Background: Self-efficacy has been widely studied in suicide, both in its causality and treatment effects. However, the evidence of causality is still ambiguous and there is a lack of generalizability. The aim of this study was to examine the relationship between self-efficacy and suicidal ideation through a population cohort study. Methods: The community-based household survey using a self-administered questionnaire was conducted in a rural area of Japan, Happo Town, in Akita Prefecture with community residents aged 30 and over at two respective time points by local health volunteers. The baseline survey was conducted in 2010 with a response rate of 88.9% (n = 6044). Among them, 3812 residents met the inclusion criteria for the follow-up survey in 2012 where the response rate was 75.3% (n = 2869). Exposure variables to suicidal ideation included demographic details, depression and self-efficacy. The Institutional Review Board and the Ethics Committee of Akita University approved the study protocol and all subjects signed informed consent. Results: A total number of 2105 participants (76.4%) without suicidal ideation in the baseline study were enrolled into the follow-up study, and 8.2% of them had developed suicidal ideation. These participants with suicidal ideation were significantly less likely to be married/cohabitant; they had worse subjective health, poorer self-perceived economic status, stronger depressive mood, and lower self-efficacy scores. The odds ratio of the self-efficacy scores at follow-up survey for participants who had developed suicidal ideation were about 2 times lower than at baseline (95% confidence interval = 1. 53 - 3.06). After adjusting for all confounding factors, the association was still significant (OR = 1.66, 95% CI = 1.15 - 2.42). Conclusion: This result suggests that suicidal ideation may be prevented by increasing self-efficacy. We suggest that self-efficacy can be an effective tool for identifying people with suicidal ideation, and increasing self-efficacy can be strategically beneficial for larger suicide prevention.
Suicide is the 15th leading cause of death, with more than 800,000 people dying of suicide each year, thus accounting for 1.4% of all deaths worldwide [
Various studies have concluded that suicide is a complicated issue, and that risk factors comprise social, cultural, environmental, personal, financial, or mental health factors [
On the other hand, low self-efficacy is often reported to be associated with higher levels of depressive and anxiety symptoms [
On the basis of these findings, we hypothesize that self-efficacy may predict suicide ideation. To the best of our knowledge, the causal relationship between self-efficacy and suicide ideation has not been explored in a population cohort study.
The aim of this study therefore was to examine the relationship between self-efficacy and suicidal ideation through a population cohort study.
Participant flow is shown in
A set of self-administered questionnaires was distributed door-to-door to the community residents age 30 and over at two respective time points by local health volunteers. The questionnaire could be filled in at home by the person him/herself or by family members. The signed and sealed questionnaire was then collected by the health volunteers. The baseline survey was conducted in 2010 with a response rate of 88.9% (n = 6044), and 4066 questionnaires were retained by excluding the anonymous questionnaires. After further excluding residents age 85 and over to avoid gender and selection biases, questionnaires from 3812 residents remained for analysis. The follow-up survey was conducted in 2012 with a response rate of 75.3% (n = 2869).
Basic demographic characteristics including age, gender, education, job status, marital status, and household structure were collected in the survey. Objective health status was measured by indicating one or more diseases under treatment. Subjective health status was measured by a simple question, “what do you think about your
health”. The answer choices consisted of “excellent”, “good”, “not so good”, and “poor”. Excellent and good were considered healthy, not so good and poor were considered not healthy. Self-perceived economic status was also measured by a simple question, “what do you think about your economic circumstance” with four choices of answers, “excellent”, “good”, “not so good”, and “poor”. Excellent and good were considered wealthy, not so good and poor were considered not wealthy.
Depression was measured by the K6 (The Kessler 6-Item Psychological Distress Scale) [
Self-efficacy was measured by the GSES (General Self-Efficacy Scale) [
Suicidal ideation was measured by a simple question, “have you ever wished to die” that was requested from the participants without suicidal ideation at baseline and during the follow-up study. The choices of answer were “no”, “a little”, or “yes”. Participants who answered “a little” or “yes” were classified as the group with suicidal ideation thoughts.
First, gender difference of demographic characteristics and psychosocial, economic, and health factors at baseline were confirmed using Chi-square test for categorical variables and Kruskal-Wallis test for age. Next, to examine the association between suicidal ideation and demographic characteristics and psychosocial, economic, and health factors at follow up, univariate analyses using the Chi-square test were performed while applying the Yates correction wherever necessary. Variables that appeared to be the risk factors for suicidal ideation, as suggested by the univariate analysis, were entered into the multivariate logistic regression model to calculate the odds ratio (OR) and its confidence interval (CI). Statistical significance was set at p < 0.05. Data were analyzed using SPSS (SPSS Inc.) Version 20.0.
Basic demographic characteristics of individual participants were traceable for the follow-up study. Privacy of the participants was protected, that is, personal information was not disclosed and not used elsewhere. The participants also had all rights to refuse participation or choose not to disclose some specific information. The Institutional Review Board and the Ethics Committee of Akita University approved the study protocol and all subjects signed informed consent.
Suicidal ideation was assessed at baseline and in the follow-up study. During baseline, 346 participants that had indicated suicidal ideation (13.6%, male = 136, female = 210) were excluded, leaving 2105 participants for statistical analysis. During the follow-up survey, 8.2% of the participants responded having suicidal ideation (n = 172, male = 70, female = 102).
This study involved residents of an entire community thus allowing generalizability of the findings. People with high self-efficacy at the baseline survey were 2 times less likely to develop suicidal ideation within the 2 years until the follow-up survey. This finding was not affected by adjusting confounding factors such as gender, age, marital status, self-perception of economic status, self-perception of health, and depression. This result suggests that suicidal ideation may be prevented by increasing self-efficacy. However, when the data were analyzed separately for each gender, the association between self-efficacy and suicidal ideation remained significant in females but not in males. This could be explained by the population structure in our study, which is typical for an aging society.
In our study, female participants were generally older and had higher needs for medical attention. This difference could explain why females had lower self-efficacy and higher levels of depressive mood. Furthermore, females after menopause (aged 50 - 69 years) had significantly lower suicidal ideation after adjusting for self-ef- ficacy, subjective health status, and depressive mood. When health further deteriorates with age, self-efficacy
Variable | Male | Female | |
---|---|---|---|
n = 959 (45.6%) | n = 1146 (54.4%) | p-value | |
Age (mean ± SD) Ten-year age groups (%) 30 - 39 40 - 49 50 - 59 60 - 69 70 - 79 80 - 84 Household structure (%) Living alone Living with others Spouse (without children) Two generations and more Others Marital status Single Married/cohabitant Separated Divorced Widowed Educational Elementary school Middle school High school College, university, graduate school Others Job Status (%) Self-employed White collar Blue collar Engineer Sales/service Housewife/husband Unemployed Others Self-perceived household economic status (%) Good to excellent Poor to not so good Objective health status (%) None Having disease under treatment Subjective health status (self-rated health) (%) Good to excellent Poor to not so good Mental health (K6 score) (%) Normal (K6 score < 9) Depressive mood (K6 score ≥ 9) Self-efficacy (%) High (26 - 40) Low (10 - 25) | 59.8 ± 13.3 8.9 13.8 24.0 26.7 21.1 5.6 4.8 29.4 61.9 3.8 11.1 81.5 1.0 3.4 3.0 3.7 27.6 50.6 17.1 1.1 29.1 12.4 15.3 4.1 3.7 0.0 21.6 13.7 35.8 64.2 48.2 51.8 82.1 17.9 93.5 6.5 57.6 42.4 | 61.0 ± 13.5 8.5 12.8 21.0 26.1 24.9 6.7 23.2 23.2 61.2 4.3 4.5 69.2 1.9 4.0 20.3 4.9 32.4 44.0 17.4 1.3 18.9 6.7 3.6 4.6 3.2 19.0 18.0 26.2 38.8 61.2 39.8 60.2 77.2 22.8 88.7 11.3 47.0 53.0 | 0.062 0.235 <0.001 <0.001 0.027 <0.001 0.093 <0.001 0.003 <0.001 <0.001 |
Notes: SD = standard deviation. Chi-square test for categorical variables. Kruskal-Wallis test for age.
Variable | Considered suicide during past 24 months | ||
---|---|---|---|
No n = 1933 (91.8%) | Yes n = 172 (8.2%) | p-value | |
Gender (%) Male Female Ten-year age groups (%) 30 - 39 40 - 49 50 - 59 60 - 69 70 - 79 80 - 84 Household structure (%) Living alone Living with others Marital status Married/cohabitant Single Separated, divorced, widowed Educational Elementary school Middle school High school College, university, graduate school Others Job Status (%) Self-employed White collar Blue collar Engineer Sales/service Housewife/husband Unemployed Others Self-perceived household economic status (%) Good to excellent Poor to not so good Objective health status (%) None Having disease under treatment Subjective health status (self-rated health) (%) Good to excellent Poor to not so good Mental health (K6 score) (%) Normal (K6 score < 9) Depressive mood (K6 score ≥ 9) Self-efficacy (%) High (26 - 40) Low (10 - 25) | 92.7 91.1 86.3 91.8 91.1 94.2 91.8 92.4 91.9 91.8 92.8 90.3 88.8 93.3 91.1 92.5 90.8 88.0 94.1 92.3 93.0 88.9 90.1 91.4 91.7 90.1 94.0 90.4 91.7 91.9 93.0 87.2 92.8 81.7 94.3 88.5 | 7.3 8.9 13.7 8.2 8.9 5.8 8.2 7.6 8.1 8.2 7.2 9.7 11.2 6.7 8.9 7.5 9.2 12.0 5.9 7.7 7.0 11.1 9.9 8.6 8.3 9.9 6.0 9.6 8.3 8.1 7.0 12.8 7.2 18.3 5.7 11.5 | 0.104 0.032 1.000 0.031 0.670 0.451 0.002 0.872 <0.001 <0.001 <0.001 |
Notes: SD = standard deviation. Chi-square test for categorical variables.
Variable | Crude | Model 1 | Model 2 |
---|---|---|---|
OR (95% CI) | OR (95% CI) | OR (95% CI) | |
Self-efficacy High (26 - 40) Low (10 - 25) Gender (%) Male Female Ten-year age groups 30 - 39 40 - 49 50 - 59 60 - 69 70 - 79 80 - 84 Marital status Married/cohabitant Single Separated, divorced, widowed Self-perceived household economic status Good to excellent Poor to not so good Subjective health status (self-rated health) Good to excellent Poor to not so good Mental health (K6 score) Normal (K6 score < 9) Depressive mood (K6 score ≥ 9) Hosmer-Lemeshow test | 1 2.16 (1.53 - 3.06)** | 1 2.03 (1.42 - 2.91)** 1 1.14 (0.80 - 1.64) 1 0.55 (0.30 - 1.03) 0.59 (0.33 - 1.03) 0.37 (0.20 - 0.69)** 0.52 (0.28 - 0.97)* 0.70 (0.30 - 1.62) 1 1.07 (0.58 - 1.98) 1.64 (1.04 - 2.56)* χ2 = 4.79 (df = 7) p = 0.685 | 1 1.66 (1.15 - 2.42)** 1 1.13 (0.78 - 1.64) 1 0.56 (0.30 - 1.06) 0.60 (0.34 - 1.07) 0.37 (0.20 - 0.70)** 0.53 (0.27 - 1.02) 0.71 (0.30 - 1.69) 1 1.13 (0.61 - 2.12) 1.58 (0.99 - 2.52) 1 1.41 (0.95 - 2.10) 1 1.55 (1.02 - 2.35)** 1 2.22 (1.38 - 3.55)** χ2 = 10.22 (df = 8) p = 0.250 |
Notes: OR = odds ratio; CI = confidence interval. *p < 0.05; **p < 0.01.
Variable | Crude | Model 1 | Model 2 |
---|---|---|---|
OR (95% CI) | OR (95% CI) | OR (95% CI) | |
Self-efficacy High (26 - 40) Low (10 - 25) Ten-year age groups 30 - 39 40 - 49 50 - 59 60 - 69 70 - 79 80 - 84 Marital status Married/cohabitant Single Separated, divorced, widowed Self-perceived household economic status Good to excellent Poor to not so good Subjective health status (self-rated health) Good to excellent Poor to not so good Mental health (K6 score) Normal (K6 score < 9) Depressive mood (K6 score ≥ 9) Hosmer-Lemeshow test | 1 1.77 (1.06 - 2.97)* | 1 1.70 (1.00 - 2.89) 1 0.62 (0.22 - 1.75) 0.95 (0.39 - 2.30) 0.66 (0.26 - 1.71) 0.65 (0.23 - 1.84)* 0.84 (0.20 - 3.53) 1 1.24 (0.57 - 2.68) 1.16 (0.44 - 3.09) χ2 = 4.40 (df = 8) p = 0.848 | 1 1.44 (0.83 - 2.51) 1 0.57 (0.20 - 1.63) 0.85 (0.34 - 2.08) 0.57 (0.22 - 1.52) 0.61 (0.21 - 1.79) 0.71 (0.16 - 3.12) 1 1.33 (0.61 - 2.90) 1.27 (0.47 - 3.42) 1 1.31 (0.72 - 2.39) 1 1.80 (0.94 - 3.46) 1 1.53 (0.63 - 3.68) χ2 = 7.08 (df = 8) p = 0.528 |
Notes: OR = odds ratio; CI = confidence interval. *p < 0.05; **p < 0.01.
Variable | Crude | Model 1 | Model 2 |
---|---|---|---|
OR (95% CI) | OR (95% CI) | OR (95% CI) | |
Self-efficacy High (26 - 40) Low (10 - 25) Ten-year age groups 30 - 39 40 - 49 50 - 59 60 - 69 70 - 79 80 - 84 Marital status Married/cohabitant Single Separated, divorced, widowed Self-perceived household economic status Good to excellent Poor to not so good Subjective health status (self-rated health) Good to excellent Poor to not so good Mental health (K6 score) Normal (K6 score < 9) Depressive mood (K6 score ≥ 9) Hosmer-Lemeshow test | 1 2.46 (1.52 - 3.98)** | 1 2.29 (1.40 - 3.76)** 1 0.51 (0.23 - 1.11) 0.41 (0.19 - 0.86)* 0.24 (0.11 - 0.55)** 0.44 (0.20 - 0.97)* 0.58 (0.20 - 1.67) 1 0.91 (0.30 - 2.73) 1.84 (1.08 - 3.15)* χ2 = 2.792 (df = 8) p = 0.947 | 1 1.84 (1.10 - 3.08)* 1 0.57 (0.25 - 1.27) 0.46 (0.21 - 0.98)* 0.27 (0.12 - 0.63)** 0.48 (0.21 - 1.11) 0.68 (0.23 - 2.04) 1 0.95 (0.32 - 2.88) 1.68 (0.96 - 2.93) 1 1.44 (0.85 - 2.47) 1 1.41 (0.82 - 2.43) 1 2.59 (1.46 - 4.60)** χ2 = 6.85 (df = 8) p = 0.553 |
Notes: OR = odds ratio; CI = confidence interval. *p < 0.05; **p < 0.01.
declines and thus raises the probability for suicidal ideation. This is supported by previous findings, where self- efficacy was found to be an important predictor for suicidal ideation in a residential care home [
Our study suggests that self-efficacy influenced suicidal ideation risk factors. We also found that people who had lower self-efficacy concurrently suffered from stronger depressive mood and poorer subjective health. This finding agrees with our prediction, as self-efficacy is believed to have a strong correlation with mental distress [
The strong relationship between suicidal ideation and hopelessness as well as mental illness cannot be denied. However, in reality, mental illness including depression is often under-diagnosed and under-treated [
Future studies need to examine the impact of increasing self-efficacy in suicide prevention for different age groups and develop specific strategies accordingly. For example, a middle-aged man may need more mental health support to increase self-efficacy at the work place, while elderly people may need more opportunities for social involvement, and community empowerment may serve these needs.
Japanese people often hold back their negative thoughts and emotions in front of others; hence, it is difficult to identify depression and hopelessness at an early stage, which makes suicide prevention even more difficult. It makes sense to use self-efficacy as a countermeasure for preliminary screening of suicidal ideation, as we understand now that self-efficacy is not only a key element of depression and hopelessness, but also has a general impact on suicidal ideation. Modifying self-efficacy is expected to induce behavioral change and perhaps alter the attitudes towards life. Self-efficacy can be improved in all aspects of life, and it plays a major role in shaping the early development of thoughts and personal characteristics. Early self-efficacy intervention should be considered for early school education as well as a family approach.
LimitationsThis study has several shortcomings. Although the self-administered, signed, and sealed questionnaire promised a certain level of privacy, there may have been some resistance towards the idea of signing as an identifiable individual. This may have resulted in a bias in the findings, where people with poor mental health may have refused to participate in the study to avoid identification. Study subjects were community residents, which allow a certain generalizability of the findings; however, they are limited to the population living in rural Japan, where crop production is the major economic activity and main source of income. The time span between baseline and follow-up study might have been too short (2 years), as mental health problems develop over an extended period of time. We suggest that a future study should consider a longer time span to evaluate the relationship between self-efficacy and mental health as well as its effects on suicidal ideation.
In our study, self-efficacy was associated with suicidal ideation. Low self-efficacy may be regarded as a predictor of suicidal ideation, and increasing self-efficacy can be strategically beneficial for broader suicide prevention.
We thank for the cooperation of all staff of the Department of Health and Welfare of the Happo town Government Office, and Ms. Yong for assisting with the manuscript. This work was supported by JSPS KAKENHI Grant Numbers 20590632, 23590773.