Literature review

Many researchers have investigated suicidal ideation among general and specific populations in other countries. Kaplan et al. (1997) [11] conducted a study on the relationship between physical abuse during adolescence and suicide attempts. The results of that study demonstrated that abused adolescents were significantly more prone to risk factors related to adolescent suicide than those who have not been abused in their adolescence, family disintegration, depression, disruptive behavior disorders, and substance abuse and dependence. Moreover, it was found that women in Asia who were exposed to violence appear to have an increased risk of suicidal behavior and mental disorders [12–16]. In another study by Wikman et al. (2022) [17] a substantial proportion of women with confirmed premenstrual dysphoric disorder reported current suicidal ideation in the late luteal phase. Furthermore, Katzenmajer-Pump et al. (2022) [18] showed the importance of recognizing “worthlessness” for suicide prevention in adolescents with attention-deficit/hyperactivity disorder (ADHD). This study revealed that the symptoms of depression and anxiety mediate the relationship between ADHD and suicidal thoughts and planning. A study done in China examined the risk factors and clinical characteristics of suicidal ideation in Chinese patients with depression and the independent risk predictors included crying, helplessness, worthlessness, hopelessness, unusually restless, self-harm, mood-incongruent psychosis, feeling losing control of oneself, hypersomnia, sensory system complaints, derealization, guilt, suicidal attempts, male gender, the total course of depression [19]. Moreover, a study done in India found a significant positive association between the consumption of alcohol by husbands and the presence of depressive symptoms and suicidal thoughts in their wives [20]. Another study in India found that in contrast to married working women, single working women experience higher levels of suicidal thoughts. The same study indicated that, compared to working women between the ages of 46 and 55, women in the 35–43 age range reported more suicidal thoughts and the women employed in the private sector exhibited a higher rate of suicidal ideation compared to those in the public sector [21]. A study done by Kiliçarslan et al. (2024) determined several significant factors associated with the suicidal tendency of Turkish married women, women with a previous relationship, and women currently in a relationship, including age, educational level, health status, number of children, the sector in which the spouse or partner works, the drinking status of the spouse or partner, and household income level [22].

Several studies have been conducted on suicides and suicidal ideation among women in Sri Lanka where most existing literature has focused on the general population, youth, and/or pregnant women [8, 9, 23, 24]. In a study in 2006, sociological risk factors for youth suicidal behavior were found to include poor father-child interactions, low educational attainment, poor peer interaction, easy access to pesticides and medicinal drugs, a mother being away or abroad, breakups with romantic partners, and parents who exhibited alcoholism [23]. Genetic and other risk factors for suicidal ideation and the relationship with depression were examined in a study done in Sri Lanka [9] in which various factors were found to be associated with suicidal thoughts, such as being female, ending a marriage, low level of education, residing in an urban area, losing a parent at a young age, low standard of living, and experiencing stressful life events in the previous year, etc. The same study also found that suicidal ideation was closely related to depression, abnormal fatigue, and use of alcohol and tobacco. Furthermore, a significant impact of genetic and non-shared environmental factors on suicidal thoughts in both men and women was observed. For women, the genetic factor that contributes to suicidal ideation was mostly influenced by depression. However, in men, there was a distinct heritable component linked to suicidal thoughts that was not related to depression. Rajapakse et al. (2020) conducted a case-control analysis to investigate the connection between exposure to childhood adversity and hospital admissions for non-fatal self-poisoning in Sri Lanka. The results indicated that being physically or emotionally abused or neglected as a child, witnessing domestic violence, living with a family member who is mentally ill or suicidal, and experiencing parental death, separation, or divorce during childhood, etc. increased the chances of self-poisoning in adulthood [25]. Palfreyman (2021) studied the prevalence of common perinatal mental illnesses and suicidal tendencies among women residing in urban areas of Sri Lanka and the analysis revealed that the most significant factors associated with depression and suicidal ideation and/or behaviors during the current pregnancy were exposure to intimate partner violence and a history of suicidal ideation and/or behaviors [24]. Samaraweera et al. (2010) identified that the active suicidal ideators were high/frequent among physically ill women, and with higher levels of helplessness and hopelessness. Moreover, Bandara et al. (2022) found that women exposed to past-year domestic violence in Sri Lanka are 4.08 times more likely to self-harm than women not exposed to domestic violence [12]. In a study done by Bandara et al. (2024), suicide and intimate partner violence (IPV) were found to be correlated highlighting the importance of prioritizing trauma-informed approaches and addressing IPV in post conflict settings as suicide prevention initiatives in Sri Lanka [13]. Although these studies have explored factors associated with suicidal ideation in Sri Lanka, nearly all of them have utilized small sample sizes from specific locations or target groups. Furthermore, there have no studies examined the factors associated with suicidal ideation among married women or women cohabiting with a male partner in the Sri Lankan context.

This study aims to fill significant gaps in both local and global literature by examining, for the first time, the prevalence of suicidal ideation among married and cohabiting women in Sri Lanka, and identifying the factors associated with it. Utilizing a nationally representative large sample, this research ensures greater accuracy and provides novel insights into the subject. The originality of this study lies in its focus on a previously under-researched population, offering unique findings that can guide the development of targeted interventions. Furthermore, the results will be instrumental in identifying appropriate measures to reduce suicidal ideation in women vulnerable to sexual abuse, thereby offering valuable information to decision-makers, health authorities, and policymakers.

Data source, sampling, and data collection

The data for this study were secondary data obtained on 26/10/2023 from the 2019 Women’s Well-being Survey (WWS) conducted in Sri Lanka by the Department of Census and Statistics (DCS). This was the first national-level survey conducted in Sri Lanka on violence against married and unmarried and cohabiting females [26]. The data were collected in this survey by using the Computer-Assisted Personal Interviewing Technique (CAPI) in the form of a questionnaire administered by female staff from the DCS. The WHO Multi-Country Study on Women’s Health and Life Experiences questionnaire version 12.04.01 (dated 10 July 2018) was used for this survey [26]. The survey used a multi-stage sampling approach, stratified by sector (urban, rural, and estate). The sampling frame used in the survey consisted of 65,000 census blocks, where 252 were selected as the primary sampling unit (PSU), and 10 housing units were selected from each selected PSU as the secondary sampling unit (SSU). The selection was done using a systematic random sampling method. Eligible women who were 15 years or older in each household were interviewed.

The study sample consists of secondary data from 1462 females, who were 15 years or older and currently married or living with a male partner, extracted from WWS 2019.

Ethical approval

The survey was conducted by the DCS using the WHO’s ethical and safety recommendations for research on violence against women as a guide [26] and consent from each respondent was obtained before continuing. The detailed measures are given in the "Women’s Well-being Survey—2019 Final Report" published by the DCS Sri Lanka. Ethical approval is not required for this study as this is an analysis of secondary data where all data are anonymized by DCS to conceal the characteristics and the identity of the respondents. Therefore, the authors had no access to information that could identify individual participants during or after data collection.

Variables

Independent variables.

The study used 33 categorical independent variables, including socio-demographic variables such as the woman’s age, marital status, and highest education level, as well as variables indicating psychological distress symptoms such as depression, hopelessness, worthlessness, and experiences of physical, sexual, psychological violence, and economic violence by the women at different stages. Detailed descriptions of the independent variables are given in Table 1. The selection of independent variables was based on an analysis of relevant literature and the availability of those variables in the dataset considered.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Details of independent variables.

https://doi.org/10.1371/journal.pone.0312753.t001

Statistical analysis

A bivariate analysis was performed to investigate the association between each independent variable and the dependent variable using the chi-square test of association. Variables that were significant at 5% in the chi-square test of association were selected to fit a binary logistic regression model with backward elimination to identify the effect of the association between the dependent variable and the selected independent variables, as well as the most contributing risk factors of suicidal ideation. The odds ratios were then interpreted for the significant variables. The performance of the final model was checked using a receiver operating characteristic (ROC) curve and the area under the curve (AUC). Finally, the model assumptions were tested. Data analysis was performed using Statistical Analysis System (SAS) On Demand for Academics. Moreover, p < 0.05 was considered the acceptable statistical level of significance, and 95% confidence intervals (CI) were used for the odds ratio.

Baseline characteristics

The baseline characteristics of the study sample are shown in Table 2. Of the 1,462 study participants, 13.2% (193) of the married or cohabiting women reported having suicidal ideation. Although a predominant number of women lived in rural areas (79.69%) at the time of the survey, the highest prevalence of suicidal ideation was found in the estate sector (18.75%). By the time of the survey, 99.59% of the women in the sample were married, and the highest prevalence of suicidal ideation was observed among women who lived together with a partner (83.33%). Additionally, 25.1% of the study sample consisted of women aged between 40 and 49 years, and 79.48% of women had up to secondary education. Women aged between 40–49 years and women with no education reported the highest prevalence of suicidal ideation (17.71% and 27.27%, respectively). While 67.03% of women were not working at the time of the survey, 70.31% of women had their household income coming from household members’ work. Moreover, the highest prevalence of suicidal ideation was observed among working women (18.05%) and women with no fixed income (23.08%). Of the sample, 21.47%, 21.00%, 22.09%, 15.19%, 19.84%, and 27.57% of women reported feeling nervous, restless, depressed, worthless, hopeless, and that everything was an effort, often. The prevalence of suicidal ideation was also highest among those who often felt likewise (26.75%, 32.57%, 31.89%, 33.33%, 32.07%, and 26.55%, respectively). Moreover, 26.2%, 17.85%, 6.02%, and 15.05% of respondents reported that they had faced psychological, physical, sexual, and economical violence from their current husband or partner, respectively. The highest prevalence of suicidal ideation was reported among women who had faced psychological, physical, sexual, and economical violence from their current husband or partner (29.77%, 39.85%, 47.73%, and 25.91%, respectively). Furthermore, 7.66% and 25.44% of women experienced physical violence since they were 15 years old and before 15 years old, respectively. The prevalence of suicidal ideation was highest among those who had encountered physical violence (33.04% and 16.67%, respectively). Additionally, 3.76% and 1.71% of women experienced sexual violence since 15 years old and before 15 years old, respectively, and they reported the highest prevalence of suicidal ideation (23.64% and 48.00%, respectively). Of the respondents, 82.63% reported that they could count on the members of their birth family for support, and the highest prevalence of suicidal ideation was observed among women who reported otherwise (18.90%). Additionally, more than half of the women in the study sample had their first marriage or union with a man during 20 to 29 years of age (63.47%). The prevalence of suicidal ideation was highest among women who had their first marriage or union with a man during 10 to 19 years of age (20.63%). Of the sample, 72.09% were found independent and not living with their parents, in-laws, or relatives. Those who lived with their parents, in-laws, or relatives reported the highest prevalence of suicidal ideation (37.50%). Moreover, 64.57% of women reported health problems such as weak vision, hearing, mobility, memory, communication, etc., while the rest reported no health issues. The prevalence of suicidal ideation was highest among those who reported having health problems (14.94%). Additionally, 92.89% of women had given birth to a child, and the prevalence of suicidal ideation was highest among them (13.62%). Of the sample, 77.43% and 87.55% of women had husbands/partners with secondary education and employment, respectively. Women living with husbands/partners with no education and those who were with working husbands/partners reported the highest prevalence of suicidal ideation (21.21% and 13.28% respectively). It was also mentioned that 56.57% and 2.8% of their husband/ partner consumed alcohol and drugs, respectively, and the prevalence of suicidal ideation was also high among them (17.41% and 36.59%, respectively). Around 4% of women confirmed that their current husband/ partner had an affair with another woman, and suicidal ideation prevalence was also high among them (52.73%). Moreover, 3.9% reported often quarreling with their current husband or partner, and they reported the highest prevalence of suicidal ideation as well (50.88%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Bivariate analysis of dependent and independent variables.

https://doi.org/10.1371/journal.pone.0312753.t002

Effect of associated factors on suicidal ideation

The adjusted odds ratios (AOR) were calculated to determine the strength of the association between suicidal ideation among married/ cohabiting women and the independent variables that remained in the final model after adjusting for all variables in the final model (Table 3). It was found that a woman’s working status had a significant association, with unemployed women having 50% reduced odds of suicidal ideation (AOR: 0.504; 95% CI: 0.347–0.732] compared to working women. As expected, the odds of suicidal ideation were almost three times higher for women who reported that their current husband or partner had an affair with another woman (AOR: 2.909; 95% CI: 1.456–5.812) compared to women with partners who did not report infidelity. In addition, the odds of suicidal ideation were three and two times higher for women who reported feeling restless (AOR: 3.107; 95% CI: 1.936–4.986) and worthless (AOR: 2.44; 95% CI: 1.484–4.014), respectively, often when compared to women who reported not having such feelings at any time. The odds of suicidal ideation were almost 60% higher among women whose husbands or partners consumed alcohol (AOR: 1.595; 95% CI: 1.073–2.371) compared to women who had a teetotaler. A significant association was observed between psychological and physical violence by the current husband or partner and suicidal ideation, where the odds of suicidal ideation were two and four times higher among women who faced psychological violence (AOR: 2.090; 95% CI: 1.405–3.109) and physical violence (AOR: 3.794; 95% CI: 2.501–5.754), respectively, from the current husband or partner. Moreover, the odds of suicidal ideation were 72% higher among women who encountered physical violence from age 15 until the study (AOR: 1.722; 95% CI: 1.001–2.964) compared to women with no such encounters. A highly significant association between suicidal ideation and childhood sexual violence was observed. (AOR: 4.721; 95% CI: 1.851–12.046). Odds of having suicidal ideation of women who got married/ started living together with a man at an age between 20–29 years were less when compared with 10–19 age category (AOR of 0.563; 95% CI: 0.381, 0.832) and this shows the protective effect of marrying at a later age compared to younger age.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Odds ratio estimates.

https://doi.org/10.1371/journal.pone.0312753.t003

Fig 1 demonstrates the performance of the fitted logistic regression model for predicting suicidal ideation among Sri Lankan women who were married or living with a man. The AUC of the logistic regression model was 0.8560 (95% CI: 0.8265, 0.8855), which indicated a very good predictive ability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Receiver operating characteristic (ROC) curve for logistic model.

https://doi.org/10.1371/journal.pone.0312753.g001

Model assumptions

The assumption of no multicollinearity among the independent variables was tested using the variance inflation factor (VIF). The results showed a mean VIF value of 1.5, indicating the absence of high multicollinearity among the independent variables. The assumption of independence of errors was tested for overdispersion by calculating the ratio of deviance and Pearson statistics values to the degrees of freedom of the final model. The ratio of deviance (0.59) and the ratio of Pearson statistic (1.00), both being less than 1, indicated no overdispersion. The linearity assumption in logistic regression implies that each continuous predictor is linearly related to the logit of the outcome [27]. Since all the variables are categorical, this assumption did not need to be tested. Finally, the presence of outliers was checked by examining the standardized residuals between the predicted and observed outcomes. Residuals greater than 3 or less than -3 are considered potential outliers [28]. In this study, out of 1462 cases, 43 outliers were identified.

Limitations of the study

This study is a cross-sectional study as it only used data from the 2019 WWS Sri Lanka. As a result, causal inference between the predictor variables and suicidal ideation cannot be considered. The study utilized variables that were significant and considered in previous literature that were also available in the study dataset. However, other factors that may be associated with suicidal ideation and that were not included in the dataset were not considered. Another limitation is that the observations in the dataset were limited, but there were many explanatory variables, which could lead to over-fitting of the models and incorrect predictions. Furthermore, the information provided in the study was based on self-reported data, which carries the risk of being unreliable or deceptive, and there is also a possibility of recall bias being present.

Implications for practice and/or policy

The findings of the current study found a strong link between depression symptoms such as worthlessness, restlessness and suicidal ideation in married or cohabiting women in Sri Lanka. Successful psychological, pharmacological, and/or social intervention can lead to a decrease in suicidal thoughts. Unfortunately, many women in Sri Lanka are hesitant to seek mental health services due to their stigma and discrimination, which can exacerbate their mental health problems and increase the risk of suicide.

It is crucial to encourage women to seek help when necessary and address the stigma associated with mental health services in society. In addition, authorities should implement measures to decrease intimate partner violence and, physical and sexual abuse during childhood and adolescence, educate parents, teachers, and school children about various types of abuse and the process of taking appropriate legal actions to prevent recurrence. The authorities should also provide support to victims to recover from the psychological and physical trauma caused by such abuse; hence, many stakeholders need to be involved in managing such complicated cases. It is crucial to promote relationship counseling for couples as a mean to enhance the bond between women and their intimate partners.