Suicidal Behavior and Associated Factors Among Inmates at Dessie City Correctional Facility, Dessie, Ethiopia | BMC Psychiatry

Study framework

The study was conducted at Dessie City Correctional Center in North Eastern Ethiopia from February 16 to March 5, 2020. Dessie City has one correctional facility, which was established in 1958. This correctional facility is located in South Wollo, Amhara Regional State, North Eastern Ethiopia, 401 km from Addis Ababa, the Ethiopian capital, and 482 km from Bahir Dar, seat of Amhara Regional State. The town contains 18 kebeles with a total population of 219,978 (99,822 male and 120,156 female), according to data from the South Wollo Area Statistics Office for 2016-2017. According to the 2015 global prison population data, Ethiopia’s population is incarcerated at 111,050. [25]. During the study period, there were approximately 1250 prisoners. There is only one clinic serving inmates in prison and it does not provide psychiatric services. Murder, physical assault, attempted murder, rape and kidnapping, robbery and burglary, and political concerns all result in the imprisonment of intimates.

study design

A cross-sectional study design in an institutional setting was conducted.


Source population

All adult inmates (18 years and older) of the Dessie City Correctional Facility.

Study population

All adult inmates at Dessie Correctional Institution were included in the sample during the data collection period.

Detainees who were seriously ill and unable to communicate at the time of data collection.

Eligibility criteria

In our study, the inclusion criteria were all inmates aged 18 and over, while seriously ill inmates unable to communicate during data collection were excluded from the study.

Sampling procedure and sampling techniques

Determination of sample size

To determine the sample size of the study population, the following assumption was made. The actual sample size for the study was determined using a single population proportion formula, assuming a 5% margin of error and a 95% confidence interval at alpha ( α = 0.05), and the proportion of the population (23.2%) was taken from a study conducted at Jimma Correctional Institution [17]. So, based on the above information, the total sample size was calculated as follows:

Thus, the sample size was determined as follows;

$$mathrm n=frac{left(mathrm Z;mathrmalpha;/2right)^2mathrm Pleft(1-mathrm Pright)}{mathrm d^ 2}=quadquadquadmathrm n=frac{left(1.96right)^2timesleft(0.232left(1-0.232right)right)}{left(0.05 right)^2}=274$$

Where n = sample size;

ᾳ = 95% confidence interval = 1.96.

p = prevalence of suicidal behavior 23.2% (study conducted at Jimma Correctional Institution).

d = marginal error = 0.05.

Therefore, the final sample size adding 10% non-respondents = 302.

Sampling technique

After the sampling frame was developed, a systematic random selection process was used to select study participants. The sampling interval (K) was obtained by dividing the study population by the final sample size as follows; 302 (=frac{N}{n} , k=frac{1250}{302}=4.02=) 4. Accordingly, the actual participant was randomly drawn every four intervals in the sampling frame until the desired sample size was reached. The first subject of study was chosen by drawing lots from a list of 1 to 4 candidates. Accordingly, from the first study unit, volunteers were selected at every four intervals.

Operational definitions

suicidal behavior

Incarcerated individuals who achieved an SBQ-R score ≥ 7 were labeled as high risk and those below 7 as low risk [26].

The Depression

Was assessed using the Hospital Anxiety and Depression Scale (HADS). Participants who scored > 8 were considered to have depression [27].


Using the Hospital Anxiety and Depression Scale (HADS), normal (score 7) [27].

Stressful life events

Defined as having experienced one or more stressful life events in the past year [28, 29].


Using the Oslo-3 scale, inmates with an average score of 8.1 (good support) [30].

Has ever used a substance

Use of at least one substance (alcohol, tobacco and khat) during lifetime [31].

Data collection method and instruments

Data were collected using an interviewer-administered questionnaire, which has four sub-units. Socio-demographic factors, clinical factors, psychosocial factors and substance-related factors were developed after an extensive review of the literature and similar study tools.

Suicidal behavior among inmates was measured using the Suicidal Behavior Questionnaire-Revised, which included the four items listed below (SBQ-R). Item 1 of the SBQ-R assesses lifetime suicidal thoughts and attempts, item 2 assesses the frequency of suicidal ideation in the past year, item 3 defines the threat of suicidal behavior and the Item 4 assesses the likelihood of future self-reported suicidal behavior. Its sensitivity and specificity were 80% and 91%, respectively, on a scale of 3 to 18. The presence of suicidal behaviors in community prisoners is explained if a score ≥ 7 for the SBQ-R [26]. Further, from question number one, it was categorized into suicidal ideation, plan/intent, and attempt for discussion purposes. Suicidal ideation (If the respondent answers the question, have you ever thought / briefly thought about killing yourself? If yes, the patient has suicidal ideation) [26, 32]. Suicidal plan/intention (If the respondent answers the question did you at least once intend to commit suicide? If yes, the patient has a suicidal plan/intention) [32]. Suicide attempt (defined as; if respondent answers question, have you ever attempted suicide? If yes, patient has attempted suicide) [32]. The internal consistency (Cronbach alpha) of (SBQ-R) in this study was 0.83.

The questionnaire contains variables used to assess dependent variables, including sociodemographic factors (age, gender, religion, ethnicity, marital status, level of education, occupational status, clinical factors (depression and anxiety) and psychosocial factors (support social, stressful life events, length of stay and use of a psychoactive substance [31].

The Hospital Anxiety and Depression Scale (HADS) is made up of two subscales, one of which assesses depression and the other anxiety. Respondents rated each item on a four-point scale (0 to 3), resulting in possible scores ranging from 0 to 21 for each of the two subscales [27]. A score of 0 to 7 is considered “normal” according to the HADS manual. In a study comparable to others, a threshold of >8 was applied for measures of depression and anxiety [33]. In this study, the internal consistency was 0.79 for the depression subscales and 0.81 for the anxiety subscales.

The Stressful Life Events Screening Questionnaire (SLESQ) is a self-assessment of exposure to one or more stressful life events. Events such as unemployment, marital difficulties, spouse, child or parent’s death, etc. For each event, respondents are asked to indicate whether the event occurred by (“yes” or “no”) [28, 29]. In this study, the internal consistency (Cronbach’s alpha) of stressful life events was 0.81. The Oslo-3 scale was used to measure social support [30]. In this study, social support is measured based on the average score of the participants. Inmates with an average score of 8.1 or lower received poor social support, while those with a score of 8.1 or higher received good support. The internal consistency (Cronbach alpha) of Oslo-3 social support in this study was 0.87. The internal consistency (Cronbach alpha) of Oslo-3 social support and stressful life event in this study was 0.87 and 0.81 respectively.

Data quality control

Data were collected through face-to-face interviews by three trained psychiatric nurses and a supervisor from Integrated Clinical and Community Mental Health (ICCMH). The questionnaire was properly prepared and edited to ensure quality, and it was translated into the local language (Amharic) so that all participants could understand it, and then translated back into English. Data collectors and supervisor received two-day training on the purpose of the study, tools, how to collect data, sampling techniques, and how to handle ethical issues, including confidentiality . Pre-tested on 5% of the Kombolcha City Correctional Center one week prior to the main data collection to identify potential issues in the proposed study, such as data collection methods and data collector performance, and was not included in the main study. Data collectors were supervised regularly, and supervisors and the principal investigator reviewed field questionnaires daily. The data collected was edited and entered into the computer from a paper then double checked and processed in a timely manner.

Data processing and analysis

Data were collected, cleaned, and stored for consistency on a computer using Epi-Data version 3.1, then exported to SPSS statistical software version 26 for analysis. The researchers used frequency, proportion and other descriptive statistics. To identify the independent factors associated with suicidal behavior, binary logistic regression was used to determine the association between the explanatory factors and the outcome variable, then all the independent variables with a p– values ​​less than 0.25 were entered in the final model (multivariate logistic regression). Hosmer Lemeshow’s quality test was used to assess the fitness hypothesis. AORs with 95% CI were used to measure the strength of association. Finally, variables with a P– value