Development and validation of the Redeemer’s University Suicidality Scale

Background The need for a culturally suitable scale for suicidality within the multilingual Nigerian society necessitated this research interest. Aim The study is a development and validation of the Redeemer’s University Suicidality Scale (RUSS). Setting South western Nigeria. Methods This comprised of initial generation of items; face and content validity, item refinement and administration of RUSS to 150 university undergraduates, using exploratory factor analysis at the first, second and third stages. In the fourth stage, 184 undergraduates responded to the 20-item RUSS, Suicide Ideation Scale (SIS) and General Health Questionnaire (GHQ-12). Data gathered at this stage were analysed for congruent validity, reliability and norms. Results The principal component analysis extracted four components from items whose eigenvalues exceeded one. Twenty-one of the 25 items loaded best in the first, two in the second and one on the third component(s). Only items in the first component were retained. Item-total correlation further showed that the values of one item in the first component fell below the very good discrimination and was deleted from the scale. The RUSS has a Cronbach’s alpha of 0.93. Congruence validity coefficient of r = 0.881 (p < 0.001) and r = 0.605 (p < 0.001) was observed between RUSS and SIS and between RUSS and GHQ-12, respectively. Conclusion The RUSS is gender-sensitive, has acceptable psychometric properties and is recommended as a diagnostic tool for assessing suicidal behaviour in adolescents and adults. Contribution This article contributes to the development of a culture sensitive measure for suicidality.


Introduction
http://www.sajpsychiatry.org Open Access Suicidal behaviour encompasses a variety of behaviours that have the same goal of killing oneself. Suicidal behaviour, however, differs from non-suicidal self-injury, which is a deliberate destruction of one's own bodily tissue with no lethal intent, although the two behaviours might lead to dying.
Suicidal ideation, planning and attempts are three essential suicidal behaviours. Suicidal ideation occurs when a person considers or plans to commit suicide. 3 Suicidal ideation encompasses a wide spectrum of behaviours, from fleeting thoughts to long-term plans. Suicidal ideations, often known as suicidal thoughts or ideas, is a broad term that refers to a variety of thoughts, aspirations and preoccupations with death as well as the inclination or desire to terminate one's own life. 9,10 Suicidal ideation can range from passive (wanting to die) to active (wanting to kill oneself or thinking of a specific method for doing it).
A suicide plan is defined as 'a proposed technique of carrying out a design that may result in self-injury; a systematic development of a plan of action that may result in selfinjury'. 11 Finally, attempting suicide is a planned act of ending one's life. It is a self-inflicted, possibly harmful behaviour with a nonfatal outcome for which there is evidence of a desire to die (either explicit or implicit). 11 Overdosing or ingesting is the most common means of suicide attempt among teenagers, followed by hanging or suffocation and using a sharp object (e.g. cutting). 12,13 Different theories have been used to explain suicidal behaviour. One of them is the ideation-to-action framework proposed by Klonsky and May. 14 According to this model, the development of suicidal ideation and the progression from suicidal ideation to attempts should be seen as distinct processes having distinct predictors and explanations. The ideation-to-action framework sheds more light on what the traditional approaches (psychache, social isolation, escape from aversive selfcognitions, hopelessness and so on) could not explain (i.e. offering different explanations for the development of suicidal ideation and the progression from ideation to attempt). The traditional differs from the ideation-to-action model in that the former treats suicide as a risk and a unitary construct. At the same time, the latter distinguishes ideation predictors from the progression from ideation to behaviour.
The ideation-to-action framework, which specifies: (1) the development of suicidal ideation and (2) the transition from ideation to suicide attempts, is one perspective in the study of suicide. Suicidal ideation, suicide attempts and suicide are all included in this framework, 15 with suicidal ideation being the first phase of suicide. As a result, any self-initiated or committed activities with the aim or expectation of dying, including active or passive self-inflicted acts, are considered suicide. 16 Another significant advancement is a growing collection of research distinguishing between characteristics that predict suicide thoughts and those that predict attempted suicide. 17 According to studies, most suicidal behaviours occur when people are thinking about it. 18 Suicidal behaviour is influenced by various biological, psychological, social, cultural and spiritual variables.
Suicidal ideation is defined as a desire or consideration of suicide. 19 In the literature, there are two types of suicidal ideation (passive and active). Passive suicidal ideation happens when someone aspires to die or believes they might die but has no intention of actually doing so. Active suicidal ideation, on the other hand, refers to the act of thinking about, intending to and planning to commit suicide. 19 Suicidal ideation is a significant risk factor for suicide in adolescents and young people. 20,21 Suicidal ideation has been reported by almost a third of teenagers aged 12-20 years in the United Kingdom. 22 Suicidal behaviour is common among those suffering from mental illnesses such as depression, schizophrenia, bipolar disorder and substance misuse. 19 However, it is not always associated with an underlying disease or a genuine desire to die. 23 Despite the global rise in suicide rates, Rukundo et al. 24 found a data vacuum in suicidal behaviour among children and adolescents in poor and middle-income African nations.
Like many other countries, Nigerian research shows an alarming rate of suicide deaths. 25,26 For instance, according to 2019 statistics, there is a 3.5 suicide rate per 100 000 Nigerians. 27 Based on an estimated 2021 population of over 211 million people, the death by suicide figure in 2021 was about 7400. 27 Most suicide deaths occur in underdeveloped nations such as Nigeria 17 and are often undetected 28 and grossly unreported. 6,29,30 According to the World Health Organization's (WHO) 2018 rating, 3 Nigeria has the highest suicide rate in Africa and ranks fifth in the world in terms of acute suicide cases, trailing only South Korea, Russia, India and Japan. 31

Justification for the study
The rate of death by suicide is high in Nigeria. 3

Study sample
The sample size calculation was based on the table of sample size determination published by Glenn. 34 For a sample size for ± 10% precision levels where confidence level is 95%, and p = 0.5 for 4000, a population sample size of 98 was calculated. 34 Based on a previous pilot study conducted using 180 participants, there was a 35.9% prevalence of suicidal behaviour. To account for attrition, a sample size of 150 and 184 participants who consented was used for the exploratory factor analysis (EFA) and the determination of the psychometric properties of the Redeemer's University Suicidality Scale (RUSS), respectively. For the inclusion and exclusion criteria, only registered students who were university undergraduates and currently on full-time study bases, found within the selected university campus, were selected for the study. This inclusion was meant to provide valid research outcomes by controlling for possible error in the responses caused by students who were not undergraduate students or students who were not on a full-time study basis but might be around the university campuses at the time of instrument administration. Also, only students willing to answer and return the questionnaires to the researcher were included in the study. This inclusion criterion enabled the researcher to ensure that all the prospective participants were well enlightened and understood the purpose of the study; as such, motivation and true responses were guaranteed.
Studies at the international 35 and Nigerian levels 36,37 show that the age category 18-29 years is reported to have the highest prevalence of suicidal behaviour. Also, suicide is more prevalent among undergraduates, 32 hence choosing this population for the study. Previous studies have suggested being young as a risk factor for suicidal behaviour. 38

Study setting
This study was carried out among undergraduate students at a private university in south-western Nigeria.

Study instruments
The participants responded to the RUSS, the Suicide Ideation Scale (SIS) 39 and the General Health Questionnaire 12 (GHQ-12). 40 The SIS, originally developed for and validated with nonclinical samples of young adults, was designed as a brief measure of ideation for use among clinical and nonclinical populations. The SIS is a 10-item self-report scale designed to assess the severity or intensity of suicide ideation. Responses on the SIS are scored on a Likert-type scale with anchors at 1 (never or none of the time) to 5 (always or a great many times) based on how the respondent has felt or behaved over the past year. Items in the SIS are: Items were scored by direct scoring, and the total score was calculated by finding the sum of 10 items. The total score ranges from 10 to 50. Rudd 39  The GHQ-12 was designed to assess psychological distress. 40 The 12 items of the GHQ-12 are scored on a 4-point severity and frequency scale (0-3). Scores of the items of the GHQ-12 are added to derive the total score of psychological distress. The GHQ-12 has acceptable psychometric properties for the Nigerian population. 43 As a measure of psychological distress, 40 the GHQ-12 was used to validate suicidality, which is also a manifestation of psychopathology. The researchers collected data for the study.

Statistical analysis
Data were analysed using the Statistical Package for the Social Sciences (SPSS) version 23 (IBM Corporation, Armonk, New York, United States). Factor analysis was used for the exploratory factor analysis (EFA). Reliability analysis was used to find out the Cronbach's alpha, Guttman splithalf and Spearman-Brown coefficients. Finally, Pearson correlation analysis was used to determine the congruence validity of the scale under development. Thirty-seven items were generated using three of the sources recommended by Furr (2011) and Streiner (2015). These sources are an examination of items in some instruments that measure suicide, a review of theories of suicide for potential ideas and research findings of previous studies. Clinical features of depression and suicidality in the DSM-5 were also considered. 44,45,46 Validity of the Redeemer's University Suicidality Scale

Item generation for the Redeemer's University Suicidality Scale
The generated items were subjected to both face validity and content validity assessment by a panel of experts made up of one psychiatrist and six clinical psychologists. The experts reviewed items for content regarding their relevance to the content domain, completeness and accuracy. They also reviewed for item sensitivity to evaluate potential item bias. 47 The items were evaluated using the content validity ratio (CVR) of a four-point scale of 1 = not relevant, 2 = somewhat relevant, 3 = quite relevant but needs rewording and 4 = highly relevant. 48 All items checked as 'not relevant' were eliminated. Using 75% item inclusion (i.e. 75% of the experts must approve an item for inclusion), those checked as 'needs rewording' were reworded, and items checked as highly relevant were retained. Thirty items were retained after CVR. Lastly, the 30 items were re-presented to the experts for review on a two-point response format of 'yes' and 'no'. Items checked as 'yes' were included while those checked as 'no' were removed. Using the 75% item inclusion, 25 items were retained. The justification for employing the face validity method was that the expert technique is acceptable for content validity 49 when combined with a comprehensive identification process as outlined above. This process resulted in a final version of 25 valid items used for item refinement.

Item refinement
One hundred and fifty university undergraduates (male, 79; female, 71) were purposively drawn from a Nigerian private

Exploratory factor analysis
According to Pallant 51 for factor analysis to be considered appropriate, Bartlett's test of sphericity (BTS) should be significant (p < 0.05), with 0-1 range for the Kaiser-Meyer-Olkin (KMO) index and 0.06 set as the minimum value for suitable factor analysis. The results of the tests are presented in Table 1.     Table 3.
The four components extracted summarised in Table 3 were because the items loaded on these four components with eigenvalues exceeding 1. The eigenvalues of the four components range from 14.018 to 1.021, with a percentage ranging from 51.918 to 3.782. However, only one dimension was retained, as items in the other three components either loaded more than once, rendering them complex structures or standalone items. The eigenvalues of the included component explained a total of 51.9% of the total variance. The RUSS was validated using the congruent validity technique to ascertain its relationship with existing measures: the SIS 39 and the GHQ-12. 40 Significant positive correlation coefficients were observed between the RUSS and the SIS (r = 0.881, p = 0.000) and between the RUSS and the GHQ-12 (r = 0.605, p = 0.000). This finding made the RUSS valid as a diagnostic tool for measuring suicidality among the general Nigerian population.

Reliability of the Redeemer's University Suicidality Scale Participants
A fresh sample of 182 university undergraduates (male, 89; female, 93) was purposively drawn from a Nigerian private university. The participants' ages ranged between 16 and 31 years (mean = 19.72; s.d = 3.61). A cross-sectional method was used to select the participants from all levels of study in the seven faculties in the chosen institution. The responses of these participants were used to ascertain the reliability of the scale. The 20 items of the RUSS that were extracted from EFA in the item refinement process were administered to the participants. Data were analysed using the SPSS.

Ethical considerations
This study relied on human subjects for its investigation. As a result, the Helsinki Declaration was followed in terms of research ethics for human beings.

Results
Values of the corrected item-total correlations (point-biserial) indicated discriminations in the items of the RUSS. Values between 0 and 0.19 imply poor discrimination, 0.2 and 0.39 indicate good discrimination, while ≥ 0.4 suggests very good discrimination. As observed in Table 4, the item with a value between 0 and 0.19 in the RUSS is 'I do not think killing myself   Table 4 further shows the Cronbach's alpha and the item-total statistics of the 20-item RUSS. The scale achieved a reliability coefficient of 0.93, a Guttman split-half reliability of r = 0.825 and a Spearman-Brown coefficient of r = 0.828.

Calculation of norms for the Redeemer's University Suicidality Scale
The cut-off points for the RUSS were determined using the 95% confidence interval (CI) method. As summarised in Table 5 The final draft of the validated RUSS and the scores' interpretations are itemised in Table 6 and Table 7, respectively.
Item 9 (with ' †' symbol) is reverse scored. The total score is arrived at by adding up the items' scores. The higher the score, the greater the severity of suicidal behaviour. Table 7 is a summary of the interpretation of the scores based on group and individual samples. The individual samples are categorised by gender.

Discussions
This study is a development and validation of the RUSS. The scale measures suicidal behaviours (suicidal ideation,     56 proposed seven steps, while DeVellis 57 proposed eight steps. Four main features are common in most of these studies: item generation, item refinement, reliability studies and validity studies. The development process of the RUSS followed these steps. The DSM-5 and the ICD-10 were used to review clinical aspects of suicidal behaviour (ideation, intent and planning). 44,45,46 The generation of items relating to the agreed themes resulted in 37 items used for scale purification purposes. Because of the RUSS's ability to measure opinions, beliefs and attitudes, a six-point Likert scale was utilised. In other words, each RUSS item is a declarative statement, hence the decision to employ a Likert answer format. 57 The RUSS purification was carried out using a combination of reliability analysis and EFA, as proposed by Flynn and Pearcy 58 and Pecheux and Derbaix. 59 A team of specialists checked the content validity of the initial items generated by the authors. According to Streiner et al., 48 content validity represents current knowledge in the construct of interest. It is also a vital sign of an instrument's validity and shows how viable and practicable it is. 60 The development of the RUSS provided a foundation for further investigation of its validity and reliability. Cronbach's alpha was 0.93, and item-total correlation was 0.13-0.79. This finding implies a good item inter-relatedness, unidimensionality and homogeneity of the construct 61,62 among the Nigerian population. To put it another way, the Cronbach's alpha, Spearman-Brown coefficient and Guttman split-half coefficient scores were not too high to make some items redundant. 63,64 In conclusion, the high alpha score indicates that the RUSS has a high level of reliability. The RUSS was verified using the concurrent validity approach, as Cronbach and Meehl 65 advised. Two standardised scales, one each for assessing suicidal ideation and psychological distress in the general population, were positively linked with the RUSS. The RUSS is a suitable measure of suicidal behaviour for adolescents and adults in Nigeria and other climes with similar sociocultural circumstances, based on its EFA and acceptable psychometric qualities.

Conclusion and recommendations
The RUSS was created by extracting a single-factor scale with 20 items through initial item generation, expert appraisal (face and content validity) of the initial pool of items and EFA for item purification. Internal consistency (reliability coefficient) of the RUSS items is acceptable. The SIS and the GHQ-12 had strong positive correlations with the RUSS, indicating a satisfactory congruent validity coefficient.
Based on the findings of this study, the 19-item RUSS showed good internal consistency and validity scores to measure suicidal behaviour. This analysis indicates that the RUSS is reliable and valid for the Nigerian population. Norms for the group and individual (male and female) samples are indications that the scale is gender-sensitive and can be selfadministered both individually and in group studies. Therefore, it is recommended as a diagnostic tool for suicidal behaviour among adolescents and adults in clinical settings. It can also be used to measure suicidality in group research settings to provide information on suicidal behaviour among the general public to enable policymaking in mental healthcare domains. Again, taking into cognisance the Nigerian sociocultural setting in its item generation and development, the RUSS is recommended not only for Nigerian use but also for use in countries with similar sociocultural situations.

Limitations of the study
This study was conducted in the context of the Nigerian population's distinct psychosociocultural environment. Without scale revalidation, generalising the findings and using this scale on other populations with different sociocultural traits should be addressed with caution.