Suicide Ideation Intervention in Nigeria

The effect of comprehensive suicide behavior prevention intervention among in-school adolescents Nigeria: a stepped-wedge cluster randomized trial

This study aims to evaluate the feasibility and effectiveness of a theory-based comprehensive suicide behavior prevention intervention among in-school adolescents in the Ikenne local government area (LGA) of Ogun State, Nigeria. The methodology of the study is a Stepped-wedge cluster-randomized trial of a suicide behavior intervention among fifteen secondary schools in Nigeria. The study chose the stepped-wedge design because although the investigators expect the intervention to reduce suicidal ideation, the feasibility of a full implementation may be challenging due to political, logistical, and ethical constraints. The design fulfills a secondary role of serving as a scientific tool to incorporate a fair method to determine the order of intervention rollout under logistic impediments. The implementation of our intervention is feasible only at the cluster level.

Investigators will conduct a Stepped-Wedge Cluster Randomized Trail among 15 secondary schools over one year. The participating schools will cross in a one-way switch over from the control to the intervention period. The sequence of switch over for the clusters will follow a stratified randomization scheme. All the schools will receive the intervention by the end of the study. All the groups will be recruited at the study onset and will act as control until the point of randomization into the intervention group. To mitigate contamination, the study will provide clear information on the purpose of the research to the participants. The clusters will get the details of intervention at the point of switching into the experimental group. The study will educate the intervention group against contamination. Also, the investigators will assess the exposure of the control group to the intervention before switching them. The study will report the level of contamination.

The study will recruit 15 secondary schools with a minimum of 100 enrolled students and do not have an existing suicide behavior intervention program. The school boards need to approve participation. Eligible students should be aged 10-19 years, give consent, and not have plans of leaving the school in the next year. Parental/guardians' assent is also a prerequisite for participation in the study. The study will not exclude non-participating students from the delivery of the intervention. The study will randomly select, by balloting, Five (three publicly and two privately-owned) schools in each of the three areas of the LGA. The selection of students will follow a probability sampling method. Eligible schools will be randomized into three sequences of five schools each, using a computer-generated stratified randomized sampling scheme. The stratification, which will be according to the area and ownership of the schools, will be supervised by the PIs. The nature of the experiment and intervention makes blinding unrealistic. However, investigators will ensure the conduct of the measurement of the study outcome objectively.

The intervention is a four-pronged comprehensive suicidal behavior prevention intervention in schools. The prongs consist of evidence-based interventions that use the Social Ecological Model (SEM). The SEM theorizes that there are multiple inter-related levels of influences on human behavior, and a successful preventive program should address factors at each stratum. The intervention addresses three levels; individual, interpersonal, and social. Firstly, the research will collaborate with relevant stakeholders to design and obtain a commitment from the schools to a suicidal behavior prevention policy and protocol document. Secondly, the intervention will train staff and students on best practices to improve students' connectedness thereby, creating a positive school environment. Thirdly, the study will identify and train gatekeepers on the proper response to suicidal thoughts and other suicidal behaviors; and serve as the link between the students and a school-linked mental health specialist. The research will train the gatekeepers to use school-based screening tools. Lastly, this research will educate all students on suicidal behaviors.

This prong will create awareness, promote case-finding, and provide information about available mental health resources. Also, the intervention will address the stereotypes that promote stigma and improve the students' coping abilities. The entire intervention will incorporate peer support and student involvement at all stages of planning and implementation.

The primary outcome is suicidal ideation, as assessed by the Positive Inventory of the Positive and negative Suicide Ideation (PANSI-PI). PANSI is a 14-item scale consisting of two subscales, the 6-item positive scale (PANSI-PI) and the 8-item negative (PANSI-NSA). Higher scores on PANSI-NSI scale and lower marks on PANSI-PI) symbolize the higher frequency of suicidal thoughts. The secondary outcomes are grit, gratitude, and mental health status. The study will use the gratitude questionnaire(GQ-6), the grit scales, and the General Health Questionnaire (GHQ-12) to assess gratitude, determination, and mental health status, respectively.

The sample size was determined using the Shiny CRT Calculator, a power and sample size tool for cluster-randomized trials. The aim was to have three sequences with five clusters each, assuming an intra-cluster correlation coefficient of 0.01 (0 to 0.02) and individual autocorrelation of 0.08. The study desires a minimum statistical power of at least 90%, considering a type I error of 5% and a 10% increase in the mean PANSI-PI score from 17.71 (_6:35) attributable to the intervention. Some allowance will be made for a school dropout rate of 23.12%. The derived sample size was 30 per cluster.

The study will analyze the data using Stata/SE 15.1 statistical software. The statistical analysis is rooted in the principle of intention-to-treat and will use random effect mixed modeling to address missing data while assuming that missingness is at random. The research will use the linear mixed model for continuous outcomes and the generalized linear mixed model for categorical outcomes and adjust for the possible confounding effects of individual-level characteristics that our randomization did not capture. Also, the study will adjust for age, sex, number of years in the school, ethnicity, and secular trends (the school term), and will use multilevel analysis to adjust for cluster effect and the correlation between repeated measurements.

The expected outcome of the research includes a suicidal behavior prevention policy and protocol template for secondary schools in Nigeria and an effective suicide behavior prevention program that is scalable for Ogun State, Nigeria.