wanting to be
employed, due to providing care to family members.21 Given
Japan’s aging society and the continued surge in older people
requiring long-term care, the number of workers who provide care
for family members might also increase. Companies and the
national government should consider promoting balance being
working and caregiving.
Being the child-in-law of a care recipient was negatively associ-
ated with psychological distress. A previous study in Japan showed
that the caregiver being a daughter-in-law was significantly related
to the caregiver not being depressed, which is consistent with the
current study.22 This finding might be explained as follows. First,
traditionally in Japan, women have long offered care to parents-
in-law as part of the Confucian ethic.23 However, with the imple-
mentation of long-term care insurance and as family structure has
changed from three-generation households to nuclear family
households, this tradition might be on its way out.23 Previous
studies reported that being a daughter-in-law was a risk factor for
discontinuing caregiving at home.8,24 Thus, compared with
spouses and children, children-in-law might be more likely to give
up caregiving earlier before becoming depressed, which might
have caused selection bias in the present study. Another possible
explanation is that children-in-law might have less psychological
distress, because they are more socially connected. In the post-
hoc analysis of the present study, children-in-law were more likely
to have someone to help with caregiving and to use social services.
However, other social interactions, such as neighborhood rela-
tionships25 and participation in social activities,26 which have also
been reported to be inversely associated with caregiver depression,
were not included in the present study and these could be
unmeasured confounders.
The finding that the prevalence of psychological distress is
higher in women and younger people is consistent with a previous
meta-analysis.9 However, as this corresponds with data from gen-
eral demographic studies, female sex and younger age might be
risk factors for psychological distress widely found among the
general population, and not specific to caregivers.19 We also found
© 2021 Japan Geriatrics Society.
K6 ≥ 5 P-value
n = 4821
n (%)
1362 (28.3) <0.001
647 (13.4) 0.016
210 (4.4) 0.050
336 (7.0) <0.001
869 (18.0) 0.841
518 (10.7) <0.001
229 (4.8) 0.893
423 (8.8) 0.200
959 (19.9) 0.758
150 (3.1) 0.836
1252 (26.0) 0.025
511 (10.6) 0.229
that longer times spent providing care and a caregiver’s own poor
health status were associated with psychological distress. These
results are in accordance with previous studies and unsurprising,
as caregiving is very physically demanding, which could lead to
psychological distress.6,8
We found that several diseases in care recipients were associ-
ated with caregiver’s depression. Caregiver burden for patients
with dementia has been investigated in many previous studies,27
and findings suggest that behavioral disturbance in patients was
the most important factor contributing to family caregiver bur-
den.8,22 As the population continues to age, it is becoming more
important to pay attention to caregivers of family members with
dementia. However, the present results suggest that it might also
be necessary to focus on caregivers of family members with lower
respiratory tract disease or diabetes, possibly due to higher risks of
exacerbations and hospital admissions, or the burden of managing
home oxygen therapy and insulin. Previous studies in other coun-
tries also reported a high prevalence of depression in caregivers of
chronic obstructive pulmonary disease patients,28 and that
patients with diabetes require more time for informal care than
those without, leading to a substantial caregiver burden.29
The strengths of the current study were the large representa-
tive sample of community-based family caregivers and their care
recipients, which allowed us to generalize the present findings to
community-based long-term care settings across Japan. This was
also the first study in Japan to examine psychological distress in
caregivers that focused on a wide range of caregiver- and care
recipient-related factors, including diseases in care recipients.
However, we also acknowledge several limitations. First, as this
was a cross-sectional study, the temporal relationship between
factors and the outcome variable (i.e. psychological distress) could
not be determined. Additionally, the study could not identify peo-
ple who might have stopped caregiving due to psychological dis-
tress, indicating a problem with sample selection. A longitudinal
study might be required to assess whether there are changes in
caregivers’ psychological distress over time and to identify vari-
ables that may influence changes in caregiving status. Second, the
survey did not include some relevant information, such as the use
of antidepressants and past history of psychiatric treatment, which
could potentially impact the prevalence of psychological distress.
Third, care recipients’ medical diagnoses were based on care
recipient or caregiver self-report. Therefore, misclassification of
these diagnoses is possible, which could weaken the observed
| 7
 Y Sun et al.

Table 3 Multivariable logistic regression analysis for Kessler’s Psychological Distress Scale (K6) scored ≥5

Variable Adjusted odds ratio (95% CI) P-value

Survey year
2007 1 (Reference)
2010 0.92 (0.81–1.05) 0.215
2013 1.01 (0.89–1.14) 0.881
2016 1.00 (0.89–1.13) 0.971
Caregiver characteristics
Age category (years)
20–54 1 (Reference)
55–64 0.65 (0.58–0.74) <0.001
65–74 0.54 (0.46–0.63) <0.001
≥75 0.50 (0.40–0.62) <0.001
Sex (female vs male) 1.35 (1.21–1.51) <0.001
Relationship to care recipient
Spouse 1 (Reference)
Child 0.98 (0.81–1.18) 0.851
Child-in-law 0.75 (0.61–0.92) 0.006
Other 0.87 (0.68–1.13) 0.293
Employment (employed vs 0.88 (0.80–0.97) 0.010
unemployed)
Household expenditure/month
<150 000 JPY 1 (Reference)
150 000–250 000 JPY 0.98 (0.88–1.09) 0.693
>250 000 JPY 0.99 (0.89–1.10) 0.873
Self-rated health status
Very good 1 (Reference)
Good 1.96 (1.61–2.38) <0.001
Fair 3.21 (2.71–3.79) <0.001
Not very good/poor 9.48 (7.91–11.37) <0.001
Use of social services (yes vs no) 0.99 (0.89–1.11) 0.905
Someone to help with care (yes vs no) 1.04 (0.96–1.14) 0.313
Care time spent per day
Help only when needed 1 (Reference)
2–3 h 1.13 (0.99–1.30) 0.070
About half the day 1.34 (1.17–1.54) <0.001
Almost all day 1.40 (1.25–1.58) <0.001
Other 1.15 (0.99–1.33) 0.064
Care recipient characteristics
Age category (years)
40–74 1 (Reference)
75–84 0.95 (0.82–1.09) 0.465
85–95 0.91 (0.77–1.07) 0.258
≥95 0.79 (0.61–1.01) 0.063
Sex (female vs male) 0.94 (0.84–1.05) 0.255
Care need level
Support levels 1–2 1 (Reference)
Care need levels 1–3 1.11 (1.00–1.24) 0.058
Care need levels 4–5 1.02 (0.88–1.18) 0.839
Disease (yes vs no)
Stroke 1.08 (0.98–1.20) 0.131
Dementia 1.16 (1.05–1.28) 0.004
Cardiac disease 1.06 (0.94–1.20) 0.352
Cancer 1.13 (0.92–1.39) 0.254
Lower respiratory tract disease 1.25 (1.06–1.49) 0.010
Joint disease 1.04 (0.93–1.16) 0.488
Diabetes 1.16 (1.00–1.33) 0.042
Parkinson’s disease 0.96 (0.79–1.16) 0.665
Vision or hearing impairment 0.96 (0.83–1.12) 0.596
Fractures 1.08 (0.97–1.20) 0.158

(Continues)

8 | © 2021 Japan Geriatrics Society.

 Psychological distress in caregivers
Table 3 Continued
Variable
Spinal cord injury
Senility
Other
associations. Fourth, because the study data were anonymized, we
could not determine whether the same caregiver-care recipient
dyads participated in more than one wave of cross-sectional sur-
veys. Ideally, the interperson correlation (if the same person par-
ticipated in the survey several times) should have been statistically
adjusted in our analysis, but we could not do so because of the
lack of a personal identifier in our data. Fifth, social interaction
factors other than the presence of other caregivers and the use of
social services were not collected in the survey. Future surveys
might need to include social interaction factors in more detail,
such as neighborhood relationships24 and participation in social
activities,25 which are probably associated with psychological dis-
tress in family caregivers.
To our knowledge, this is the first study of psychological dis-
tress in caregivers that focused on a wide range of factors pertinent
to both caregivers and their care recipients, using a large represen-
tative dataset in Japan. As there have been no national surveys or
average statistics on caregiver characteristics representing Japan,
the present research could contribute to policymaking and prac-
tice. Policymakers should disseminate information to health pro-
fessionals and social workers, so that they can recognize risk
factors for psychological distress in caregivers. It is also necessary
to establish a system that allows caregivers to undertake self-care
for themselves, such as educational materials and connection with
other caregivers.30 Health professionals and social workers need
to provide caregiver assessment and appropriate interventions,
especially for those at risk.30 Such efforts might help alleviate care-
givers’ psychological distress, ultimately resulting in benefits for
both caregivers and care recipients.
Acknowledgements
This research was supported by a grant-in-aid from the Ministry
of Health, Labour and Welfare; Health and Labor Sciences
Research Grant, Japan; Comprehensive Research on Aging and
Health (H30-choju-ippan-007). We thank Editage (www.editage.
jp) for English language editing.
Disclosure statement
The authors declare no conflicts of interest.
References
1 Statistics Bureau of Japan. Survey on time use and leisure activities.
[Cited 28 Oct 2020.] Available from URL: http://www.stat.go.jp/english/
data/shakai/2016/pdf/timeuse-a2016.pdf.
2 Tamiya N, Noguchi H, Nishi A et al. Population ageing and wellbeing:
lessons from Japan’s long-term care insurance policy. Lancet 2011; 378:
1183–1192.
3 Miyawaki A, Kobayashi Y, Noguchi H, Watanabe T, Takahashi H,
Tamiya N. Effect of reduced formal care availability on formal/informal
care patterns and caregiver health: a quasi-experimental study using the
Japanese long-term care insurance reform. BMC Geriatr 2020; 20: 207.
4 Zarit SH, Reever KE, Bach-Peterson J. Relatives of the impaired elderly:
correlates of feelings of burden. Gerontologist 1980; 20: 649–655.
© 2021 Japan Geriatrics Society.
Adjusted odds ratio (95% CI) P-value
0.91 (0.72–1.17) 0.473
1.07 (0.97–1.19) 0.194
1.05 (0.91–1.21) 0.492
5 Carretero S, Garcés J, R odenas F, Sanjosé V. The informal caregiver’s
burden of dependent people: theory and empirical review. Arch Gerontol
Geriatr 2009; 49: 74–79.
6 Kobayakawa M, Ogawa A, Konno M et al. Psychological and psychiatric
symptoms of terminally ill patients with cancer and their family care-
givers in the home-care setting: a nation-wide survey from the perspec-
tive of bereaved family members in Japan. J Psychosom Res 2017; 103:
127–132.
7 Saito T, Kondo N, Shiba K, Murata C, Kondo K. Income-based
inequalities in caregiving time and depressive symptoms among older
family caregivers under the Japanese long-term care insurance system: a
cross-sectional analysis. PLoS One 2018; 13: e0194919.
8 Kurasawa S, Yoshimasu K, Washio M et al. Factors influencing care-
givers’ burden among family caregivers and institutionalization of in-
home elderly people cared for by family caregivers. Environ Health Prev
Med 2012; 17: 474–483.
9 Pinquart M, Sörensen S. Gender differences in caregiver stressors,
social resources, and health: an updated meta-analysis. J Gerontol B
Psychol Sci Soc Sci 2006; 61: 33–45.
10 Ministry of Health, Labour and Welfare of Japan. Comprehensive sur-
vey of living conditions. [Cited 28 Oct 2020.] Available from URL:
https://www.mhlw.go.jp/english/database/db-hss/cslc.html.
11 Ministry of Health, Labour and Welfare of Japan. The overview for
comprehensive survey of living conditions (in Japanese). [Cited 28 Oct
2020.] Available from URL: https://www.mhlw.go.jp/toukei/list/20-
21kekka.html.
12 Ministry of Internal Affairs and Communications. Statistics act 2007.
[Cited 28 Oct 2020.] Available from URL: https://www.soumu.go.jp/
main_content/000626376.pdf.
13 Furukawa TA, Kawakami N, Saitoh M et al. The performance of the
Japanese version of the K6 and K10 in the world mental health survey
Japan. Int J Methods Psychiatr Res 2008; 17: 152–158.
14 Prochaska JJ, Sung HY, Max W, Shi Y, Ong M. Validity study of the K6
scale as a measure of moderate mental distress based on mental health
treatment need and utilization. Int J Methods Psychiatr Res 2012; 21: 88–97.
15 Fukuda Y, Nakao H, Imai H. Different income information as an indi-
cator for health inequality among Japanese adults. J Epidemiol 2007; 17:
93–99.
16 Idler EL, Benyamini Y. Self-related health and mortality: a review of
twenty-seven community studies. J Health Soc Behav 1997; 38: 21–37.
17 Tsutsui T, Muramatsu N. Care-needs certification in the long-term
care insurance system of Japan. J Am Geriatr Soc 2005; 53: 522–527.
18 Kikuzawa S. Elder care, multiple role involvement, and well-being
among middle-aged men and women in Japan. J Cross Cult Gerontol
2015; 30: 423–438.
19 Nishi D, Susukida R, Usuda K, Mojtabai R, Yamanouchi Y. Trends in
the prevalence of psychological distress and the use of mental health
services from 2007 to 2016 in Japan. J Affect Disord 2018; 239: 208–213.
20 Sugihara Y, Sugisawa H, Shibata H, Harada K. Productive roles, gender,
and depressive symptoms: evidence from a national longitudinal study
of late-middle-aged Japanese. J Gerontol B Psychol Sci Soc Sci 2008; 63:
227–234.
21 Japan Cabinet Office. Gender equality bureau. White paper on gender
equality: 2019 (in Japanese). [Cited 28 Oct 2020.] Available from URL:
http://www.gender.go.jp/about_danjo/whitepaper/r01/zentai/html/
honpen/b1_s03_02.html.
22 Washio M, Arai Y. Depression among caregivers of the disabled elderly
in southern Japan. Psychiatry Clin Neurosci 1999; 53: 407–412.
23 Arai Y, Zarit SH. Exploring strategies to alleviate caregiver burden:
effects of the national long-term care insurance scheme in Japan. Psy-
chogeriatrics 2011; 11: 183–189.
24 Arai Y, Washio M, Kudo K. Factors associated with admission to a geri-
atric hospital in semisuburban southern Japan. Psychiatry Clin Neurosci
2000; 54: 213–216.
25 Noguchi T, Nakagawa-Senda H, Tamai Y et al. Neighbourhood rela-
tionships moderate the positive association between family caregiver
burden and psychological distress in Japanese adults: a cross-sectional
study. Public Health 2020; 185: 80–86.
| 9
 Y Sun et al.

26 Oshio T, Kan M. How do social activities mitigate informal caregivers’ Appendix S2 Comparisons and multivariable logistic regression
psychological distress? Evidence from a nine-year panel survey in Japan. analysis for Kessler’s Psychological Distress Scale (K6) score ≥5 in
Health Qual Life Outcomes 2016; 14: 117.
27 Montgomery W, Goren A, Kahle-Wrobleski K, Nakamura T, Ueda K.
male caregivers
Alzheimer’s disease severity and its association with patient and care-
giver quality of life in Japan: results of a community-based survey. BMC Appendix S3 Comparisons and multivariable logistic regression
Geriatr 2018; 18: 141.
analysis for Kessler’s Psychological Distress Scale (K6) scores ≥5
28 Mi E, Mi E, Ewing G et al. Associations between the psychological
health of patients and carers in advanced COPD. Int J Chron Obstruct in female caregivers
Pulmon Dis 2017; 12: 2813–2821.
29 Lebrec J, Ascher-Svanum H, Chen YF et al. Effect of diabetes on care- Appendix S4 Characteristics of the study sample by caregiver’s
giver burden in an observational study of individuals with Alzheimer’s
disease. BMC Geriatr 2016; 16: 93.
relationship to the care recipient
30 Swartz K, Collins LG. Caregiver Care. Am Fam Physician 2019; 99:
699–706.

How to cite this article: Sun Y, Iwagami M,

Supporting Information
Watanabe T, et al. Factors associated with psychological

Additional supporting information may be found in the online distress in family caregivers: Findings from nationwide data
version of this article at the publisher’s website: in Japan. Geriatr. Gerontol. Int. 2021;1–10. https://doi.org/

Appendix S1 Comparisons and multivariable logistic regression 10.1111/ggi.14250

analysis for Kessler’s Psychological Distress Scale (K6) scores ≥13

10 | © 2021 Japan Geriatrics Society.