Child Protection and Household Vulnerability : A Longitudinal Analysis of Child Rights and Protection Amongst Vulnerable Households and their Families in Rural Uganda

This paper explores child protection based on three factors; child abuse, child labour and child substance abuse. The analysis is based on a cohort of 17,848 vulnerbale households and the children in therein studied over a period of four years under the “sustainable comprehensive repsonses for vulnerable children and tehir families project” implemented in rural Uganda between 2011 and 2017. We find overall child protection factors mentioned above improved over the four years alongside household vulnerability. Howveer we observe that compared to the rest of the vulnerable households, the children who had experience child abuse, used susbanctes and alcohol and experience child labour were less likely to transition out of vulnerability compared to the others. We also found out that these children were less likely to be enrolled in school over the foru years and were more likely to be absent from school. We conclude that poor child protection indtcaors amongst the vulnerable children compound their household and individual vulnerability and increase the likelihood that their basic rights will be violated. We recommend that child protection issues, and child rights specifically should be brought at the forefront of all child protection interventions.


Introduction
Vulnerability is a widely used but elusive concept (Whitney, 2015).The World Bank (2000) defines vulnerability as a high probability of a negative outcome or an expected welfare loss above a socially accepted norm.Pelling (2003) defines vulnerability as exposure to risk and an inability to avoid or absorb potential harm.Vilagrán de León (2006), look at vulnerability in terms of its categories.In this context, he defines physical vulnerability as the vulnerability of the physical environment; social vulnerability as experienced by people and their social, economic, and political systems; and human vulnerability as the combination of physical and social vulnerability.Both vulnerability and its antithesis, resilience, are determined by physical, environmental, social, economic, political, cultural and institutional factors (Prevention Consortium, 2007).Whereas child vulnerability alludes to vulnerability amongst children, it means different things to different actors supporting children.In Uganda, the Uganda Bureau of Statistics, UBOS (2012) defines vulnerability as a measure of the extent to which a community, individual, structure, service, or geographic area is likely to be damaged, disrupted, or impoverished by the impact of a disaster or hazard because of its nature or location.On the other hand, the national Orphans and Vulnerable Children policy (2011) defines vulnerable children as including orphans, children below the age of 18 years who have lost one or both parents and other children and who are likely to be in a risky situation, and /or likely to suffer significant physical, emotional, or mental harm that may result in violating their human rights.Without a harmonized definition of vulnerability or child vulnerability, it is not surprising that according to UNICEF (2016), the global community risks failing millions of children if it does not focus on the most disadvantaged in its new 15-year development roadmap.Eriksen and Kelly (2007) conclude that vulnerability is a human state or condition that cannot easily be measured in an objective manner.In this article, vulnerability is defined as child deprivation measured at the household level against five thematic wellbeing parameters: child protection, access to legal services, food security and nutrition, economic wellbeing, and access to critical services (SCORE, 2012).Our measure of vulnerability incorporates both individual attributes of the child in question and household members as well as the characteristics of the household where the index child resides.
Regardless of the definitional challenges, what remains clear is that child protection is a factor in mitigating child vulnerability outcomes.Kelly Clarke (2016), for example, states that violence against children increases their level of vulnerability because children who are victims of abuse have worse educational outcomes compared to their non-abused peers and are at increased risk of financial and employment problems in later life.The protection of children from all forms of abuse is a fundamental right guaranteed by the Convention on the Rights of the Child and other international human rights treaties and standards; yet abuse remains an all-tooreal part of life for children around the globe (UNICEF 2017).Africa has one of the highest levels of child exposure to violence against children, for example, with a past year prevalence of 50 percent according to The World Health Organization (2017).Child abuse can take several forms, including physical, sexual, psychological, child labour and neglect.The World Health Organization (2017) defines child abuse as child maltreatment, which means abuse and neglect of children that takes physical form.These definitions and explanations are consistent with indicators associated with child vulnerabilities in Uganda.This article uses longitudinal data to analyse child and household vulnerability in relation to child protection issues: child abuse, child substance use, child labour and child school enrolment and their effect on household vulnerability.We discuss the effect of these protection issues at household level, how they compound child vulnerability and how they impact on the right to education.

Child vulnerability and protection in Uganda
Consistent with the definitions above, the present study considers child labour as work that is mentally, physically, socially and or morally dangerous and harmful to children and/or interferes with the child's education.Child abuse is any recent act or failure to act on the part of a parent or caretaker/guardian that results in death, serious physical or emotional harm, sexual abuse or exploitation; or an act or failure to act which presents an imminent risk of serious harm.
Child alcohol and substance abuse is defined as child intake of substances; prenatal exposure of a child to harm due to the mother's or father's use of an illegal drug or other substance; manufacture of the substance in the presence of a child; selling, distributing, or giving illegal drugs or alcohol to a child; and use of a controlled substance by a caregiver that impairs the caregiver's ability to adequately care for the child.
Children in Uganda are the most vulnerable societal group despite representing a significant majority of the population (UNICEF, 2015).According to UNICEF (2017), most children have experienced some form of violence and abuse, For example, more than 8 million children are vulnerable to harm and sexual abuse is the most common form of violence, with gender as a major risk factor.A 2012 UNICEF and Ministry of Education survey in northern Uganda (Acholi and Karamoja) shows that 78 per cent of primary and 82 per cent of secondary school-aged children had been sexually abused.In terms of child labour, despite the enactment of laws against child labour and exploitation, the Uganda Bureau of Statistics (UBOS) (2017), shows that nearly 45 percent of children in vulnerable households (defined as living below the poverty line in this case) are forced out of school to work and supplement their parents' incomes.
In terms of alcohol and drug abuse, although the data are scarce, available preliminary evidence shows increasing cases of child alcohol and drug abuse.For example, Hope and Beyond (2017), one of the organizations that supports alcohol and drug abuse victims shows that the number of children attending their programs rose from 60 in 2011 to 600 in 2017 in the capital city alone.
Apart from definitional challenges, the understanding of child vulnerability and child protection issues such as child abuse, alcohol and substance abuse and child labour is limited.This is largely due to cross sectional research designs that cannot be used to analyze behaviours over time.Secondly, research is based largely on one-dimensional definitions of vulnerability, focusing on individuals including those affected by HIV/Aids and those living below the poverty line.The present study takes a holistic definition of vulnerability, looking at the household as well as individual parameters categorized into five different themes.Longitudinal assessments determine variable patterns over time regarding improvements in child vulnerability over four years, between households with children that experienced child abuse, alcohol and substance abuse, and child labour with those that did not report these cases.The study compares changes in vulnerability status of these children after four years and analyzes the effect of these child protection issues on the child's right to education based on school enrolments and absenteeism.
A case is made for more inclusive program approaches to ensure children affected by child abuse realize their full rights to education and development opportunities.

Data source
Panel data sourced from moderately and critically vulnerable children and their households in the 35 rural districts of Uganda were used for this study.This data was collected between 2011 to 2016 through the Sustainable Comprehensive Response for Vulnerable Children and their Families project (SCORE 2013).The data were compiled from the same vulnerable children and their households annually over four years representing the four different stages of the evaluation mentioned below.The same population was assessed over four years and standing for four stages of assessment.Whereas in each vulnerable household there were on average, six vulnerable children, the data refers to one child who was considered the most vulnerable by the household members and the social worker.The number of participants assessed in Stage 1, Stage 2, Stage 3, and Stage 4 of the evaluation were 18,532, 13,209, 18,615, and 14,882, respectively.The variation in figures implies that all participants were not assessed at every stage of the evaluation; thus, the panels were not balanced.

Variables and Measurements
The Vulnerability Assessment Tool (VAT) was used annually to assess child and household vulnerability (SCORE, 2011).The level of household vulnerability was assessed using three outcomes: critical, moderate and slight or no vulnerability.The tool assesses vulnerability based on five major thematic areas: (i) economic wellbeing, (ii) child protection and access to legal services, (iii) food security and nutrition, (iv) access to critical services including health; (v) an allowance for the social worker to make their own personal impression about vulnerability given the complexity of the subject matter.Each of the thematic areas of the VAT yields a maximum of 30 vulnerability points while the assessor's impression yields a maximum of 10 points.To this end, the highest possible vulnerability score is 130.All thematic areas enter linearly in the overall score; households with a VAT score between 40 and 53 are considered moderately vulnerable while those with a VAT score above 54 are considered critically vulnerable.On the other hand, those with a VAT score below 40 are considered to be slightly or not vulnerable.Vulnerability transition is considered as a shift from a higher level of vulnerability to a lower one.This involves a shift from critical vulnerability to a lower level i.e. moderate or slight vulnerability or a shift from moderate vulnerability to slight vulnerability.
In the present study, the dependent variable, "vulnerability improvement," saw a decrease in vulnerability status between the baseline year (2011) and the end line year (2014).On the other hand, schooling status was evaluated using an outcome; that is, whether a child was in school at the time of the evaluation.A similar approach was adopted in modeling absenteeism; that is, whether a pupil was absent from school for at least a month in a term.The independent variables were characteristics of children (sex, involved in child labour, child abuse and alcohol or substance use).Table 1 presents a detailed description of these variables.

Data Analysis
The analysis was conducted using STATA 13.0 at three stages: 1) a descriptive summary of all the characteristics of children and their households was undertaken using frequency distributions.The purpose of the analysis was to provide a description of the children and households as well as a description of the vulnerability status of the children and household, before and after four years; 2) differentials in child protection indicators such as child abuse, child substance and alcohol abuse and child labour were assessed using the Pearson Chi-square test.The purpose of the analysis was to select variables for further analysis at the later stage.All variables with a probability value of 0.5 and below were considered for further analysis at the third stage.The appropriateness of using the fitted model in the analysis was assessed using the Pearson Chi-square goodness of fit test (Hosmer & Lemeshow, 1980).In the analysis at the multivariable stage, no interaction effects of the independent variables on either schooling status or absenteeism were investigated.However, the project worked with all the children in the household.

Results
The following section presents results regarding the characteristics of the vulnerable children and other members of the households.It also presents the improvement in vulnerability over the four years of implementation.We then compare the improvement from vulnerability of vulnerable households which had children affected by child abuse, child labour, substance abuse or lack of school enrollment as compared to those without these factors.The section concludes with data on general effects of child abuse, with specific reference to child abuse and its effect on improvement of child protection vulnerability and on education indicators within the vulnerable households.Nearly seven in every ten children in these households (68%) are either orphaned or have at least one parent absent.Less than a quarter (17%) of the households have at least one person living with HIV/AIDS and at least one in every 10 households (10.1%) has a guardian and a child (7.1%) living with a disability.Nearly one in every ten households was headed by either a child below 18 years (0.3%) or an elderly person above 65 years (10.1%).A few of the households were child headed.The vulnerable households under study have a household size typically ranging from one to 27 members.The average household size was 6 members, above the national average of 4.7 members (UBOS, 2016).Whereas the median number of dependents was two, several households had up to 20 dependents, including children and the elderly above 65 years, significantly constraining available meagre resources.

Vulnerability Status
The figure below shows the vulnerability status of the households/children from year (stage) 1 in 2011 to year (stage) 4 in 2014.This is the baseline situation collected in 2011 for all the vulnerable children who were enrolled in the project.As mentioned earlier, vulnerability was measured using a vulnerability assessment tool that categorized the target population into three vulnerability levels namely, slight/no vulnerability, moderately vulnerable, and critically vulnerable.The population that was neither moderately vulnerable nor critically vulnerable were not targeted for intervention and were not followed up.The assessment results of the pre-test (first year, 2011) show that out of 17,484 households, close to two-thirds were critically vulnerable while a third were moderately vulnerable.Critical vulnerability depicts the worst cases of vulnerability.Nearly two-thirds of the households were critically vulnerable.Figure 1 below shows that over the four years, there was notable improvement in household vulnerability.
The findings show that critical vulnerability decreased from 63.6% to 4.3% in four years.
In terms of the child protection indicators, namely child abuse, child abuse, alcohol and substance use), the study results indicate that at the time of enrollment (at pretest in 2011), one quarter of the vulnerable children in these households (26%) were involved in some form of child labour.This figure is slightly more than a third (38%) ofcases of child abuse including physical, emotional and psychological; nearly one in ten (7%) reported substance and drug

Regression Analysis
Regression analysis below shows that when issues of child protection are addressed, transition out of vulnerability is more likely.For example, in Table 2, when households which hitherto had child abuse, alcohol and substance abuse or child labour were involved in interventions mentioned above, they were more likely to transition out of vulnerability.Vulnerable children involved in child labour were less likely (p<0.05) to be enrolled in school over the four years compared to other children.These children were 15% less likely to be enrolled in school compared to the children that were not involved in any form of child labour (OR = 0.85).This implies that children involved in any form of child labour were less likely to be in school compared to those with no involvement.In addition, vulnerable children involved in any form of substance or alcohol abuse were less likely to be enrolled in school over the four years (p<0.05).Children involved in any form of substance or alcohol use had 44% reduced odds of being in school compared to those not involved (OR = 0.56).This implies that children involved in any form of substance or alcohol use were less likely to be in school compared to those not involved.Furthermore, findings suggeest that children affected by child labour were more likely to be absent from school (p<0.05).Children involved in any form of child labour had increased odds of school absenteeism compared to those with no involvement (OR = 2.76).This implies that children involved in any form of child labour are more likely to be absent from school compared to those with no involvement.Similar, we found out that children who were involved in substance and alcohol abuse were more likely to be absent from school (p<0.05).
Children involved in any form of substance or alcohol use, and especially male children, had increased odds of school absenteeism compared to those not involved (OR = 1.64).This implies that children involved in any form of substance or alcohol use were more likely to be absent from school compared to those not involved.

Discussion
This study set out to analyze child protection in the context of a vulnerable rural population.It aimed at demonstrating the effect of child abuse, child labour and substance above on overall household vulnerability.The study compared improvement in child vulnerability over four years between households with children who experienced child abuse, alcohol and substance abuse, and child labour with those who did not report these cases and noted changes in vulnerability status of these children after four years.The study also sought to demonstrate the compounding effect of child vulnerability, and specifically the factors mentioned above, on the recent years; most of the abusers are young, poor or both (UNICEF, 2015), with some children as young as 11 years.

Conclusions
In the context of vulnerable populations, negative child protection indicators and child rights are a challenge.Child abuse, child labour, and alcohol and substance abuse reinforce each other to deny vulnerable children other rights such as that of education.Findings of this study indicate that all of these negative child protection factors compound household and child vulnerability.The more a household has, the more likely that household will remain vulnerable.
As a compendium of indicators of children's rights, these findings demonstrate that vulnerability makes it harder for children's rights to be realized in the context of a rural population.
Figure 2: Child protec5on indicators amongst the vulnerable households over four stages of assessment

Table 1 :
Description of variables adopted in the assessment

TABLE 2 :
Child Protection Indicators As Determinants Of Transition From Vulnerability