Determinants of suicidal ideation among patients with mental disorders visiting psychiatry outpatient unit in Mekelle town, psychiatric clinics, Tigray, Northern Ethiopia: a case–control study

Background Globally, more than 450 million people suffer from a mental or behavioral disorder. Psychiatric disorder and its duration, physical illness, family history of mental illness, previous suicidal attempt, unemployment, poor social support, and psychotic symptoms are of the main reasons enabling patients to be suicidal ideates. The purpose of this study is to identify the determinants of suicidal ideation among patients with mental disorders in Mekelle, Ethiopia. Methods Case–control study design was employed with a total of 221 study subjects (74 cases and 147 controls) in Mekelle, Ethiopia. Suicidal ideation was measured by the Suicidal Behavior Questionnaire-Revised (SBQ-R) scale. Bivariate and multiple logistic regression analyses were performed to determine between the independent and dependent variables. Results This study revealed that family suicide history (AOR = 6.87, 95% CI [1.138–41.531, P = 0.036), previous attempts history (AOR = 27.457, 95% CI 10.417–72.368, P = 0.0001), family mental illness history (AOR = 3.029, 95% CI 1.088–8.431, P = 0.034), major depressive disorders (AOR = 11.182, 95% CI 2.382–52.501, P = 0.002), and psychiatric comorbid disorders (AOR = 12.245, 95% CI 1.898–78.986, P = 0.008) were significant factors of suicidal ideation. Conclusion Family mental illness history, family suicide history, previous suicide attempt history, major depressive disorders, and psychiatric comorbid disorders were significant factors of suicidal ideation.


Introduction
Globally, more than 450 million people suffer from a mental or behavioral disorder; the total number of people living with depression and anxiety disorders in the world is estimated to 322 million (4.4%) and 264 million (3.6%), respectively. These disorders are common mental disorders and suicide is common in mental disorder especially in depressive patients [1].
There are over 35,000 deaths per year in the United States attributed to suicide and is currently ranked the tenth overall cause of death in the United States. Almost 95% of all persons who commit or attempt suicide have a diagnosed mental disorder. Psychiatric disorders like depressive disorders account for 80% of this figure and schizophrenia accounts for 10%. Psychiatric patients' risk for suicide is 3 to 12 times that of non-patients, but the degree of risk varies, depending on age, sex, diagnosis, and patient status [2]. Suicidal ideation refers to thoughts of serving as the agent of one's own death and it may vary in seriousness depending on the specificity of suicide plans and the degree of suicidal intent [3].
Identifying determinants of suicidal ideation will be helpful to mental health professionals for routine assessing, designing effective prevention and intervention methods that are important in preventing the effect of suicidal ideation and save the patient's life.

Study areas and period
This study was conducted in 2019 at Mekelle town, psychiatric clinics (Ayder Comprehensive Specialized Hospital and Mekelle General Hospital). According to the Ethiopian central statistical agency report; the total population in 2011 has been 367,470 (186,045 males and 181,376 females), which is located in Tigray regional state 783 km away from Addis Ababa, the capital city of Ethiopia to the North. Psychiatric services are given by psychiatrists, general practitioner, M.Sc in mental health, B.Sc in psychiatric nursing and clinical psychologists. More than 30 health professionals work in a psychiatry clinic, out of them, two psychiatrists and two clinical psychologists. The current flow of psychiatric patients on average was 946 and 337 patients per month in Ayder Comprehensive Specialized Hospital and Mekelle General Hospital, respectively. Ayder Comprehensive Specialized Hospital had 24 beds for inpatient services, but Mekelle General Hospital has no inpatient service. The study period was from April 15 to May 30, 2019.

Study design
An institutional-based unmatched case-control study was conducted.

Source population and study population Source population
All psychiatry outpatient unit visitors were in Mekelle town, psychiatric clinics, Tigray, Northern Ethiopia.

Study population
All psychiatric sampled outpatient unit visitors were in Mekelle town, psychiatric clinics, Tigray, Northern Ethiopia.

Inclusion criteria
• Case: psychiatry outpatient unit visitors, clients who had suicidal ideation (scored 8 and above out of 18) and aged 18 years and above, patients attending treatment at psychiatry clinic during the data collection period were included in the study. • Control: psychiatry outpatient unit visitors, clients who had no suicidal ideation, responded to suicidal behavior questions and scored less than 8 out of 18.

Exclusion criteria
Clients who were unable to communicate and unable to sign verbal informed consent and those who had decision incapacity were not be included in the study.

Sample size and sampling procedure Sample size
The sample size was calculated using a double proportion formula taken from a study conducted in Jimma showed that one of the risk factors for suicidal ideation was having a family history of mental illness (AOR = 2.25) and P = 40% [18]. Based on these assumptions, the sample size was calculated as follows: two-sided confidence level = 95%, power = 80%, ratio of cases to controls = 1:2 and odds ratio = 2.25. Then, the desired sample size using Epi info Stata calculation, odds ratio = 1:2, case = 74 and control = 147, the final sample size was 221 (Table 1).

Sampling technique and procedure
The systematic sampling technique was employed. The study participants were proportionally allocated in both psychiatric clinics. Since the ratio of case to control was 1:2, every case 2 controls were selected (Fig. 1).

Data collection procedure
Face-to-face interview method using a structured questionnaire was used in this study to identify determinants of suicidal ideation such as (1) sociodemographic information, (2) biopsychosocial (3) suicidal ideation, (4) suicidal attempts and its methods. Also, the chart was reviewed to check the psychiatric and other medical illness diagnoses. The SBQ-R is a self-report measure of suicidal ideation. This shortened version of the SBQ consists of four questions and used a Likert type scale to assess suicidal behavior history, current suicide status and self-appraisal and expectancies about the future likelihood of engaging in suicidal behavior. The magnitude of overall suicidal ideation and behaviors (as defined by SBQ-R a total score ≥ 8 for adult clinical population) and the total score ranges from 3 to 18.
Data were collected by five B.Sc psychiatry professionals having a previous experience of data collection. The principal investigator checked the filled questionnaires for consistency and completeness daily. The questionnaire was translated from English to Tigrigna language by expert and retranslated to English by another proficient in English. This primary version was made to compare with the original English version to resolve inconsistencies and then the data collectors, who are Tigrigna native speakers, collected data in the Tigrigna questionnaire.

Study variables Dependent variables
Suicidal ideation.

Independent variables
Socio-demographic characteristics, medical and substance-related characteristics, mental disorders, illness and hospital related characteristics, psychosocial-related characteristics, suicidal attempt-related characteristics.

Operational definitions
Suicidal ideation: after the onset of the illness, the patient experiences the degree of having wished, ideation, expectancies about the future likelihood of engaging in suicidal behavior, and plan to kill oneself, and based on SBQ-R scale scores ≥ 8 from 18, at least 2 points out of 4 from suicidal thought or attempt question 1 and the total scale scores range from 3 to 18 [19].

Data quality assurance
The pre-test was conducted on a sample of 5% (7 controls and 4 cases) of the total study population in Wukiro General Hospital before 2 weeks of data collection and a common understanding was reached between the data collectors to avoid inter-rater variability. The pre-test questionnaires were not included in the analysis as part of the main study. Data collection was collected within 30 working days. Regular supervision by the principal investigator was carried out. Each day during data collection, filled questionnaires' were checked for completeness and consistency. Incomplete and unclear questionnaires were returned to the data collectors to get it corrected.

Data analysis procedure
After data collection, filled questionnaires were coded. The data were entered using Epi data version 4.2 statistical software in order to minimize error that occurs during data entry and exported to SPSS; and analyzed using SPSS version 23. Data cleaning was performed to check for frequencies, accuracy, and consistencies and missed values and variables. The finding of this study was presented using text, figure and tables form from the result of frequencies and crosstabs.
A bivariate logistic regression model analysis was done to see the association between the explanatory and outcome variables. Henceforth, multivariable logistic regression analysis was employed by selecting only variables with a P-value < 0.25 in the bivariate analysis. The odds ratio with 95% CI was used to measure the strength between dependent and independent variables at P-value < 0.05 to determine the level of statistical significance. Variables with the P-value less than 0.05 (i.e., 95% CI) in multivariate regression were declared to be potential predictors for suicidal ideation.

Sociodemographic characteristics
In this study, a total of 221 participants were invited and fully participated in the study providing a response rate of 100%. Out of them, 147 and 74 were controls and cases, respectively. Thirty-eight 38(25.9%) controls and 34(45.9%) cases were 18-27 years old. Sixty (40.8%) controls and 41(55.4%) cases were single. A hundred twentyfour (84.4%) non-suicidal participants and 57(77.0%) suicidal participants were Orthodox by religion, and A hundred thirty-six (92.5%) controls and 64(86.5%) cases were Tigray by ethnicity. Twenty-seven (18.4%) controls and 23(31.1%) cases educated college and above.  Table 2). Regarding the educational status, this study found that 112 educated non-suicidal participants and 60 suicidal participants that had an elementary school and above. Of these, 35(31.3%) respondents who had no suicidal ideation and 20(33.3%) respondents who had suicidal ideation had stress-related education. Out of 53 nonsuicidal participants and 31 suicidal participants that had a job, 24(45.3%) non-suicidal and 19(61.3%) suicidal had stress in their job. Thirty-nine (26.5%) and 26(35.1%) participants had a history of family/parent/death and were controls and cases, respectively. Among those who had a history of sexual/physical violence, 13(8.8%) and 13(17.6%) had no and had suicidal ideation, respectively. Forty-nine (33.3%) participants who had no suicidal ideation and 34(45.9%) participants who had suicidal ideation had a history of residential change. From those who had a conflict with who had a good relationship, 59(40.1%) and 40(54.1%) were non-suicidal respondents and suicidal respondents, respectively.

Patient's psychiatric disorders related characteristics
This study showed that diagnosis of patients, 43(29.3%) controls and 38(51.4%) cases were major depressive disorders. Fifty-seven (38.8%) controls and 20(27.0%) cases were schizophrenia. Participants that their diagnosis was psychotic disorder and/or mood disorders with psychotic feature that associated with suicide, 4(3.7%) controls and 11(21.6%) cases had a hallucination that tells/commands him/her to commit suicide. Participants whose diagnosis was either psychotic disorder or major depressive/bipolar I disorder with psychotic feature responded that they had no psychotic features associated with suicide were 102(95.3%) controls 36(70.6%) cases.
Regarding illness onset mode, 63(42.9%) controls and 31(41.9%) cases were their mode of illness onset and duration of illness was less than 3 months. Sixty-eight (46.3%) non-suicidal respondents and 35(47.3%) suicidal respondents, their duration of illness was 3 months to 2 years duration.

Suicide attempt-related characteristics
This study revealed that 4(2.7%) controls and 11(14.9%) cases had a family history of suicide. Only 2(1.4%) nonsuicidal respondents and 24(32.4%) suicidal respondents had current thought about killing oneself ); and 20(13.6%) non-suicidal participants and three-fourths, 56(75.7%) suicidal participants had previously suicide attempted history. From 24 suicidal respondents that have current suicidal ideation, 14(58.3%) had the intention to kill themselves via poison, fall from a hill, drown, and availability of lethal methods during that time.

Determinants of suicidal ideation Sociodemographic determinants
Marital status (being divorced or widowed) is 0.2 times less likely to have suicidal ideation than married (AOR = .20, 95% CI .046-.879). On the other hand, educational status, age, and alcohol were not significantly factors in this study.

Biopsychosocial determinants
This study showed that patients who had family mental illness history are 3.03 times more likely to have suicidal ideation than those had not family mental illness history (AOR = 3.029, 95% CI 1.088-8.431), P = 0.034) and having family suicide history is 6.87 times more likely to have suicidal ideation than participants without history of family suicide (AOR = 6.87, 95% CI 1.138-41.531, P = 0.036). Patients who had previous attempts history are 27.46 times more likely to have suicidal ideation than those patients did not attempt (AOR = 27.457, 95% CI 10.417-72.368, P = 0.0001). On the other hand, physical illness was not a significant factor in this study. Major depressive disorders include major depressive disorder, a major depressive disorder with a psychotic feature, and severe major depressive disorder with the psychotic feature. Bipolar disorders include bipolar I disorder, bipolar I disorder with the psychotic feature Other psychotic disorders include schizophreniform disorder and brief psychotic disorder and substance-induced psychotic disorder. Other psychiatric disorders include somatic symptom disorder GAD, PTSD, MDD with PTSD, schizophrenia comorbid with post-traumatic symptoms, schizophrenia comorbid with depression, SAD, GAD comorbid with SAD, GAD comorbid with MDD; and OCD Other F includes those who visit less frequency follow-up and send their relatives for follow-up

Psychiatric disorder determinants
This study revealed that major depressive disorders are 11.18 times more likely risk to have suicidal ideation than those patients their diagnosis is bipolar disorders (AOR = 11.182, 95% CI 2.382-52.501, P = 0.002). In addition to major depressive disorder, other psychiatric disorders (those who have comorbid diagnosis that is MDD comorbid with PTSD, PTSD, schizophrenia comorbid with PTSD, SAD comorbid with GAD, OCD, GAD comorbid with MDD, schizophrenia comorbid with depression) were 12.25 times more likely risk to have suicidal ideation than patients whose diagnosis is bipolar disorders (AOR = 12.245, 95% CI 1.898-78.986, P = 0.008). On the other hand, psychiatric disorders like schizophrenia, other psychotic disorder were not significant factors in this study (Table 7).

Discussion
This study showed that a family history of suicide attempt was a significant predictor of suicidal ideation. This finding is in line with others studies done in Ethiopia, Addis Ababa and Dessie [10,17]. Also, in this study, having a family mental illness history is a significant predictor of suicidal ideation. This finding was consistent with studies done in Eastern Nepal and Ethiopia, Jimma [13,18].
This study also revealed that a previous suicide attempt history was significant for suicidal ideation. This finding Table 7 Multiple logistic regression analysis of determinants of suicidal ideation of patients with mental disorders visiting psychiatry outpatient unit in Mekelle town, psychiatric clinics, Northern Ethiopia, 2019 (n = 221, control = 147, and case = 74) Italic values indicate variables which show significant determinants with suicidal thought at multivariate analysis MDDs include major depressive disorder, a major depressive disorder with psychotic features and severe major depressive disorder with a psychotic feature Other psychiatric disorders include GAD, MDD comorbid with PTSD, PTSD, schizophrenia comorbid with PTSD, social phobia, SAD comorbid with GAD, OCD, GAD comorbid with MDD, schizophrenia comorbid with depression, and somatic symptom disorder Other psychotic disorders include schizophreniform disorder, brief psychotic disorder, and substance-induced psychotic disorder. Bipolar disorders include bipolar I disorder and bipolar I disorder with psychotic features was consistent with findings in America; Malaysia; Ethiopia, Addis Ababa and Gonder [4,16]. This study showed that major depressive disorders were a significant predictor of suicidal ideation. This finding was in line with studies done in America; in America, Washington DC; China; Lithuania; Finland, Helsinki; Japan; Germany, Jena; Eastern Nepal; South Africa; and in Ethiopia, Jimma [4,6,8,9,11,12,[14][15][16]18].
This study also revealed that other psychiatric disorders (those who have a comorbid diagnosis that is MDD comorbid with PTSD, PTSD, schizophrenia comorbid with PTSD, social phobia comorbid with GAD, OCD, GAD comorbid with MDD, schizophrenia comorbid with MDD) were significant factors for suicidal ideation. This finding was consistent with findings in America, Washington DC; Germany, Jena; Ethiopia, Addis Ababa and Jimma [6,15,17,18].

Conclusion
Marital status, family mental illness, family suicide history, previous suicide attempt, major depressive disorders and other psychiatric disorders that are those comorbid were significant factors of suicidal ideation.

Recommendation
• Psychiatric professionals should assess patient suicidal risk assessment routinely and should put the diagnosis with suicidal if the client is suicidal so that every professional focuses on treatment besides the medication. • Educate the family/caregivers of suicidal patients with previous attempts, current suicidal ideation and intention to have closely followed up. • It is also recommended researchers to conduct further research studies.

Limitation of the study
Despite providing valuable baseline data, there are also some limitations encountered: • Case-control nature of the study design: unmatched case-control does not control confounding factors. • In this study, only adult psychiatry patients were included, so it is difficult to generalize all psychiatry patients because those who were unable to communicate, children and adolescents psychiatry patients; and decision incapacity consent were not included in the study. So, those excluded participants might be highly suicidal.