Lee et al.

BMC Psychiatry (2024) 24:485 BMC Psychiatry

https://doi.org/10.1186/s12888-024-05919-5

RESEARCH Open Access

Childhood adversity and suicidal ideation

in older Korean adults: unraveling the
mediating mechanisms of mental health,
physical health, and social relationships
Jin-kyung Lee1, Jinhee Lee2, Moo-Kwon Chung1,3, Taeksoo Shin4, Ji Young Park5, Kyoung-Joung Lee6, Hyo-Sang Lim7,
Sangwon Hwang8, Erdenebayar Urtnasan9, Yongmie Jo3 and Min-Hyuk Kim2*

Abstract
Background Suicide rates in older adults are much higher than those in younger age groups. Given the rapid
increase in the proportion of older adults in Korea and the high suicide rate of this age group, it is worth investigating
the mechanism of suicidal ideation for older adults. Generally, adverse childhood experiences are positively associated
with suicidal ideation; however, it is not fully understood what mediating relationships are linked to the association
between these experiences and current suicidal ideation.
Methods The data from 685 older Korean adults were analyzed utilizing logistic regression, path analyses, and
structural equation modeling. Based on our theoretical background and the empirical findings of previous research,
we examined three separate models with mental health, physical health, and social relationship mediators. After that,
we tested a combined model including all mediators. We also tested another combined model with mediation via
mental health moderated by physical health and social relationships.
Results The univariate logistic regression results indicated that childhood adversity was positively associated with
suicidal ideation in older adults. However, multivariate logistic regression results demonstrated that the direct effect
of childhood adversity became nonsignificant after accounting all variables. Three path models presented significant
mediation by depression and social support in the association between childhood adversity and suicidal ideation.
However, combined structural equation models demonstrated that only mediation by a latent variable of mental
health problems was statistically significant. Social relationships moderated the path from mental health problems to
suicidal ideation.
Conclusions Despite several limitations, this study has clinical implications for the development of effective
strategies to mitigate suicidal ideation. In particular, effectively screening the exposure to adverse childhood
experiences, early identification and treatment of depressive symptoms can play a crucial role in weakening the
association between childhood adversity and suicidal ideation in older adults.

*Correspondence:
Min-Hyuk Kim
mhkim09@yonsei.ac.kr
Full list of author information is available at the end of the article

© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use,
sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included
in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The
Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available
in this article, unless otherwise stated in a credit line to the data.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 2 of 14

Keywords Older adults, Suicidal ideation, Adverse childhood experiences, Mediation, Mental health, Physical health,
Social relationships

Introduction manner of suicidal ideation would provide useful infor-

Globally, 703,000 people die by suicide every year [1] and mation in creating effective and meaningful policies.
suicidal behaviors are higher in older adults than in other Thus far, whereas much of the research on adverse
age groups [2, 3]. In particular, South Korea showed the childhood experiences and suicidality has focused on
highest elderly suicide rate (54.8 per 100 000) among the adolescents and young adults, there is a growing body
members of the Organization for Economic Cooperation of research showing the association between adverse
and Development (OECD), and the rate was estimated childhood experiences and suicidal ideation throughout
to be approximately triple the OECD’s average (18.4 per an individual’s life [14–17]. Theoretically, the interper-
100 000). As the proportion of elderly people has risen sonal theory of suicide [18] explains that individuals with
sharply, suicide in the elderly, alongside its enormous adverse childhood experiences feel burdensome to their
economic burden, is expected to become a serious social family from an early developmental stage and struggle
issue [4]. Previous research has shown that suicidal desire to develop a strong sense of belonging with their parents
in older adults is more dangerous than in other popula- due to child maltreatment. Sachs-Ericsson et al. [14] pro-
tions. This population, who are more likely to experi- posed that adverse childhood experiences may increase
ence isolation, generally tends to retain positive attitudes the risk for suicide in late adulthood via several mecha-
toward life but once they are gripped by suicidal ideation, nisms, such as biological processes, psychiatric and
even the first suicidal behavior is likely to involve lethal health comorbidities, and psychosocial development.
methods, making them more likely to be found later and According to a recent white paper on suicide preven-
vulnerable to death [5]. Given the higher likelihood of tion in South Korea [19], suicide rates tend to increase
developing suicidal ideation and behaviors among older with age. Overall, the suicide rate in those 80 and older
adults, identifying risk factors and underlying mecha- was the highest at 61.3 per 100 000, 3 times the OECD
nisms is crucial to developing strategies to prevent sui- average in the corresponding age group. The suicide rate
cidal thoughts in older people. for ages 70–79 was the second highest at 41.8 per 100
It is well-known that adverse childhood experiences 000, 2.5 times higher than the corresponding OECD
significantly increase suicidality. Recent meta-analy- average statistics. The suicide rate for ages 60–69 was
ses reviewing international studies demonstrated that relatively lower at 28.4 per 100 000, 2.1 times higher
adverse childhood experiences have been associated than the corresponding statistics for the OECD [19].
with suicidality including suicidal ideation, attempts, and Given that there is a significant time between exposure to
events [6–11]. Academically, suicidal ideation is defined adverse childhood experiences and suicide in late adult-
as passive or active thoughts about death of oneself, and a hood, and that adverse childhood experiences influence
suicidal attempt indicates a self-injurious behavior based the overall developmental processes across an individu-
on suicidal ideation [12]. A suicidal event occurs when al’s life, it would be worthwhile to explore which media-
a self-injurious attempt with suicidal ideation worsens tors play a significant role in the relationship between
and causes an emergent situation [12]. As Van Orden et adverse childhood experiences and suicidal ideation.
al. [13] summarized, it is difficult to investigate suicidal Given that adverse childhood experiences influence sui-
attempts in a community sample due to the low rate of cidal ideation, the ideal solution would be the preven-
suicidal attempts among the whole population, the high tion of such experiences in the first place. However, the
possibility of exclusion from clinical trials due to safety occurrences of adverse childhood experiences are often
concerns, and the unavailability of psychological assess- uncontrollable, which signifies the existence of individu-
ments for those who died by suicide. However, although als at high risk for suicidality. Then, the primary concern
suicidal ideation is not always linked to lethal suicidal in an intervention policy should be on figuring out signif-
attempts, experiencing suicidal ideation significantly icant mediators and breaking the chain. To date, several
increases the risk of suicidal attempts and, in fact, all sui- previous studies have proposed that various factors could
cidal attempts are predicated on the existence of prior be mediators in the association between adverse child-
suicidal ideation. Among clinical cases, it is not difficult hood experiences and suicidality, ranging from mental
to see that persistent suicidal ideation increases the like- disorders (e.g. depression, anxiety) and physical health
lihood of recurrent suicidal attempts even when a prior problems (e.g. cancer, cerebrovascular disease) to social
attempt has been successfully halted. Thus, to effectively relationships (e.g. social support) [15, 16, 20–33]. How-
prevent a completed suicide, an investigation of one’s ever, these studies tend to focus on only a few mediating
variables fallen in one specific area rather than examining

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 3 of 14

them across mental, physical, social areas and compar- of thwarted belongingness. Poor social relationships
ing their effects to find the most significant mediators. To intensify the effects of social isolation, which increases
design efficient and effective policies to reduce suicide, thwarted belongingness as well. Poor social relationships
we need further research to explore which mediators increase the likelihood of a history of serious relational
are more relevant in the mechanism of suicidal ideation conflicts or experiences of neglect with others. Therefore,
amongst a high-risk population. Furthermore, while the others may perceive the task of meeting the aged popu-
suicide rate is highest among older adults and tends to lation’s basic needs as a serious burden. This increases
rise with age, previous studies have predominantly con- older adults’ perceived burdensomeness.
centrated on young or middle adulthood [15, 16, 20–33]. Taken together, this research aims to examine mental
Thus, the specific factors that mediate the pathway from health, physical health, and social relationships as media-
childhood adversity to suicidality in older adults remain tors of the positive association between adverse child-
incompletely understood. Given that risk factors of sui- hood experiences and suicidal ideation among older
cidal ideation can be different by age, an integrative adults. The four major hypotheses investigated in this
investigation to figure out significant mediators in older research are shown below:
adults would fill the knowledge gap and help to draw pol-
icy implications for effective interventions to reduce the Hypothesis 1.1 Older adults with adverse childhood
rate of suicide amongst older adults. experiences are likely to have suicidal ideation than those
Regarding the development of suicidal thoughts and without adverse childhood experiences.
behaviors in older adults, the interpersonal theory of
suicide [13, 18] provides a persuasive perspective with Hypothesis 1.2 Several characteristics such as depressive
numerous clinical examples. According to the theory [13, symptoms, anxiety, and binge drinking that affect mental
18], “thwarted belongingness” and “perceived burden- health mediate the association between adverse child-
someness” are two critical factors that significantly trig- hood experiences and suicidal ideation in older adults.
ger suicidal ideation. This theory especially focuses on
explaining the mechanism of suicidal behaviors, but it Hypothesis 1.3 Physical health problems such as car-
is invaluable for understanding the development of sui- diovascular disease or cancer mediate the association
cidal ideation. Although this theory highlights the mer- between adverse childhood experiences and suicidal ide-
its of an integrated view to enhance our understanding ation in older adults.
of suicidality, unfortunately, there has not been enough
empirical research conducted thus far considering all Hypothesis 1.4 Social relationships, including the range
these critical factors together. In explaining further of social networks and the degree of perceived social sup-
mechanisms between childhood adversity and suicidal port, mediate the association between adverse childhood
ideation in older adults, we will investigate three media- experiences and suicidal ideation in older adults.
tion models covering mental health, physical health, and
social relationship areas based on previous research [14, Methods
34–39]. Our attention will be given to variables con- Data
nected to thwarted belongingness and perceived bur- Data collected from 685 older Korean adults were ana-
densomeness among older adults. From our perspective, lyzed. Participants were required to be at least 55 years
mental health problems contribute to a distorted per- old and not suffer from substance use disorder, dementia,
ception which leads them to view their circumstances or serious illness in order to complete a face-to-face inter-
more negatively and increases thoughts such as believing view for 90 min. The participants were recruited from
one has no one to rely on (thwarted belongingness) and the large national cohort study, the Korean Genome and
everyone would be happier if one was gone (perceived Epidemiology Study-Cardiovascular Disease Association
burdensomeness). Experiencing physical health prob- Study (KoGES_CAVAS), via phone. The KoGES_CAVAS
lems requires expensive costs and presents many restric- was a multi-sited study with approximately 28,500 adults
tions on their independence, which in turn can increase in South Korea. Among the KoGES_CAVAS, 3,620 adults
their perceived burdensomeness. Also, restricted physi- lived in Wonju, a large local community mixed with
cal and financial independence as consequences of poor urban and rural areas. Out of 3,620 adults, 1,896 people
health conditions presents a barrier to their active par- changed or did not report available cell phone numbers.
ticipation in social activities as much as other healthy Out of 1,724 adults who reported available cell phone
individuals. Even when making social plans with friends, numbers, we recruited 685 participants after excluding
older adults with poor health conditions often experience 1,039 people who did not meet the inclusion criteria.
situations in which they have to cancel due to their dete- Different from KoGES_CAVAS, we designed new semi-
riorated health conditions. This can increase the sense structured interviews to collect data about older adults’

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 4 of 14

mental health from the participants in order to under- Next, anxiety was measured by the General Anxi-
stand the mechanisms of late-life depression [40]. After ety Disorder-7 screening tool (GAD-7) [43]. Under the
agreeing to participate in this semi-structured interview, prompt “Over the last two weeks, how often have you
the participants visited a campus once from December been bothered by the following problems?”, seven items
2020 to April 2021. After being provided an explanation including “Trouble relaxing” and “Worrying too much
of this study, they voluntarily signed the written consent about different things” were asked to each participant.
form. Subsequently, they had a 1:1 session with a trained Each item was scored from 0 for “Not at all” to 3 for
researcher to complete the interview. Data about partici- “Nearly every day”. A sum score was created using seven
pants’ demographic characteristics, close relationships, items (Cronbach’s alpha = 0.87).
and physical and mental health were collected during this Last, binge drinking was defined as consuming 5 or
interview. more drinks on an occasion based on the definition by
the Centers for Disease Control and Prevention [44]. It
Measures was coded as a binary variable (1 = binge drinking, 0 = no).
Childhood adversity
The severity of adverse childhood experiences was mea- Physical health problem mediators
sured by the short form of the Early Trauma Inventory The presence of six common physical health problems
[41]. This scale asked about each participant’s general was determined based on the report of ‘having been diag-
traumas, physical punishment, emotional abuse, and nosed by the doctor’ from participants. High blood pres-
sexual events before the age of 18. In this study, a total sure, hyperlipidemia, diabetes mellitus, cardiovascular
score was constructed from 27 binary items such as “Did disease, cerebrovascular accident, and cancer were coded
you ever witness violence towards others, including fam- as a binary variable to indicate if a participant had ever
ily members?”, “Were you ever slapped in the face with been diagnosed with the disease (1 = yes, 0 = no).
an open hand?”, and “Were you often put down or ridi-
culed?” (Cronbach’s alpha = 0.75). Social relationship mediators
Social networks and perceived social support were tested
Suicidal ideation as social relationship mediators. Social support was con-
Suicidal ideation in older adults was measured by the structed by averaging 12 items from the Multidimen-
last item of the Patient Health Questionnaire-9 (PHQ-9) sional Scale of Perceived Social Support [45](Cronbach’s
[42]. This item asked how often a participant had been alpha = 0.92). The items of social support included “There
bothered over the past two weeks by thoughts that they is a special person who is around when I am in need”,
would be better off dead or hurting themselves in some “I get the emotional help and support I need from my
way. Original responses ranged from “0 Not at all” to “3 family”, and “I can count on my friends when things go
Nearly every day”. When delineating suicidal ideation, the wrong”. Each item was scored from “1 Strongly disagree”
contrast between never experiencing a suicidal thought to “5 Strongly agree”.
and experiencing one is qualitatively distinct, unlike The variable of social networks was constructed by
variations in the frequency of suicidal thoughts. Original aggregating 18 items from the Lubben Social Network
responses indicating “Not at all” were recoded as 0 while Scale [46](Cronbach’s alpha = 0.85). The items of social
responses of “Several days”, “More than half the days”, and networks included “How many relatives do you see or
“Nearly every day” were all recoded as 1. hear from at least once a month?”, “How many friends
do you feel at ease with that you can talk about private
Mental health problem mediators matters?”, and “How many neighbors do you feel close
Depression, anxiety, and binge drinking were tested as to such that you could call on them for help?”. Each item
mediators related to mental health functioning. The was scored from 0 to 5 in which a higher score indicated
degree of depressive symptoms (PHQ-8) was constructed a wider social network and more frequent contacts with
by aggregating eight items (Cronbach’s alpha = 0.76) others.
from the PHQ-9 [42] with the exclusion of the last item
regarding suicidal ideation. Under the prompt “Over the Covariates
last two weeks, how often have you been bothered by any Sex was coded as 1 for males and 0 for females. Age was
of the following problems?”, eight items including “Little coded as a continuous variable with an ordinal scale. Par-
interest or pleasure in doing things” and “Feeling down, ticipants in their 50s were coded as 1, those in 60s were
depressed, or hopeless” were asked to each participant. coded as 2, those in 70s were coded as 3, and those in 80s
Each item ranged from 0 for “Not at all” to 3 for “Nearly were coded as 4. Education was also coded with an ordi-
every day”. nal scale. Participants who completed elementary school
or had a lower education level were coded as 1, those who

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 5 of 14

completed middle school or high school were coded as 2, within three separate models in adherence with previous
and those with a community college or higher education research. Covariates such as sex, age, education, house-
level were coded as 3. In terms of marital status, mar- hold income, and marital status were controlled for all
ried participants were coded as 1 and unmarried partici- endogenous variables. For mediation models, bootstrap-
pants were coded as 0. Household monthly income was ping with 5,000 iterations [48–50] was applied.
reported in the official currency of South Korea (Won). Since we are interested in comprehensively investigat-
Log transformation was applied for household income to ing the mediation mechanisms of suicidal ideation for
adjust for its high skewness. older adults, we also tested a combined model by inte-
grating all mediators across three models. For better vis-
Statistical analysis ibility, we present the combined model in which three
As a preliminary analysis, we first checked descriptive latent mediator variables (mental health problems, physi-
statistics including means, standard deviations, fre- cal health problems, and social relationships) were con-
quencies, and proportions for all variables. For group structed. Considering the insufficient factor loadings of
comparison by suicidal ideation, independent t-tests the six binary physical health problems, only the latent
and chi-square tests were performed. We also ran mul- variable of physical health problems was replaced with
ticollinearity test, but we did not find any problematic the observed continuous variable, indicating the number
issues (mean VIF = 1.31, minimum VIF = 1.04, maximum of physical health problems of a participant. Contrary to
VIF = 1.93). our expectation, binge drinking showed a non-signifi-
As main analyses, univariate and multivariate logistic cant path to explain suicidal ideation and a poor factor
regression models were tested to estimate crude effects loading of a latent variable of mental health problems.
and adjusted effects of the independent variables on sui- We excluded binge drinking as an indicator of the latent
cidal ideation. Given that the number of individuals with mediator in this combined model.
suicidal ideation is small in our data, we conducted sensi- After testing simple mediation paths in the combined
tivity analyses with the same set of all the variables in the model, we also explored the joint effects to see if the
multivariate logistic regression, and the results are pre- mediation by mental health problems differed by physi-
sented in the Appendix. cal health problems or social relationships. Thus, we
According to the literature [14, 34–39], three path changed two nonsignificant mediators of physical health
models regarding mental health, physical health, and problems and social relationships into moderators. We
social relationships were tested with different sets of added their interactions with mental health problems on
mediators. Based on the empirical evidence that depres- the path explaining suicidal ideation in this combined
sion, anxiety, and alcohol use disorders significantly model.
mediated the relationship between childhood adversity For model fit indices in three path models and two
and suicidal ideation [34], the first path model tested structural equation models, we reported frequently used
three mediators (depression, anxiety, binge drinking) that model fit indices [51–53]: Chi-Squares (χ2), root mean
were related to mental health dysfunctions. Given that square errors of approximation (RMSEA), comparative
childhood adversity is associated with the occurrence fit indexes (CFI), and standardized root mean square
of physical health problems [35, 47] and physical health residuals (SRMR). To indicate a good model fit, it is rec-
problems increase suicidal ideation in adulthood [36], ommended to have nonsignificant χ2 [51], RMSEA ≤ 0.07
the second path model consisted of six mediators (high [52], CFI ≥ 0.95 [53], and SRMR ≤ 0.08 [53]. STATA 17.0
blood pressure, hyperlipidemia, diabetes mellitus, cardio- for descriptive statistics and logistic regressions, Python
vascular disease, cerebrovascular accident, and cancer) for sensitivity analyses, and Mplus 8.9 for path and struc-
that were common physical health problems among older tural equation models were used.
Korean adults. Based on the previous research demon-
strating that social relationships significantly mediated Results
the association between childhood adversity and suicidal Descriptive statistics
ideation [20, 38, 39], the third path model investigated Table 1 indicates the descriptive statistics of the sample.
the role of two mediators (social support, social net- Out of 685 older adults, 31 people (4.53%) reported that
work) that represented the quality and quantity of social they had experienced suicidal ideation during the worst
relationships. two-week period in the past month. On average, people
All three mediation models had childhood adversity with suicidal ideation tended to report higher scores in
as an exogenous variable and suicidal ideation as the last childhood adversity, depressive symptoms, and anxiety
endogenous variable. To make it easy to compare this than others. There was little difference in binge drink-
research with other research investigating mediators, we ing between people with vs. without suicidal ideation.
tested mediation paths through each mediator variable Among the whole sample, approximately 17% reported

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 6 of 14

Table 1 Descriptive statistics

Total With suicidal ideation Without suicidal ideation Independent t-test (t) or chi-square test (χ2)
(N = 685) (n = 31) (n = 654)
Suicidal ideation N (%) N (%) N (%)
Yes 31 (04.53%) 31 (100.0%) -
No 654 (95.47%) - 654 (100.0%)
M (SD) M (SD) M (SD)
Adverse childhood experiences 2.85 (2.95) 4.32 (3.49) 2.78 (2.90) t = -2.87**
Depressive symptoms 0.33 (0.42) 1.07 (0.68) 0.29 (0.37) t = -10.81***
Anxiety 1.38 (2.76) 6.03 (5.19) 1.16 (2.38) t = -10.33***
Binge drinking N (%) N (%) N (%) t = 0.14
Yes 117 (17.08%) 5 (16.13%) 112 (17.13%)
No 568 (82.92%) 26 (83.87%) 542 (82.87%)
High blood pressure N (%) N (%) N (%) t = -1.25
Yes 301 (43.94%) 17 (54.84%) 284 (43.43%)
No 384 (56.06%) 14 (45.16%) 370 (56.57%)
Hyperlipidemia N (%) N (%) N (%) t = -3.77***
Yes 267 (38.98%) 22 (70.97%) 245 (37.46%)
No 418 (61.02%) 9 (29.03%) 409 (62.54%)
Diabetes mellitus N (%) N (%) N (%) t = -3.05**
Yes 140 (20.44%) 13 (41.94%) 127 (19.42%)
No 545 (79.56%) 18 (58.06%) 527 (80.58%)
Cardiovascular disease N (%) N (%) N (%) t = -2.73**
Yes 89 (12.99%) 9 (29.03%) 80 (12.23%)
No 596 (87.01%) 22 (70.97%) 574 (87.77%)
Cerebrovascular accident N (%) N (%) N (%) t = -2.71**
Yes 37 (05.40%) 5 (16.13%) 32 (04.89%)
No 648 (94.60%) 26 (83.87%) 622 (95.11%)
Cancer N (%) N (%) N (%) t = -0.44
Yes 72 (10.51%) 4 (12.90%) 68 (10.40%)
No 613 (89.49%) 27 (87.10%) 586 (89.60%)
M (SD) M (SD) M (SD)
Social support 3.92 (0.70) 3.31 (0.93) 3.95 (0.68) t = 5.02***
Social network 42.76 (13.18) 35.97 (14.61) 43.09 (13.04) t = 2.95**
Sex N (%) N (%) N (%) t = 1.16
Male 290 (42.34%) 10 (32.26%) 280 (42.81%)
Female 395 (57.66%) 21 (67.74%) 374 (57.19%)
Age N (%) N (%) N (%) χ2 = 4.96
50s 68 (09.93%) 2 (06.45%) 66 (10.09%)
60s 317 (46.28%) 12 (38.71%) 305 (46.64%)
70s 240 (35.04%) 11 (35.48%) 229 (35.02%)
80s 60 (08.76%) 6 (19.35%) 54 (08.26%)
Education N (%) N (%) N (%) χ2 = 10.64**
≤ Elementary 200 (29.20%) 17 (54.84%) 183 (27.98%)
Middle or high 319 (46.57%) 8 (25.81%) 311 (47.55%)
≥ College (2-3yrs or 4yrs) 166 (24.23%) 6 (19.35%) 160 (24.46%)
Marital status N (%) N (%) N (%) t = 2.81**
Married 612 (89.34%) 23 (74.19%) 589 (90.06%)
Unmarried 73 (10.66%) 8 (25.81%) 65 (09.94%)
M (SD) M (SD) M (SD)
Household income 2,912,381 2,141,290 2,948,931 t = 1.92
(Won; monthly) (2,288,407) (1,645,286) (2,309,001)
*
p < .05, **p < .01, ***p < .001

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 7 of 14

consuming 5 or more cups of alcohol in one setting. 301 (Crude OR = 2.94, p < .01), and cerebrovascular accident
participants (43.94%) reported they had been diagnosed (Crude OR = 3.74, p < .05) to be positively associated with
with high blood pressure, and 267 participants (38.98%) suicidal ideation. As expected, social support (Crude
reported they had ever had hyperlipidemia. 140 partici- OR = 0.36, p < .001) and social network (Crude OR = 0.96,
pants (20.44%) reported they had experienced diabetes p < .01) were negatively associated with suicidal ide-
mellitus, and 89 participants (12.99%) reported they had ation. However, in the adjusted model after accounting
suffered from cardiovascular disease. 72 participants for all independent variables, only the direct effects of
(10.51%) reported they had been diagnosed with cancer, depressive symptoms (adjusted OR = 4.91, p < .01) and
and 37 participants (5.40%) reported they had experi- anxiety (adjusted OR = 1.14, p < .05) among key variables
enced cerebrovascular accidents. Older adults with sui- remained statistically significant in explaining suicidal
cidal ideation showed higher incidence rates in all six of ideation.
these health issues compared to those without suicidal
ideation. However, older adults with suicidal ideation Path model with mental health problem mediators
reported lower average scores in social networks and The results from the first path model with mental health
perceived social support than those without suicidal ide- mediators are shown in Fig. 1. This was a saturated
ation. In the group with suicidal ideation, the proportions model, so the model fit indices were χ2(df) = 718.04(30),
of female, elderly, and lower education level individu- p = .000, RMSEA = 0.000, CFI = 1.000, and SRMR = 0.002.
als were higher than those in the group without suicidal When mediation paths by depression, anxiety, and binge
ideation. Additionally, the group with suicidal ideation drinking were tested, the results demonstrated that child-
reported lower household income than the group with- hood adversity was positively associated with depression
out suicidal ideation. (β = 0.22, p < .001), and depression was positively associ-
ated with suicidal ideation (β = 0.34, p < .001). The indirect
Logistic regression predicting suicidal ideation effect of childhood adversity via depression was statisti-
Table 2 indicates the crude and adjusted effects of all cally significant in explaining suicidal ideation (β = 0.07,
independent variables on suicidal ideation. Examina- p < .01). In terms of anxiety, childhood adversity was
tion of the crude effects revealed childhood adversity significantly associated with anxiety in late adulthood
(Crude OR = 1.15, p < .01), depressive symptoms (Crude (β = 0.30, p < .001) but anxiety was not significantly asso-
OR = 9.71, p < .001), anxiety (Crude OR = 1.32, p < .001), ciated with suicidal ideation. Different from our hypoth-
hyperlipidemia (Crude OR = 4.08, p < .001), diabetes mel- esis, childhood adversity was not statistically significant
litus (Crude OR = 3.00, p < .01), cardiovascular disease in predicting binge drinking, and binge drinking was
not significantly associated with suicidal ideation. When
Table 2 Crude and adjusted effects on suicidal ideation including all these mediation paths, the direct effect of
Y = Suicidal ideation childhood adversity became statistically nonsignificant in
Crude OR Adj. OR explaining suicidal ideation.
(95% CI) (95% CI)
Adverse childhood experiences 1.15** (1.04, 1.26) 1.03 (0.90, 1.18)
Path model with physical health problem mediators
Depressive symptoms 9.71*** (5.22, 18.06) 4.91** (1.83, 13.13)
The results from the second path model with physical
Anxiety 1.32*** (1.22, 1.43) 1.14* (1.00, 1.30)
health mediators are given in Fig. 2. As expected, child-
Binge drinking 0.93 (0.35, 2.48) 1.07 (0.28, 4.07)
High blood pressure 1.58 (0.77, 3.26) 0.93 (0.35, 2.43)
hood adversity was positively associated with suicidal
Hyperlipidemia 4.08*** (1.85, 9.01) 2.38 (0.89, 6.33)
ideation (β = 0.22, p < .001). Furthermore, older adults
Diabetes mellitus 3.00** (1.43, 6.28) 1.68 (0.66, 4.29) who had hyperlipidemia (β = 0.24, p < .01) and those with
Cardiovascular disease 2.94** (1.31, 6.60) 1.21 (0.41, 3.53) diabetes mellitus (β = 0.19, p < .01) were more likely to
Cerebrovascular accident 3.74* (1.35, 10.38) 2.77 (0.73, 10.48) experience suicidal ideation. However, childhood adver-
Cancer 1.28 (0.43, 3.76) 1.61 (0.47, 5.57) sity was not significantly associated with any of the six
Social support 0.36*** (0.24, 0.56) 1.03 (0.53, 2.02) common physical health problems (high blood pressure,
Social network 0.96** (0.93, 0.99) 0.97 (0.94, 1.02) hyperlipidemia, diabetes mellitus, cardiovascular disease,
Age 1.52 (0.97, 2.39) 1.65 (0.88, 3.12) cerebrovascular accident, and cancer). Therefore, none of
Male 0.64 (0.29, 1.37) 1.02 (0.33, 3.12) the physical health problems played an important medi-
Education 0.53* (0.31, 0.90) 0.50* (0.26, 0.96) ating role in explaining the relationship between child-
Married 0.32** (0.14, 0.74) 0.58 (0.18, 1.89) hood adversity and suicidal ideation. In this model, CFI
Household income 0.97 (0.83, 1.15) 1.24 (0.88, 1.74) had a poor fit but other model fit indices were acceptable
Constant - 0.00* (0.00, 0.25) (χ2(df) = 94.39(63), p = .006, RMSEA = 0.027, CFI = 0.762,
R2 - 0.341 and SRMR = 0.047).
*
p < .05, **p < .01, ***p < .001

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 8 of 14

Fig. 1 Path results with mental health problem mediators

*
p < .05, **p < .01, ***p < .001
Note: A black line indicates a statistically significant path (p < .05) while a gray line indicates a nonsignificant path. Covariates such as age, sex, education,
marital status, and household income were controlled for all endogenous variables

Fig. 2 Path results with physical health problem mediators

*
p < .05, **p < .01, ***p < .001
Note: A black line indicates a statistically significant path (p < .05) while a gray line indicates a nonsignificant path. Covariates such as age, sex, education,
marital status, and household income were controlled for all endogenous variables

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 9 of 14

Path model with social relationship mediators indices for this combined model with three mediators
The results from the third path model with social rela- indicated a good model fit (RMSEA = 0.064, CFI = 0.948,
tionship mediators are presented in Fig. 3. Model SRMR = 0.045).
fit indices indicated that this was a saturated model In the combined model, only the latent mediator
(χ2(df) = 280.97(21), p = .000, RMSEA = 0.000, CFI = 1.000, variable of mental health problems was significantly
and SRMR = 0.000). As anticipated, childhood adver- explained by childhood adversity (ACE) and did sig-
sity was negatively associated with social support (β = nificantly explain suicidal ideation (SI) in late adulthood
− 0.20, p < .001) and social support was negatively associ- (ACE → mental health problems: β = 0.33, p < .001; men-
ated with suicidal ideation (β = − 0.21, p < .05). The indi- tal health problems → SI: β = 0.44, p < .001). The observed
rect effect of childhood adversity on suicidal ideation continuous mediator of the number of physical health
via social support was statistically significant (β = 0.04, problems was significantly associated with suicidal ide-
p < .05). However, although childhood adversity was sig- ation, but it was not significantly explained by childhood
nificantly associated with social networks (β = − 0.13, adversity (ACE → physical health problems: β = 0.05,
p < .01), social network was not significantly associated p = .238; physical health problems → SI: β = 0.26, p < .001).
with suicidal ideation. This demonstrated that the qual- On the contrary, childhood adversity was significantly
ity of perceived social support was more important than associated with the latent mediator variable of social rela-
the quantity of social networks in explaining suicidal tionships, but this latent social relationship variable was
ideation. statistically nonsignificant in explaining suicidal ideation
in late adulthood (ACE → social relationships: β = − 0.21,
Combined structural equation models p < .001; social relationships → SI: β = − 0.12, p = .129).
In addition, a structural equation model was tested Next, we also tested moderated mediation paths in
after constructing a latent mediator variable with corre- a structural equation model by changing two nonsig-
sponding observed variables for each model (i.e. mental nificant mediator variables representing physical health
health model, physical health model, and social relation- problems and social relationships into moderator vari-
ship model) and including all these three latent variables ables. We added their interactions with mental health
within the model. As mentioned in the Methods section, (i.e., mental health problems x physical health problems,
due to poor factor loadings, binge drinking was excluded mental health problems x social relationships) on the
from the latent variable of mental health characteris- path predicting suicidal ideation in the combined model.
tics. Also, the latent variable of physical health media- By doing this, we were interested to see if the mediation
tors was replaced with the continuous variable of the via mental health differed by physical health problems
number of physical illnesses. All covariates such as age, or social relationships. In this latent interaction model,
gender, education, marital status, and household income social relationships significantly moderated the effects
were controlled for physical health, mental health, and of mental health problems on suicidal ideation (men-
social relationship mediators and the outcome variable tal health problems x social relationships → SI: β = 0.12,
(suicidal ideation) in this combined model. Model fit p < .01) while the interaction between mental health

Fig. 3 Path results with social relationship mediators

*
p < .05, **p < .01, ***p < .001
Note: A black line indicates a statistically significant path (p < .05) while a gray line indicates a nonsignificant path. Covariates such as age, sex, education,
marital status, and household income were controlled for all endogenous variables

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 10 of 14

problems and physical health problems was not statisti- implication of these results is that screening the expo-
cally significant. The significant interaction plot between sure of adverse childhood experiences should be a crucial
two latent variables (mental health problems x social component of effective strategies for suicide prevention.
relationships) was presented in Fig. 4. As shown in the recent scoping review [54], more discus-
sion is needed about which measures or what aspects of
Discussion adverse childhood experiences are more relevant to pre-
This study investigates the direct and indirect associa- venting suicidality. Previous research has discussed that
tions between childhood adversity and suicidal ideation it is difficult to separate and estimate the unique effect of
in community-dwelling older adults in South Korea. The each specific type of adverse childhood experiences since
results of the present study demonstrate that older adults child sexual, physical, emotional, and general abuse often
with adverse childhood experiences were more likely to co-occur within a household over a long period [14]. In
experience suicidal ideation. However, mental health this study, the types of adverse childhood experiences
problems, especially depression explained the path from were quantified for each participant. This is one viable
childhood adversity and suicidal ideation in late adult- approach.
hood. Social support seemed to mediate the association On the other hand, the findings from our combined
between childhood adversity and suicidal ideation, but model that the direct effect of childhood adversity
in fact, social relationships did not mediate this path but became nonsignificant after including mediation paths
rather moderated the effect of mental health problems on instill hope. In other terms, if we successfully intervene
suicidal ideation, intensifying the effect of mental health to break the mediation chain, the risk of suicidal ide-
problems. ation caused by childhood adversity would be signifi-
cantly reduced. Further research is needed to fully reveal
Direct association between childhood adversity and the mechanism of how suicidal ideation is influenced by
suicidal ideation childhood adversity, particularly among those who were
Our findings support the research hypothesis that child- exposed to greater adverse childhood experiences. To
hood adversity increases the risk of suicidal ideation comprehensively investigate the mechanism from child-
in late adulthood. This is in line with the interpersonal hood adversity to suicidal ideation in late adulthood,
theory of suicide [13]. Exposure to adverse childhood future research requires longitudinal data collection from
experiences is a significant risk factor of suicidal ide- a larger representative sample.
ation because people who were previously exposed to
more adverse childhood experiences are likely to be more Indirect association between childhood adversity and
sensitive to developing feelings of burdensomeness and suicidal ideation through mental health mediators
thwarted belongingness in a given situation [13, 18]. Our This study supports the positive association between
findings demonstrate that childhood adversity not only adverse childhood experiences and suicidal ideation
triggers the development of suicidal ideation at an earlier through mental health mediators. Previous studies using
developmental period in one’s life, but also increases the a variety of methods (including longitudinal and cross-
likelihood of suicidal ideation later in life. One clinical sectional studies) have demonstrated that depression and

Fig. 4 Moderation by social relationships on the path from mental health problems to suicidal ideation

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 11 of 14

anxiety have a mediating role in the relationship between in the combined structural equation model, older adults
adverse childhood experiences and suicidality [25–27, with more health problems were more likely to expe-
29]. Our results partially supported it by demonstrating rience suicidal ideation. This implies that the sever-
that depression is a significant mediator in this relation- ity of physical health problems would be more relevant
ship. Our findings suggest that early identification and to suicidal ideation in late adulthood than the type of
treatment of depression is very important for the preven- physical health problems. Also, we asked about the his-
tion of suicidal ideation in older Korean adults. tory of physical health problems but the current health
In terms of anxiety, the results of the present study condition would be more related to suicidal ideation. In
demonstrate that the positive association between child- our data, the discrepancy between the correlations of
hood adversity and anxiety was statistically significant, the two chronic diseases and other physical health prob-
but the association between anxiety and suicidal ideation lems concerning suicidal ideation can be explained by the
was weaker than we anticipated. Given the high factor social history of physical health problems, which differs
loading of anxiety to construct the latent mental health based on the type of disease. For example, the diagnosis
variable, and the latent variable’s significant mediation in of hyperlipidemia and diabetes mellitus was associated
the combined model, we interpret that heightened anxi- with a high risk of feeling like a burden to one’s family.
ety influenced by childhood adversity would be a con- However, in the case of cancer, our participants who have
textual factor (i.e. worsening the risk of suicidal ideation survived cancer were enthusiastic to open their diagno-
with depressive symptoms) rather than directly increas- sis and their successful treatment, showing an increase in
ing suicidal ideation. Considering the high association feelings of gratitude in their lives.
between childhood adversity and anxiety in late adult- One interesting finding is that childhood adversity was
hood, early diagnosis and clinical treatments for anxiety not significantly associated with any history of physical
disorder would be effective for those who had adverse health problems (high blood pressure, hyperlipidemia,
childhood experiences. diabetes mellitus, cardiovascular disease, cerebrovas-
Contrary to popular belief, binge drinking was not sig- cular accident, and cancer). Given that adverse child-
nificantly associated with adverse childhood experiences hood experiences influence physical health problems
or suicidal ideation. Most previous research has reported and physical health problems are risk factors for suicide
that adverse childhood experiences are associated with among older adults in previous studies [10, 14, 57], we
alcohol abuse later in life, including alcohol dependence hypothesized that these chronic diseases would be pos-
or binge drinking, which affects suicidal behaviors [10, sible mediators in the path from adverse childhood expe-
15, 22, 25, 27, 55]. However, in our study, binge drinking riences to suicidal ideation. However, even the number
was not associated with either adverse childhood expe- of physical health problems was not predicted by child-
riences nor with suicidal ideation. This is in line with hood adversity. Our results suggest that the effects of
another previous research reporting nonsignificant medi- childhood adversity on physical health might involve
ation by alcohol misuse [38]. One thing we were con- more complex and individual life factors. Previous stud-
cerned was that our participants, especially men, might ies have suggested various mechanisms: For example,
have reported less alcohol use than the general Korean the effects of childhood adversity on physical health
population [56]. It is not certain whether our participants can vary depending on the number or types of adverse
consume less alcohol than average or if the participants childhood experiences [10, 11]. It is also likely that the
underreported their consumption, and this could possi- relationships between childhood adversity and the devel-
bly have affected our conclusion. Further study is needed opment of chronic diseases differ by sex, disability, sever-
to confirm the role of alcohol use problems in the con- ity of pain, or comorbid depression and anxiety [57].
nections between adverse childhood experiences to sui- Altogether, these factors would make the relationships
cidal ideation. between childhood adversity and physical health prob-
lems more complex. Future research is recommended to
Indirect association between childhood adversity and delve into the roles of these various factors and physical
suicidal ideation through physical health mediators health problems between childhood adversity and sui-
Our research hypothesis that physical health problems cidal ideation.
would significantly increase suicidal ideation in older
adults is partially supported in this study. When we Roles of social relationships in the association between
tested separate mediation paths by six common physical childhood adversity and suicidal ideation
health problems in the path model, only hyperlipidemia The results from the path model with social relationship
and diabetes mellitus were significantly positively asso- mediators demonstrate that perceived social support
ciated with suicidal ideation. However, if we tested the mediated the relationship between childhood adver-
mediation by the number of physical health problems sity and suicidal ideation in late adulthood, but social

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 12 of 14

networks did not. Specifically, more exposure to child- of causal relationships. Second, most of the data was
hood adversity was associated with poorer social sup- measured retrospectively which inherently contains some
port and weaker social networks. However, weaker social amount of recall or social desirability bias. Although we
networks were not significantly associated with higher interviewed each participant in private to prevent this
suicidal ideation when poorer social support was. This bias, it still may not be wholly avoidable. Third, we used
highlights the importance of social support as a mediator one item of the PHQ-9 as the outcome variable, sui-
between childhood adversity and suicidal ideation [25, cidal ideation. This question about suicidal ideation cap-
28, 33]. Future research is recommended to investigate tures the short-term suicidal desire of participants at the
if the underlying mechanism by which social functioning moment of data collection. Given that suicidal ideation
influences suicidal ideation in older adults with adverse tends to fluctuate over time, this may affect the extent
childhood experiences is different from the mechanism of the association, but we were not able to capture that
in those without adverse childhood experiences. If the fully. Fourth, our study had a small sample size, which
underlying mechanism is different, more tailored meth- prevented us from further subgroup analysis stratified
ods of intervention focused on social relationships would by gender, adverse childhood experiences, or subscales
need to be developed. of social networks. Further study utilizing longitudinal
Meanwhile, the new findings from our combined data with a larger sample size is recommended. Lastly,
models suggest an interesting possibility that social rela- there might be some other variables that have not been
tionships play a significant role as a moderator in the considered in this study but can influence the association
mediation path from childhood adversity to suicidal between childhood adversity and suicidal ideation.
ideation through mental health problems. Despite dif-
ferent significance in mediation effects in the basic path Conclusions
model, high factor loadings of perceived social support Despite several limitations, this study contributes to
and social network in the combined structural equa- broader research on the mechanisms of suicidal ideation
tion model reflected the fact that the directions of social in older adults. Our data from a community sample indi-
support and social network are basically similar. The cates that more exposure to childhood adversity is associ-
combined structural equation model demonstrated that ated with suicidal ideation even in late adulthood. Based
greater childhood adversity was significantly associated on previous research, we tested three different mediation
with poorer social relationships, but poorer social func- models to explain the association between childhood
tioning was not directly related to heightened suicidal adversity and suicidal ideation. The major findings of
ideation in late adulthood. These are different from the this study support that mental health problems such as
previous research [28, 31] which demonstrated the sig- depressive symptoms did significantly mediate the effects
nificant mediation paths by poorer social functioning in of childhood adversity on suicidal ideation in late adult-
the relationship between childhood adversity suicidality hood. Although it is unclear the relationship between
among adolescents and young adults. In our combined childhood adversity and physical health problems, the
model, the joint effect between mental health problems association between physical health problems and sui-
and social relationships was noticeable. When a partici- cidal ideation in late adulthood is supported. The asso-
pant had a low risk of mental health problems, a higher ciation between social relationships and suicidal ideation
level of social functioning was significantly associated was weaker than we anticipated, but childhood adversity
with lower suicidal ideation. However, the role of social was still significant in explaining social relationships in
relationships as a protective factor significantly disap- late adulthood. In this study, the role of social relation-
peared when a participant experienced a high level of ships among older adults seems to be different from that
mental health problems. These results support that the of younger people, which requires further research. We
functions of social relationships among older adults believe this study makes a meaningful contribution in
would be different from younger populations. Based on revealing the mechanisms regarding suicidal ideation in
the findings, we anticipate the effects of protective fac- older adults. This study has clinical implications for the
tors would be much lower among older adults with a high development of effective strategies to prevent suicidal
risk. Further research is required to investigate the role of ideation among older adults in South Korea. Most nota-
social relationships, particularly in high-risk older adults bly, screening the degree of adverse childhood experi-
experiencing serious mental health problems. ences, early identification or treatment of mental health
problems, and obtaining more information of social
Limitations relationships in late adulthood can play a critical role in
The present study has some limitations. First, this is a breaking the association between childhood adversity
cross-sectional study; therefore, all associations reported and suicidal ideation in older adults.
in this study should be interpreted as tendencies instead

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 13 of 14

References
Supplementary Information 1. WHO. Suicide Worldwide in 2019: Global Health estimates. Geneva, Switzer-
The online version contains supplementary material available at https://doi. land: World Health Organization; 2021.
org/10.1186/s12888-024-05919-5. 2. De Leo D. Late-life suicide in an aging world. Nat Aging. 2022;2:7–12.
3. De Leo D, Giannotti AV. Suicide in late life: a viewpoint. Prev Med.
Supplementary Material 1 2021;152:e106735.
4. Kim H, Ahn JS, Kim H, Cha YS, Lee J, Kim MH, Min S. Sociodemographic
and clinical characteristics of old-old suicide attempters compared
Acknowledgements with young-old and middle-aged attempters. Int J Geriatr Psychiatry.
The authors appreciate all the participants of our research project for their 2018;33(12):1717–26.
open hearts and the precious time they willingly shared with us for research 5. Ding OJ, Kennedy GJ. Understanding vulnerability to late-life suicide. Curr
purposes. The authors acknowledge that it would be impossible to obtain Psychiatry Rep. 2021;23:e58.
this data without many research assistants working hard in the data collection 6. Angelakis I, Austin JL, Gooding P. Association of childhood maltreatment
and cleaning process. Also, a lot of administrative help from the Institute for with suicide behaviors among young people: a systematic review and meta-
Poverty Alleviation at Yonsei University is greatly appreciated. analysis. JAMA Netw Open. 2020;3(8):e2012563.
7. Angelakis I, Gillespie EL, Panagioti M. Childhood maltreatment and adult
Author contributions suicidality: a comprehensive systematic review with meta-analysis. Psychol
M-HK initiated this study and supervised the development of the manuscript. Med. 2019;49(7):1057–78.
M-HK and J-kL wrote the manuscript. J-kL analyzed the data. JL supported 8. Castellvi P, Miranda-Mendizabal A, Pares-Badell O, Almenara J, Alonso I, Blasco
the development of the manuscript. M-KC is the principal investigator of MJ, Cebria A, Gabilondo A, Gili M, Lagares C, et al. Exposure to violence, a risk
the overall research project. TS and JP contributed to designing the semi- for suicide in youths and young adults. A meta-analysis of longitudinal stud-
structured questionnaire used in this study. K-JL, H-SL, SH, and EU contributes ies. Acta Psychiatr Scand. 2017;135(3):195–211.
to developing a prediction algorithm for late-life depression using wearable 9. Baldwin JR, Wang B, Karwatowska L, Schoeler T, Tsaligopoulou A, Munafo MR,
devices. YJ managed the data collection process at the baseline survey. All Pingault JB. Childhood maltreatment and mental health problems: a system-
authors discussed the results and contributed to the final manuscript. atic review and meta-analysis of quasi-experimental studies. Am J Psychiatry.
2023;180(2):117–26.
Funding 10. Grummitt LR, Kreski NT, Kim SG, Platt J, Keyes KM, McLaughlin KA. Association
This work was supported by the Ministry of Education of the Republic of Korea of childhood adversity with morbidity and mortality in US adults: a system-
and the National Research Foundation of Korea (NRF-2020S1A5A2A03045088). atic review. JAMA Pediatr. 2021;175(12):1269–78.
11. Nelson CA, Scott RD, Bhutta ZA, Harris NB, Danese A, Samara M. Adversity
Data availability in childhood is linked to mental and physical health throughout life. BMJ.
The datasets analyzed during the current study are not publicly available due 2020;371:m3048.
to participants’ confidentiality but are available from the corresponding author 12. Turecki G, Brent DA. Suicide and suicidal behaviour. Lancet.
on reasonable request. 2016;387(10024):1227–39.
13. Van Orden KA, Witte TK, Cukrowicz KC, Braithwaite S, Selby EA, Joiner TE. The
interpersonal theory of suicide. Psychol Rev. 2010;117(2):575–600.
Declarations 14. Sachs-Ericsson NJ, Rushing NC, Stanley IH, Sheffler J. In my end is my begin-
ning: developmental trajectories of adverse childhood experiences to late-
Ethics approval and consent to participate
life suicide. Aging Ment Health. 2016;20(2):139–65.
This study protocol and amendments involving human participants were
15. Sachs-Ericsson N, Corsentino E, Rushing NC, Sheffler J. Early childhood abuse
reviewed and approved by the Institutional Review Board at the Yonsei
and late-life suicidal ideation. Aging Ment Health. 2013;17(4):489–94.
University Mirae Campus (1041849-202111-SB-180-06). All participants
16. Jardim GBG, Novelo M, Spanemberg L, von Gunten A, Engroff P, Nogueira
provided their written informed consent before participation in the study.
EL, Neto AC. Influence of childhood abuse and neglect subtypes on late-life
suicide risk beyond depression. Child Abuse Negl. 2018;80:249–56.
Consent for publication
17. Kwak M, Ahn S. Childhood adversity predicted suicidal ideation in
Not applicable.
older age: results from a national survey in Korea. Aging Ment Health.
2020;24(7):1141–8.
Competing interests
18. Joiner T. Why people die by suicide. Cambridge, MA: Harvard University Press;
The authors declare no competing interests.
2005.
19. Ministry of Health and Welfare, Korea Foundation for Suicide Prevention.
Author details
1 White paper on suicide prevention. South Korea: Ministry of Health and
Institute for Poverty Alleviation and International Development, Yonsei
Welfare, Korea Foundation for Suicide Prevention; 2023.
University, Mirae Campus, Wonju, Republic of Korea
2 20. Kobrinsky V, Siedlecki KL. Mediators of the relationship between adverse
Department of Psychiatry, Wonju College of Medicine, Yonsei University,
childhood experiences (ACEs) and symptoms of anxiety, depression, and
Wonju, Republic of Korea
3 suicidality among adults. J Child Adolesc Trauma. 2022;16:233–46.
Department of Global Public Administration, Yonsei University, Mirae
21. Gaweda L, Pionke R, Krezolek M, Frydecka D, Nelson B, Cechnicki A. The inter-
Campus, Wonju, Republic of Korea
4 play between childhood trauma, cognitive biases, psychotic-like experiences
Department of Business Administration, Yonsei University, Mirae
and depression and their additive impact on predicting lifetime suicidal
Campus, Wonju, Republic of Korea
5 behavior in young adults. Psychol Med. 2020;50(1):116–24.
Department of Social Welfare, Sangji University, Wonju, Republic of Korea
6 22. Chen WY, Chen YY, Liu HH, Kuo PH, Huang MC. The possible mediating effect
Department of Biomedical Engineering, Yonsei University, Mirae Campus,
of alcohol dependence on the relationship between adverse childhood
Wonju, Republic of Korea
7 experiences and attempted suicide. Alcohol. 2018;73:9–15.
Department of Computer & Telecommunications Engineering, Yonsei
23. Perez NM, Jennings WG, Piquero AR, Baglivio MT. Adverse childhood experi-
University, Mirae Campus, Wonju, Republic of Korea
8 ences and suicide attempts: the mediating influence of personality develop-
Department of Precision Medicine, Wonju College of Medicine, Yonsei
ment and problem behaviors. J Youth Adolesc. 2016;45(8):1527–45.
University, Wonju, Republic of Korea
9 24. Bjorkenstam E, Hjern A, Bjorkenstam C, Kosidou K. Association of cumulative
Artificial Intelligence Bigdata Medical Center, Wonju College of Medicine,
childhood adversity and adolescent violent offending with suicide in early
Yonsei University, Wonju, Republic of Korea
adulthood. JAMA Psychiatry. 2018;75(2):185–93.
25. Stansfeld SA, Clark C, Smuk M, Power C, Davidson T, Rodgers B. Childhood
Received: 14 August 2023 / Accepted: 18 June 2024 adversity and midlife suicidal ideation. Psychol Med. 2017;47(2):327–40.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Lee et al. BMC Psychiatry (2024) 24:485 Page 14 of 14

26. Cluver L, Orkin M, Boyes ME, Sherr L. Child and adolescent suicide attempts, 41. Bremner JD, Bolus R, Mayer EA. Psychometric properties of the early trauma
suicidal behavior, and adverse childhood experiences in South Africa: a inventory–self report. J Nerv Ment Dis. 2007;195(3):211–8.
prospective study. J Adolesc Health. 2015;57(1):52–9. 42. Löwe B, Kroenke K, Herzog W, Gräfe K. Measuring depression outcome with
27. Dube SR, Anda RF, Felitti VJ, Chapman DP, Williamson DF, Giles WH. Childhood a brief self-report instrument: sensitivity to change of the patient health
abuse, household dysfunction, and the risk of attempted suicide throughout questionnaire (PHQ-9). J Affect Disord. 2004;81(1):61–6.
the life span: findings from the adverse childhood experiences study. JAMA. 43. Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing gen-
2001;286(24):3089–96. eralized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166(10):1092–7.
28. Johnson JG, Cohen P, Gould MS, Kasen S, Brown J, Brook JS. Childhood 44. Binge drinking [https://www.cdc.gov/alcohol/fact-sheets/binge-drinking.
adversities, interpersonal difficulties, and risk for suicide attempts during late htm#:~:text=Binge drinking is a serious,on an occasion for women].
adolescence and early adulthood. Arch Gen Psychiatry. 2002;59(8):741–9. 45. Wilcox S. Multidimensional scale of perceived social support. Psychol Trauma.
29. Jiang L, Shi X, Wang Z, Wang S, Li Z, Wang A. Sleep problems and emotional 2010;2(3):175–82.
dysregulation mediate the relationship between childhood emotional abuse 46. Lubben JE. Assessing social networks among elderly populations. Fam Com-
and suicidal behaviors: a three-wave longitudinal study. J Affect Disord. munity Health. 1988;11(3):42–52.
2021;295:981–8. 47. McCrory C, Dooley C, Layte R, Kenny RA. The lasting legacy of childhood
30. Lee MA. Emotional abuse in childhood and suicidality: the mediating roles of adversity for disease risk in later life. Health Psychol. 2015;34(7):687–96.
re-victimization and depressive symptoms in adulthood. Child Abuse Negl. 48. Tibbe TD, Montoya AK. Correcting the bias correction for the bootstrap
2015;44:130–9. confidence interval in mediation analysis. Front Psychol. 2022;13:810258.
31. Wan Y, Chen R, Ma S, McFeeters D, Sun Y, Hao J, Tao F. Associations of adverse 49. Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect
childhood experiences and social support with self-injurious behaviour and effects in simple mediation models. Behav Res Methods Instrum Comput.
suicidality in adolescents. Br J Psychiatry. 2019;214(3):146–52. 2004;36(4):717–31.
32. Rabbany JM, Ellis S, Metts A, Burke A, Brent DA, Melhem N, Marcott S, Mann 50. Rijnhart JJM, Twisk JWR, Chinapaw MJM, de Boer MR. Comparison of meth-
JJ. Mood disorders and aggressive traits mediate effects of reported child- ods for the analysis of relatively simple mediation models. Contemp Clin
hood adversity on suicide attempt risk. Arch Suicide Res. 2022;27(4):1207–30. Trials Commun. 2017;7:130–5.
33. Stagaki M, Nolte T, Feigenbaum J, King-Casas B, Lohrenz T, Fonagy P, Personal- 51. Barrett P. Structural equation modelling: adjudging model fit. Pers Indiv Differ.
ity and Mood Disorder Research Consortium, Montague PR. The mediating 2007;42(5):815–24.
role of attachment and mentalising in the relationship between childhood 52. Steiger JH. Understanding the limitations of global fit assessment in struc-
maltreatment, self-harm and suicidality. Child Abuse Negl. 2022;128:105576. tural equation modeling. Pers Individ Dif. 2007;42(5):893–8.
34. Rhee TG, Barry LC, Kuchel GA, Steffens DC, Wilkinson ST. Associations of 53. Hu LT, Bentler PM. Cutoff criteria for fit indexes in covariance structure
adverse childhood experiences with past-year DSM-5 psychiatric and sub- analysis: conventional criteria versus new alternatives. Struct Equ Model.
stance use disorders in older adults. J Am Geriatr Soc. 2019;67(10):2085–93. 1999;6(1):1–55.
35. Ehrlich KB, Miller GE, Chen E. Childhood adversity and adult physical health. 54. Tzouvara V, Kupdere P, Wilson K, Matthews L, Simpson A, Foye U. Adverse
In: Developmental psychopathology: risk, resilience, and intervention. Volume 4, childhood experiences, mental health, and social functioning: a scoping
3rd edn. Edited by Cicchtti D. Hoboken, New Jersey: John Wiley & Sons; 2016: review of the literature. Child Abuse Negl. 2023;139:106092.
1–42. 55. Jung J, Rosoff DB, Muench C, Luo A, Longley M, Lee J, Charlet K, Lohoff FW.
36. Kye SY, Park K. Suicidal ideation and suicidal attempts among adults with Adverse childhood experiences are associated with high-intensity binge
chronic diseases: a cross-sectional study. Compr Psychiatry. 2016;73:160–7. drinking behavior in adulthood and mediated by psychiatric disorders.
37. Vasiliadis H-M, D’Aiuto C, Lamoureux-Lamarche C, Pitrou I, Gontijo Guerra Alcohol Alcohol. 2020;55(2):204–14.
S, Berbiche D. Pain, functional disability and mental disorders as potential 56. Kim SY, Kim HJ. Trends in alcohol consumption for Korean adults from 1998
mediators of the association between chronic physical conditions and to 2018: Korea National Health and Nutritional Examination Survey. Nutrients.
suicidal ideation in community living older adults. Aging Ment Health. 2021;13(2):609.
2021;26(4):791–802. 57. Fässberg MM, Cheung G, Canetto SS, Erlangsen A, Lapierre S, Lindner R,
38. Rytilä-Manninen M, Haravuori H, Fröjd S, Marttunen M, Lindberg N. Mediators Draper B, Gallo JJ, Wong C, Wu J, et al. A systematic review of physical illness,
between adverse childhood experiences and suicidality. Child Abuse Negl. functional disability, and suicidal behaviour among older adults. Aging Ment
2018;77:99–109. Health. 2016;20(2):166–94.
39. Bahk Y-C, Jang S-K, Choi K-H, Lee S-H. The relationship between childhood
trauma and suicidal ideation: role of maltreatment and potential mediators.
Psychiatry Investig. 2017;14(1):37–43. Publisher’s Note
40. Lee J-k, Kim M-H, Hwang S, Lee K-J, Park JY, Shin T, Lim H-S, Urtnasan E, Chung Springer Nature remains neutral with regard to jurisdictional claims in
M-K, Lee J. Developing prediction algorithms for late-life depression using published maps and institutional affiliations.
wearable devices: a cohort study protocol. BMJ Open. 2024;14(6):e073290.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

onlineservice@springernature.com