Identifying potential catalysts to accelerate the achievement of Sustainable Development Goals (SDGs) among adolescents living in Nigeria

ABSTRACT

Investing in adolescents in Africa holds great promise for the development of the continent. The steps involved in identifying factors linked to interventions that may accelerate the attainment of multiple SDGs for adolescents in Nigeria are described. Data from a survey to investigate the well-being of 1800 adolescents aged 10–19 years in Southwest Nigeria was analysed. A four-step process was employed: 1) Mapping of variables deemed as suitable proxies for SDG targets; 2) Mapping hypothesised protective factors (accelerators) from the study instruments. Consequently, SDG targets related to elimination of hunger, good health, gender equality and peace; and seven accelerators (safe schools, parenting support, good mental health, no survival work, food security, stable childhood, and regular physical activity) were identified; 3) evaluating associations using bivariate analysis and multivariable logistic regression, 4) calculating adjusted probabilities. The mean age of the adolescents was 15.02 ± 2.27 years (48.6% female). Good mental health, not doing survival work, safe schools, stable childhood and parental support were significantly associated with at least two SDG targets. For example, food security was significantly associated with the highest number of SDG outcomes: one SDG target related to child survival (no substance use: x² = 3.39, p = <0.001); three SDG targets related to educational outcomes (school progression: x² = 5.68, p = 0.017, ability to concentrate in school: x² = 26.92, p = <0.001, and school attendance: x² = 25.89, p = <0.001); and four SDG targets related to child protection (no risky sexual behaviours: x² = 16.14, p = <0.001, no perpetration of violence: x² = 15.74, p = <0.001, no community violence: x² = 39.06, p =<0.001, and no sexual abuse: x² = 7.66, p = 0.006). Interventions centred around good mental health, not doing survival work, safe schools, small family size, stable childhood and parental support are potential accelerators for the attainment of SDG outcomes by adolescents living in Nigeria.

KEYWORDS:

• Nigeria
• adolescents
• accelerators
• SDGs
• health
• well-being

1. Introduction

The Sustainable Development Goals (SDGs) aim to bring about a better world through promoting human rights, improving health and quality of life, and eradicating poverty, while safeguarding the natural environment (UN, 2015). They were launched in September 2015 and received commitments from governments worldwide to implement an ambitious agenda towards the attainment of 17 goals over the next 15 years. Two years after this commitment by world leaders, UNICEF assessed the progress countries were making toward achieving the SDG targets. This preliminary analysis showed that up to 75–80% of child-and-adolescent-relevant indicators in each country (Nigeria inclusive) either had insufficient data or insufficient progress to meet SDG global targets by 2030 (UNICEF, 2018a).

The mainstreaming, acceleration and policy support (MAPS) approach proposed by the United Nations Development Programme (UNDP) provides a roadmap for achieving these SDG targets through multidimensional and reinforcing interventions (UNDP, 2018). The acceleration component of the MAPS moves a step away from utilising single interventions impacting single SDG targets to the use of interventions that have catalytic effects across multiple interlinked SDG targets (UNDP, 2018). These interventions or programmes that trigger the progress across multiple SDG targets, are called accelerators. However, employing a one-size-fits-all model may not be practical, as programmes that have been shown to be effective in a specific context might be ineffective in another setting. The UNDP, therefore, emphasizes that accelerators should be country- or context-specific.

Nigeria, the most populous country in sub-Saharan Africa (SSA), is home to over 42 million adolescents aged 10–19 years, representing 22% of the population (UNICEF, 2018b). Unfortunately, the potential of adolescents living in Nigeria is threatened by diverse adversities. For instance, in Nigeria, 25.8% of adolescents aged 12 to 17 years do not have access to education; approximately 12–25% of in-school adolescents experience different levels of depression; 3 out of 4 children and adolescents face multidimensional poverty (Fatiregun & Kumapayi, 2014; Oderinde et al., 2018; Omigbodun et al., 2004; UNICEF, 2019) and Nigeria has the highest rate of child labor in the World. These difficulties hold back progress for achievement of the SDGs. Policies and initiatives that will accelerate adolescent achievements of SDGs across multiple domains have the potential of leading to sustained economic growth (UNDP, 2017).

The main aim of this paper is to provide the methods employed in identifying potential accelerators (and preliminary results obtained) from existing data on school-going adolescents in South-west Nigeria collected in the pre-SDG era. The findings of the more robust multivariate regressions and marginal effect modelling – to determine ‘definitive accelerators’ and their synergy – as well as policy implications will be detailed in a future manuscript.

2. Materials and methods

2.1. Participants and data collection procedure

The data used for this study was collected as part of a cross-sectional survey (Ibadan School Health Survey) conducted in 2004 to investigate the health of 1873 in-school children and youths aged 9–25 years in rural and urban settlements in Ibadan (Omigbodun et al., 2008, 2010; Yusuf et al., 2011, 2015). Data pertaining to 1800 adolescents aged 10–19 years was extracted from the source dataset between November and December 2020.

Ibadan, located in the Southwest region (Oyo state) of Nigeria is the third-largest metropolitan city by population in the country, and the largest indigenous city in SSA by geographical size. The city is made up of six urban settlements, which make up the urban core, and there are five rural settlements described as ‘greater Ibadan’. In 2016, the urban core had a population of 3,160,200 with children and adolescents constituting about 40% of the community (Population.city, 2016). Statistics from the Nigerian Ministry of Education in 2016 for Oyo state, showed that the net basic enrolment rates for junior and senior secondary schools were 54.55% and 47.23% respectively (Federal Ministry of Education, 2016). The Nigeria education system is based on the 6-3-3-4 formula (Omowubi, 2019). Primary education begins around age 5 for most Nigerians and children spend six years in primary school. Students spend another six years in secondary school comprising three years of junior secondary school (JSS1-JSS3 … equivalent to grades 7–9) and three years of senior secondary (SSS1-SSS3 … equivalent to grades 10–12). Students then undergo a minimum of 4 years of tertiary education (Omowubi, 2019).

In the survey, 1873 students at all levels (JSS1-SSS3, which is equivalent to grades 7–12) in secondary schools in 22 schools in urban and rural settlements in Ibadan City were interviewed. Five districts (3 urban and 2 rural) were selected using random sampling. Using proportionate probability, 20 schools were then selected from a list of schools in the five districts. Students were then selected in the schools depending on: the number of students in the school, total number of students in the class and male to female ratio. The detailed procedure for the study has been reported by Omigbodun et al. (2008).

Four instruments were used for data collection: A Sociodemographic questionnaire, The Nigerian version of the Global School Health Questionnaire, The Culture-Free self-esteem questionnaire, and The Youth Diagnostic interview schedule for Children Predict Scales (Youth DPS), Version 4.32 (See Appendix 1 in supplementary material). The nutritional status of the participants was also assessed by standardized measurements of weight and height.

2.2. Procedure

The following procedure was employed in the exploration of the secondary data.

2.2.1. Step 1 mapping of SDG targets

The study instruments were reviewed, and items deemed suitable as proxies for measuring SDG targets were identified. This was achieved through a consensus of opinions following discussions held by RT, KK, TB, OO, AA, ET, LH and OO on items on the questionnaire that were similar in construct to individual SDG targets.

Eight constructs/variables on the questionnaire were identified as proxies for measurement of 8 SDG-aligned outcomes as follows: See Table 1a & 1b for details

Target 2.2, end all forms of malnutrition, measured as Body Mass Index (BMI) for Age

Target 3.5, prevention of Substance use, measured as no self-reported use of alcohol or tobacco in the past 30 days

Target 4.1, all girls and boys complete primary education, measured as appropriate age for class based on the Nigerian Ministry of Education recommendations

Target 4.4, increase relevant skills for employment, measured as no report of school absenteeism in the past 30 days and/or no report of not being able to concentrate (in the past 12 months)

Target 5.2, universal access to sexual and reproductive health, measured as no onset of sexual intercouse on or before the age of 14 and/or use of condom during the last sexual encounter

Target 16.1, reduce all forms of violence and related deaths, measured as no report of shoplitting, snatching a purse, vandalising or burglary (target 16.a) and no report of physical attack (target 16.b) in the last 12 months

Target 16.2. End violence against children, measured as no report of rape and/or fondling of body parts.

Table 1. a. SDG-aligned outcomes, questions used and operationalised measures. b. SDG targets, questions used and operationalized measure.

SDG Target	Questions/variables used	Operationalised measure used in this study	Measure
3.4 End all forms of malnutrition	Anthropometry measurements: Height and Weight	BMI for Age Stunting	No undernourishment
3.5 Prevention of substance use	-During the last 30 days how many times days did you have at least one drink containing alcohol? -During the past 30 days, on how many days did you smoke cigarettes or any other form of tobacco, such as snuff?	No self-reported use of alcohol, or tobacco within the past 30days	No substance use
4.1 All girls and boys complete primary education	Age and class	Appropriate age for class based on the Nigerian Ministry of Education recommendations	Normal school progression
4.4 Increase relevant skills for employment	During the past 30 days, on how many days did you miss class or school without permission? Have you often had trouble concentrating or keeping your mind on what you are doing for more than a short time?	No self-reported school absenteeism within the past 30 days No self-reported history of not being able to concentrate (in the past 12 months)	Optimal school attendance Ability to concentrate in school
5.2 Universal access to sexual and reproductive health	i) Age at onset of sexual intercourse ii) The last time you had sexual intercourse did you or your partner use a condom?	No self-reported onset of sexual intercourse on or before the age of 14 and/or Use of condom during last sexual encounter	No risky sexual behaviours
16.1 Reduce all forms of violence and related deaths. 16.1a No violence perpetration 16.1b No community violence	i) Have you shoplifted, that is, stolen from a shop when you thought no one was looking? ii) Have you snatched someone else’s purse iii) Have you broken or spoilt some place on purpose? iv) Have you stolen from someone when they were not looking? v) Have you broken into a building, house or car? In the past 12 months how many times you were physically attacked?	No self-reported history of shoplifted or snatching someone’s purse or vandalizing or burglary (in the last 12 months) No self-reported history of being physically attacked in the past 12 months	No self-perpetration of violence No community violence
16.2 End violence against children	i) Has an adult ever forced you to have sexual intercourse with them? ii) Has an adult ever touched breast, private parts or made you touched parts of their private parts?	No self-reported history of rape and/or fondling with body parts	No sexual abuse

Drawing lessons from work carried out by UNICEF on the ‘progress for every child in the SDG era’, this study arranges the SDG targets into three dimensions of the Child Rights namely: survive + thrive, learning and protection.

2.2.2. Mapping of accelerator protective factors

Regarding the accelerator protective factors (accelerators), questions that closely represented practices for which interventions could be implemented were identified from the study instruments. This selection was based on available evidence in the literature on practices or interventions that affect the social, biological and psychological well-being of adolescents in Africa (Ashimolowo et al., 2010; Cluver et al., 2019; Ijadunola et al., 2015; Shapiro & Oleko Tambashe, 2001) and a consensus of opinions following discussions held by RT, KK, TB, OO, AA, ET, LH and OO. Eight potential accelerators were subsequently identified.

i) Food security (Yes or No)

The following question from the global school global health questionnaire was used: During the past 30 days how often did you go hungry because there was not enough food at home? This was rated as (never, rarely, sometimes, most of the time, always)

‘Never’ and ‘rarely’ were classified as food security while ‘sometimes’, ‘most of the time’ and ‘always’ were classified as food insecurity.

ii) No Mental health issues (Yes or No)

This was measured as the absence of a diagnosable mental health disorder namely;depression, oppositional defiant disorder, low self-esteem, and suicidal attempts.

iii) Safe schools (Yes or No)

During the past 30 days, how were you bullied most often? See the full description of the different questions asked to assess bullying in Appendix 2 in supplementary materials.

Safe school was defined as the absence of a self-reported history of any form of bullying within the last 30 days.

iv) Child not doing survival work (Yes or No)

Do you do any kind of work to earn money before or after school? (Yes, No)

v) Stable childhood (Yes or No)

Children who lived with or who were brought up by both parents were defined as having had a stable childhood.

vi) Regular physical activity (Yes or No)

Regular physical activity was defined as reporting engaging in at least 60 minutes of exercise for at least three days in a week (Physical Activity, n.d..).

vii) Parental support

Three questions (measured on a Likert scale of never, rarely, sometimes, most of the times and always) from the protective factor module of the GSHQ were used to assess parental support. The responses were translated to a Likert scale of 1 to 5 with ‘never’ accorded the lowest value (1) and always accorded the highest value (5). An additive scale was then derived by summing up the responses for the three (3) questions. The maximum score a participant could record on the scale was fifteen (15) while the minimum was three (3), with higher scores indicating a higher level of parental support. See the full description of the different questions asked to assess parental support in Appendix 2 in supplementary materials.

2.2.3. Step 3: content validity of variables measured using different items

One potential accelerator (parental support) was computed using a set of related items (three questions) on the questionnaire, which were measured on a Likert scale. The emergent scale was further subjected to a factor analysis to ascertain if the component items were an accurate measure of the construct under investigation. The three questions showed high loadings on one component (Cronbach alpha = 0.62). See Appendix 3 in supplementary material

2.2.4. Step 3: exploring the relationship between potential accelerators and SDG targets

The proportion of the different variables was assessed by determining the frequency of participants that displayed the different SDG outcomes and accelerator protective factors. This was tabulated for visual appreciation. To objectively choose factors that could be potential accelerators, bivariate analysis using Chi-squared and t-test were conducted. Chi-square analysis was carried out to assess associations between the SDG targets and categorical mapped out accelerators while the association between accelerators measured on a continuous scale and the SDG targets was assessed using a t-test. Any variable that was significantly associated with at least two SDG targets was retained as potential accelerators for exploration in the next step of the analysis.

2.2.5. Step 4 : muti-variate logistic regression

To ascertain whether potential protective factors that emerged from the bivariate analyses were independent predictors of attaining the SDG-aligned targets, multiple multivariate logistic regression analyses were performed. For each potential accelerator, logistic regression analyses were run, with each model including one SDG target and confounders. The confounders adjusted for included age, sex, socioeconomic status, and parent’s highest formal educational level, geographical location (rural versus urban), orphanhood and large family size.

2.2.6. Step 5: Benjamini Hochberg corrections

Given that multiple hypotheses were tested simultaneously using multiple logistic regressions, we controlled for the possibility of false positives findings emerging. This was done using the Benjamini-Hockberg procedure at a false discovery rate of 10%. The independent predictors which were associated with at least two (2) SDG targets after the Benjamini Hochberg corrections were identified as potential accelerators.

2.2.7. Step 7: calculation of adjusted probabilities and marginal and partial effects for different combinations of protective factors

The adjusted probabilities of experiencing the outcomes (SDG-aligned targets) were calculated for the following three scenarios:

• exposure to all potential accelerators.

• exposure to the individual potential accelerators.

• exposure to none of the potential accelerators.

For accelerators measured using a binary indicator, no receipt is equivalent to ‘No’ and receipt is equivalent to ‘Yes’. For the only continuous variable in the sample – parenting support – the exposure was the highest attainable score on the scale (15) and no exposure as the average score (8.4) of the accelerator protective factor in the sample.

2.2.8. Statistical analysis

The statistical analysis was carried out in Statistical Package for Social Sciences (SPSS) version 24 and R studio. The details of the analytical methods have been explained in section 2.2 above.

3. Preliminary results

3.1. Sociodemographic characteristics of respondents

We examined data from 1,800 adolescents. The mean age of participants was 15.02 ± 2.27 years. Almost half (48.6%) of the participants were female, and a third of the participants (75.2%) were from urban districts. Majority of the participants (89.2%) parents were still alive, and 67.0% were from families with a low socioeconomic status (See Table 2).

Table 2. Socio-demographic characteristics of respondents (N = 1800).

Characteristics	Categories	Number	Percentage (%)
Age   Total	10–14 15–19	749 1051 1800	41.6 58.4 100
Gender   Total	Male Female	925 875 1800	51.4 48.6 100
Residence   Total	Rural Urban	446 1354 1800	24.8 75.2 100
Socioeconomic status*   Total	Low High	1206 593 1799	67.0 33.0 100
Orphan status   Total	Yes No	173 1605 1778	9.7 90.3 100
Fathers highest level of education       Total	No formal/Quranic Education Primary Education Secondary/Tertiary Education Don’t know	96 255 1018 426 1795	5.3 14.2 56.7 23.7 100
Mothers highest level of education       Total	No formal/Quranic education Primary Education Secondary/Tertiary Education Don’t Know	179 315 952 334 1780	10.1 17.7 53.5 18.8 100

N< 1800 indicates missing value.

*Participants’ socioeconomic status (SES) was derived using a modified version of the Oyedeji socioeconomic classification (Oyedeji, 1985), in which the parents’ occupational status and educational status were graded on a scale of 1 to 4 (i.e. highly skilled and highly educated= 4; unemployed/don’t know and uneducated/don’t know= 1). The average of the sum of both parents’ score was obtained, and the mean score (2.97) was used as the cut-off. Participants who scored below the mean were categorised as having high SES and those who had the mean score and above were categorised as having low SES.

3.2. Frequency distribution of SDG targets and accelerator protective factors

Seventy-nine percent (79.4%) of the participants had normal BMI for age, while 82.2% had normal height for age (not stunted). The majority (79.6%) of the participants, reported that they had not used substances in the month prior to data collection (See Table 3). Missing data for variables was less than 4.5%.

Table 3. Frequency distribution of SDG targets (N = 1800).

Variables	N(%)
SDG Targets
SDG 2.2 Adequate BMI for Age	1389 (79.4)
SDG 2.2 No Stunting	1438 (82.2)
SDG 3.5 No substance use	1428 (79.6)
SDG 4.1 Optimal school progression	927 (51.5)
SDG 4.4 Ability to concentrate in class	978 (55.2)
SDG 4.4 Optimal school attendance	1238 (69.8)
SGD 5.2 No risky sexual behaviours	1463 (83.1)
SDG 16.1a No self-perpetration of violence	1230 (71.4)
SDG 16.1b No community violence	1202 (32.0)
SDG16.2 No sexual abuse	1573 (89.6)
Accelerator Protective Factors
Food secure	1221 (68.4)
No mental health issues	890 (50.5)
Safe school	1163 (66.8)
Not doing survival work	1472 (81.8)
Regular Physical activity	703 (39.9)
Stable Childhood	1301 (73.3)

The mean score for parenting support was 10.31± SD3.08, with a range of 1 to 15. The majority of participants (68.4%) were classified as being food secure, while about half (49.4%) reported having at least one mental health condition. Also, about 2 out of 10 participants (18.2%) were engaged in survival work (See Table 3). Missing values for variables was less than 3.5%.

3.3. Association between proposed accelerators and SDG targets

3.3.1. Bivariate associations between hypothesised accelerators and SDG targets (03) related to survive and thrive dimension of children’s rights

Being food secure (x² = 13.89, p= <0.001), having good mental health (x² = 20.70, p = <0.001) safe schools (x² = 27.62, p= <0.001) and having a stable childhood (x² = 6.46, p = 0.011) were associated with not engaging in substance use.

Not doing survival work (x² = 4.35, p = 0.037) was associated with having normal BMI for age.

Not doing survival work (x² = 22.53, p < 0.001) and having a stable childhood (x² = 4.55, P = 0.033) were associated with not being stunted (See Table 4).

Table 4. Association between hypothesised accelerators and SDG targets related to the survive and thrive dimension of the Child Rights (N = 1800).

SDG Targets	Normal BMI for Age			No stunting			No substance use
Accelerators	n (%)	X^2	p-value	n (%)	X^2	p- value	n (%)	X^2	p- value
Food Security Yes No	939 (79.3) 437 (79.5)	0.01	0.944	958 (81.1) 468 (84.8)	3.46	0.063	998 (82.1) 419 (74.4)	13.89	<0.001
Optimal Mental Health Yes No	696 (80.0) 691 (78.8)	3.90	0.532	713 (82.0) 723 (82.6)	0.13	0.712	851 (83.4) 575 (25.3)	20.71	<0.001
No survival work Yes No	1151 (80.3) 239 (75.1)	4.35	0.037	1208 (84.4) 231 (73.1)	22.53	<0.001	1165 (79.5) 263 (80.4)	0.15	0.697
Safe schools Yes No	895 (79.1) 452 (80.6)	0.52	0.469	930 (81.9) 460 (83.0)	0.34	0.556	965 (83.1) 415 (72.3)	27.62	<0.001
Stability of childhood Yes No	1005 (79.6) 361 (78.0)	0.57	0.450	1053 (84.3) 365 (79.0)	4.55	0.033	1054 (81.3) 358 (75.8)	6.46	0.011
Regular physical Activity Yes No	557(79.3) 845 (79.9)	0.07	0.790	570(82.8) 833(81.3)	0.89	0.40	554 (81.4) 805(77.9)	2.93	0.08

Bolded values indicate significance at P < 0.005

3.3.2. Bivariate associations between hypothesised accelerators and SDG targets (03) in the learning dimension of the child rights

Being food secure (x² = 5.68, p = 0.017), having optimal mental health (x² = 4.79, p = 0.029), not doing survival work (x² = 75.07, p= <0.001), safe schools (x² = 4.96, p = 0.026), and having a stable childhood (x² = 32.46, p= <0.001) were significantly associated with progressing well in school. Having optimal mental health (x² = 71.65, p= <0.001) and being in a safe school (x² = 16.71, p= <0.001) were significantly associated with being able to concentrate in school. Safe school (x² = 16.06, p < 0.001) was significantly associated with optimal school attendance (See Table 5).

Table 5. Association between hypothesised accelerators and SDG targets in the learning dimension of the Child Rights (N = 1800).

SDG Targets	School Progression			Able to Concentrate			Optimal School Attendance
Accelerators	n (%)	X^2	p-value	n (%)	X^2	p-value	n (%)	X^2	p-value
Food Security Yes No	651(53.3) 266 (47.2)	5.68	0.017*	717(59.5) 255(46.2)	26.92	<0.001	889 (73.6) 340 (61.6)	25.89	<0.001
Optimal Mental Health Yes No	550 (53.7) 374 (48.4)	4.79	0.029*	626(63.9) 318(43.3)	71.65	<0.001	732 (72.4) 503 (66.3)	7.73	0.005
No survival work Yes No	829 (56.3) 98 (43.7)	75.07	<0.001	778(55.7) 166(52.2)	1.30	0.253	1023 (70.3) 215 (67.8)	0.73	0.392
Safe schools Yes No	575 (49.4) 318 (55.1)	4.96	0.026	659(58.5) 256(47.9)	16.71	<0.001	839 (72.8) 357 (36.7)	16.06	<0.001
Stability of childhood Yes No	723 (55.6) 191 (40.3)	32.46	<0.001	720(56.0) 247(53.3)	0.96	0.327	915 (71.3) 309 (66.5)	3.76	0.052
Regular physical activity Yes No	375 (53.3) 531 (50.1)	1.78	0.181	372(53.6) 583(55.8)	0.84	0.358	482 (69.1) 736 (70.3)	0.31	0.580

Bolded values indicate significance at p < 0.005

3.3.3. Bivariate association between hypothesised accelerators and SDG targets (04) in the protection dimension of the child rights

Being food secure (x² = 16.14, p= <0.001), having optimal mental health (x² = 7.38, p < 0.001), not doing survival work (x² = 14.44, p < 0.001), safe schools (x² = 7.85, p = 0.005), and having a stable childhood (x² = 17.07, p < 0.001) were significantly associated with not engaging in risky sexual behaviours. Food security (x² = 15.74, p= <0.001), optimal mental health (x² = 15.18, p < 0.001) and safe schools (x² = 34.53, p < 0.001) were significantly associated with no self-perpetration of violence. Food security (x² = 39.06, p < 0.001), optimal mental health (x² = 5.45, p = 0.020) and safe schools (x² = 65.74, p < 0.001) were significantly associated with no community violence. Food security (x² = 7.66, p = 0.006), optimal mental health (x² = 18.61, p < 0.001) and safe schools (x² = 5.83, p = 0.016) were significantly associated with no sexual abuse (See Table 6).

Table 6. Association between accelerators and SDG targets in the protection dimension of the Child Rights.

SDG Targets	No risky sexual behaviour			No self-perpetration of violence			No community violence			No sexual abuse
Accelerators	n (%)	X^2	P-value	n (%)	X^2	P-value	n (%)	X^2	P-value	n (%)	X^2	P-value
Food Security Yes No	1021(85.5) 430 (77.8)	16.14	<0.001	874 (74.4) 349 (65.1)	15.74	<0.001	874 (72.7) 317 (57.6)	39.06	<0.001	1083(90.9) 477 (85.5)	7.66	0.006
Optimal Mental Health Yes No	850 (85.2) 610 (80.3)	7.38	0.007	892 (90.7) 614 (84.6)	15.18	<0.001	998 (69.3) 204 (62.6)	5.45	0.020	805 (92.7) 765 (86.4)	18.61	<0.001
No survival work Yes No	1217 (84.7) 246 (75.9)	14.44	<0.001	1013 (72.2) 217 (68.0)	2.16	0.141	1023(70.3) 215 (67.8)	0.73	0.392	1290 (89.7) 283 (89.0)	0.14	0.71
Safe schools Yes No	968 (84.5) 446 (79.1)	7.89	0.005	856 (75.8) 333 (61.9)	34.53	<0.001	856 (74.4) 311 (54.9)	65.74	<0.001	1035 (90.9) 492 (87.1)	5.83	0.016
Stability of childhood Yes No	1089 (85.5) 357 (77.7)	17.07	<0.001	901 (71.9) 317 (70.8)	0.19	0.669	870 (68.2) 313 (67.2)	0.16	0.688	1148 (90.5) 409 (88.1)	2.00	0.157
Regular physical activity Yes No	562(81.8) 874(83.8)	1.16	0.280	485(71.5) 721(71.2)	0.03	0.873	455(65.7) 727(69.7)	3.13	0.07	595 (87.2) 949 (91.0)	6.15	0.013

Bolded values indicate significance at P<0.005.

3.3.4. Association between parental support and SDG targets

Respondents who did not use substances had significantly higher scores on the parenting support scale compared to individuals who used substances (9.95 vs 10.40; t = 3.15, p = 0.753). Respondents with optimal school attendance had significantly higher scores on the parental support scale compared to individuals who did not (10.56 vs 10.56; t = 4.49, p < 0.001). Respondents who did not engage in risky sexual behaviours had significantly higher scores on the parental support scale compared to individuals who engaged in risky sexual behaviour (10.34 vs 10.11; t = 1.12, p < 0.001) (See Table 7).

Table 7. Association between parental support and SDG targets.

Accelerator	Parental Support
SDG Targets	Mean (SD)	t	df	p-value
No stunting No Yes	10.26 (3.46) 10.33 (3.27)	3.15	1750	0.753
Not underweight No Yes	10.34 (3.35) 10.30 (3.30)	0.20	1750	0.835
No substance use No Yes	9.95 (3.16) 10.40(3.33)	2.32	1782	0.020
School progression No Yes	10.15 (10.15) 10.46 (10.46)	1.98	1786	0.047
Able to concentrate No Yes	10.09 (3.21) 10.46 (3.38)	2.33	1761	0.019
Optimal school attendance No Yes	9.79 (3.14) 10.56 (3.33)	4.49	1767	<0.001
No risky sexual behaviour No Yes	10.11 (3.14) 10.34 (3.33)	1.12	1753	0.001
No self-perpetration of violence No Yes	9.71(3.29) 10.55(3.28)	4.80	1743	<0.001
No community violence No Yes	9.91 (3.15) 10.47 (3.36)	3.28	1751	0.001
No sexual abuse No Yes	9.91(3.11) 10.36(3.32)	1.73	1748	0.081

Indicate significance at P<0.005.

4. Discussion

In this paper we have described the steps/methods employed and presented the preliminary findings of the process of identifying protective factors that may lead to the attainment of multiple SDG targets from an existing dataset. This was a six-step process consisting of a mapping exercise, construct validity, bivariate analysis (using chi-square and t-test), multivariable logistic regressions, adjusting for multiple hypothesis testing, and estimation of marginal effects. The preliminary results showed that food security, optimal mental health and parenting support showed the greatest association with SDG outcomes.

The accelerator concept fronted by the UNDP is a promising and intuitive approach for the attainment of the SDGs especially for resource limited settings in SSA. However, research on this concept is still nascent (Chipanta et al., 2022; Cluver et al., 2019; Haag et al., 2022); therefore, many researchers interested in this sub-field may be poorly informed about the suitable methodological approaches to adopt. Concerns have been raised about the difficulty in replicating some scientific research, with this ‘replication crisis’ being attributed to the inadequate description of methodological details in scientific publications (Archmiller et al., 2020; Haddaway & Verhoeven, 2015; Ioannidis, 2018; Schooler, 2014). In this paper, therefore, we have attempted to expatiate on the steps required in carrying out accelerator analyses using secondary data. First, in selecting potential accelerators we went beyond relying solely on evidence in the literature by carrying out bivariate analyses to empirically select those factors/circumstances that are positively associated with at least 2 SDG-aligned targets in the data. Furthermore, we demonstrated (in the tradition of similar studies) the analytic rigour required in selecting potential accelerators by going beyond controlling for possible confounders to also accounting for multiple hypothesis testing. Also, by employing construct validity, we demonstrated what steps can be taken in unique cases such as when a potential accelerator (which is not necessarily measured on a standardised scale) is computed by merging different items of an instrument. It is, therefore, expected that these contributions would further improve the conduct of research in this growing sub-field.

In the current study, children who were food secure, were more likely not to use psychoactive substances, to have normal school progression and all the SDG targets under the protection dimension of the Child Rights (no risky sexual behaviour, no self-perpetration of violence, no community violence, and no sexual abuse). The relationship between food availability and exposure to different risks and coping mechanisms has been described in literature (Child Protection Cluster, 2017; Cluver et al., 2020). Food insecurity fuelled by the economic crisis in most SSA countries compel parents to make difficult decisions affecting their children (Child Protection Cluster, 2017). In order to meet food security needs, families turn to survival strategies such as early marriages and involvement of children in income-generating activities, which may either result in school dropout or a reduction in their educational outcomes (Child Protection Cluster, 2017). It has also been reported that, working children have lower adaptive skills, lower levels of physical health, and an increased tendency to engage in substance use and violence when compared to children who do not work (Hamdan-Mansour et al., 2013). Although, it has been clarified that not all work carried out by children and adolescents can be termed ‘child labour’ (International Labour Organization, 2023), the ‘survival work’ done by some children and adolescents in our study may well constitute child labour, which may then predispose these youngsters to some of the earlier mentioned negative outcomes. Similarly in South Africa, food insecurity has been linked to poor violence outcomes such as increased risk of sexual, physical and emotional abuse as well as community violence (Cluver et al., 2020). The association found in this study between food insecurity and perpetration and experience of violence may well be seen as giving credence to the cliché, ‘a hungry man is an angry man’.

This study also showed that, adolescents who did not have mental health concerns were less likely to use substances, more likely to have good educational outcomes and achieve all SDG targets under the protection dimension of the Child Rights (no risky sexual behaviour, no self-perpetration of violence, no community violence, and no sexual abuse). A previous paper from this data set, showed that being physically attacked and engaging in fights in the last 12 months was significantly related to suicidal ideation (Omigbodun et al., 2008). Another study carried out in Michigan to investigate the characteristics of youths who seek care at the emergency department for assault-related injuries showed that youths with assault injuries were more likely to have visited the emergency department in the past six months for mental health-related concerns (Cunningham et al., 2014). Qualitative data from the study also showed that having a bad mood and intent for retaliation was a common reason for engaging in fights (Cunningham et al., 2014). However, this association is not consistent as several other studies have yielded contrary findings revealing no clear relationship between violence and mental health concerns (Buchanan, 2008; Torrey et al., 2008).

Research findings suggest that parental support is a multi-dimensional and bi-directional construct that has clear links with social and academic outcomes of children (Amponsah et al., 2018). Majority of studies analysing the relationship between parental involvement and child school achievement have shown that children have better school outcomes if their parents show a high degree of warmth, supervision and psychological support (Amponsah et al., 2018; Cluver et al., 2019; Necsoi et al., 2012). This study was consistent with these findings as it revealed that children who had optimal school attendance and were able to concentrate in school had significantly higher mean scores on the parental support scale than their counterparts could not concentrate or who missed school at least once in the last term. Plausible reasons for this is echoed in the social cognitive theory, which suggests that children absorb messages about behaviour and socially accepted goals by observing and talking with important people in their lives (Bandura, 1977). In this regard, children are more likely to apply themselves and perform better in school, if their parents show an interest in their school work, are willing to assist them with their homework and are willing to hold them accountable for completion of assignments (Amponsah et al., 2018).

However, the protective factors that have emerged hitherto in this study may be influenced by factors such as socioeconomic status, parental level of education, geographical setting, gender, and orphanhood. For example, literature has shown that adolescents with low socioeconomic status, from rural backgrounds, and who have lost a parent are disadvantaged in terms of parental support, mental health, food security, and engaging in survival work. Therefore, to account for these possible interactions, we used these potential contributory factors as confounders in subsequent multivariate analyses – the fourth step of our analytic process, which has been mentioned in the methodology section. Results obtained after adjusting for these confounders (as well as the remainder of our analytic process) will be detailed in a future manuscript.

5. Implications of our findings for Nigeria and the sub-Saharan African region

With the advent of the COVID-19 pandemic, most countries have stalled in achieving sustainable development goals. According to the 2022 SDG index dashboard, 70% of countries in Sub-Saharan Africa are just 50% of the way to achieving the SDG targets less than ten years before the slated deadline (Sachs et al., 2022). Nigeria ranked 29^th out of 49 on the dashboard with an aggregate score of 54.27, marginally higher than the 2019 score of 47% (Enock et al., 2019; Sachs et al., 2022). In addition, another report to assess progress made on child-relevant SDG indicators showed that up to three quarters of African countries (Nigeria inclusive) either had insufficient data or showed insufficient progress to meet SDG global targets by 2030 (UNICEF, 2018b). The results in this study suggest ways to meet up with the 2030 agenda especially in resource-constrained settings, by identifying factors that could simultaneously lead to attainment of SDG targets. Potential accelerators identified span across different sectors: schools, family, and environment emphasising the need for multi-sectoral reforms and interventions.

6. Limitations and strengths

One of the major limitations of this study is that out of school adolescents who are known to have poorer outcomes (Manzuma-Ndaaba Ndanusa et al., 2021) were not included. Consequently, our results may not be generalizable to out of school children who are usually the most disadvantaged (Manzuma-Ndaaba Ndanusa et al., 2021). Nevertheless, we used a large sample size of 1800 adolescents drawn from rural and urban settings in Ibadan and environs. Also, the information was obtained by self-report and required participants’ ability to recall important information and so might have been subjected to recall bias. Even though this limitation is typical of cross-sectional studies, we expect that the rigour of questioning by interviewers – who also assured participants of utmost confidentiality – would have helped to reduce the impact of self-reporting bias.

Also, our findings are based on data from about 20 years ago, which makes it a rather historical data and, as such, it may – at first glance – seem non-relevant to the current situation in Nigeria. However, findings from more recent population-based surveys such as the Nigeria Multidimensional Poverty Index (MPI) 2022 and the National Bureau of Statistics (NBS) and United Nations Children’s Fund (UNICEF) Multiple Indicator Cluster Survey 2021 (NBS & UNICEF, 2022) show that our data is quite reflective of the current state of things. For example, the Food and Agricultural Organisation (FAO) reported a rise in the prevalence of undernourishment in Nigeria from 7.1% in 2005 to 14.6% in 2019 (FAO, 2021); experience of ‘security shocks’ which includes sexual and community violence has been estimated to be 15.7% (MPI, 2022) compared to 10.4% for sexual violence found in our study; prevalence of food insecurity has been reported to be 50.9% (MPI, 2022) compared to 31.6% found in our study; 21.3% of children aged 7–14 years were reported to have parental assistance with their homework in 2021 (NBS & UNICEF, 2022); and 20.6% of children aged 5–17 years have been reported to be involved in economic activities (NBS & UNICEF, 2022) compared to the 18.2% prevalence of engagement in survival work found in our study. Although these other more recent surveys were not restricted to adolescents, adolescents were included to some degree, thus giving some room for extrapolation of the overall findings to adolescents.

7. Conclusion

This paper provides a step-by-step description of accelerator analyses using secondary data and provides a stimulus for researchers on the African continent to analyse data sets collected through the years to add strength and voice to the needs of adolescents. Preliminary findings reveal that food security, optimal mental health, not doing survival work, safe schools, small family size, stability of childhood and parental support may be protective factors associated with attainment of at least two SDG targets. However, we are aware that, in addition to being potential causative factors, these identified protective factors may also be consequences of broader systemic issues. Therefore, subsequent analyses that adjust for confounders, multiple testing, and quantify the extent of the impact of these protective factors across multiple SDGs – when acting alone and in synergy with others – would help in identifying which of these factors will meet the criteria for accelerators as suggested by UNDP. Targeted actions to address these factors would likely result in change across multiple areas of adolescents lives simultaneously.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work was supported by the UK Research and Innovation Global Challenges Research Fund [UKRI GCRF Accelerating Achievement for Africa’s].

Appendix 1:

Instruments used during data collection

Instrument/questionnaire	Authors	Description
Socio-demographic Questionnaire	Olayinka Omigbodun	Collects information on sex, age, family structure and school related issues.
Global School Health Questionnaire (GHSQ)	World Health Organization in collaboration with Centre for Disease Control and Prevention (WHO & CDC, 2003)	Measures behavioural risk and protective factors across 10 key areas: Alcohol use Dietary behaviour Drug use Hygiene and sanitation Physical activity Sexual behaviour Dietary behaviour Parenting support and supervision Violence and unintentional injury
Culture-free Self-esteem Inventory	Battle James (Battle, 1992)	30-item questionnaire used to identify problems with self-esteem in adolescence. 10 items were selected for use during the data collection
Diagnostic Interview Schedule for Children (DISC) Predictive scales	National institute for mental health in the United States of America (Shaffer et al., 2000	Structured diagnostic interview which assesses common mental health disorders in children and adolescents such as depression, anxiety, eating disorders, suicidal behaviours and conduct disorders

Appendix 2:

Details of questions/items used in assessing some accelerator protective factors (safe schools and parental support)

A) Safe schools

The following questions were used to assess whether a school was safe or not

1. I was not bullied during the past 30 days.

2. I was kicked, pushed, shoved around or locked doors.

3. I was made fun of because of my ethnic group.

4. I was made fun of because of my religion.

5. I was made fun of with sexual jokes, comments, or gestures.

6. I was left out of activities on purpose or completely ignores.

7. I was made fun of because of how my body or face looks.

8. I was bullied in some other way.

B) Parental support

The following questions were used to assess parental support

1. During the past 30 days how often did your parents or guardians check to see if your homework was done? (Never, rarely, sometimes, most of the times, always).

2. During the past 30 days how often do your parents or guardians understand your worries and problems? (Never, rarely, sometimes, most of the times, always).

3. During the past 30 days, how often did your parents or guardians really know you were doing with your free time? (Never, rarely, sometimes, most of the time, always).

Appendix 3:

Factor analysis of parental support

Questions	Factor loadings
Parent/s know what you are doing with free time	0.783
Parent/s understand worries and problems	0.780
Parent/s check to see if homework is done	0.688