Child Labor in the State of Amazonas: Invisibility of the Notification System

Objective: To describe the characteristics of reported cases of child labor in general and to compare official reported child labor data with data on sexual exploitation and occupational accidents involving children and adolescents between 2017 and 2021. Methods: Descriptive study based on data from the Notifiable Health Conditions Information System about child labor, work accidents and sexual exploitation of children and adolescents aged between 5 and 17 years. Annual disease notification rates were calculated and grouped by mesoregions of the state of Amazonas. Results: From 2017 to 2021, there were 312 reports of situations of child labor, being 46.8% involving children and 53.2% involving adolescents. Of these notifications, 20 corresponded to child labor itself, 141 to child and adolescent sexual exploitation, and 151 to work-related accidents involving children and adolescents. The South and Center regions were the ones that most underreported child sexual exploitation as child labor. The Southwest and South mesoregions, on the other hand, were the ones that most underreported work accidents involving children and adolescents as child labor. Conclusion: Child labor in Amazonas basically stems from sexual exploitation and, even though it is underreported, it is frequent, although uneven between regions. Confronting the problem in the state necessarily involves improving the information system in order to acknowledge the actual dimension of the problem and then define intervention measures and logistics.


INTRODUCTION
Half of the world's children, or about 500 million of them, are affected by physical, sexual or psychological abuse 1,2 . According to the Ministry of Health, in 2021, Brazil registered about 83 thousand cases of child and adolescent abuse, with 1,744 notifications in the state of Amazonas 3 . The most common forms of abuse are, intrafamily violence, abuse and neglect 4 . Recognizing this issue as a health problem is relatively recent in Brazil 5 , especially regarding child labor (CL).
The invisibility of the problem related to CL stems, among other factors, from the emptying of data in the information systems of events that require compulsory notification, which impairs the effectiveness of this management tool 6 . The international literature points to factors that promote and hinder notification, but evidence on the subject is still scarce in Brazil [6][7][8][9][10] , especially studies on CL in the northern region of the country 11 .
Item 8.7 of the United Nation Sustainable Development Goals 2030 provides for a commitment to eliminate CL in all forms by 2025 and the worst forms immediately, such as hazardous work and sexual exploitation 12 . The International Labor Organization (ILO) deals with minimum age for work (Convention No. 138) and the worst forms of CL (Convention No. 182), and considers child work any work performed by children and adolescents below the minimum age established in the country. In Brazil, work below the age of 16 is prohibited, but an apprenticeship contract is admitted from the age of 14 onwards. Only from the age of 18 is night work allowed, in activities that are painful, unhealthy, dangerous or that, by nature or circumstances, can cause harm to physical, psychological, social, moral or educational development. The ban covers economic or survival activity, with or without profit, remunerated or not.
Child labor is, in addition to a violation of law, an adverse event in childhood and adolescence that has consequences for children's physical and mental health, through excessive and early exposure to work activities. As for the study, these damages occur, above all, when it comes to sexual exploitation and work accidents involving children and adolescents.
Compulsory notification covers all cases of child labor, meaning it includes hazardous work and sexual exploitation. For this reason, this article aimed to describe the characteristics of reported child labor cases in general and to compare official reported child labor data with data on sexual exploitation and work-related accidents involving children and adolescents between 5 and 17 years of age, from 2017 to 2021, in the state of Amazonas. The analysis was based on data from the Notifiable Diseases Information System (SINAN).

METHODS
This is a descriptive study developed with data from the Department of Informatics of the Unified Health System (DATASUS) on child labor, considering children and adolescents aged 5 to 17 years, from 2017 to 2021. Data on child labor in general (including sexual exploitation and work accidents involving children and adolescents) were accessed in the SINAN databases.
For reporting purposes, child labor comprises any activity carried out by children and adolescents (aged below the minimum age established in the country, according to the ILO) in a mandatory, regular, routine way, remunerated or not, in conditions that are sometimes disqualified and that puts physical, psychological, social and moral well-being at risk, limiting their conditions for healthy and safe growth and development 13 .
The sexual exploitation of children and adolescents is a form of child labor. It is characterized by the use of people for commercial and profit purposes, for the practice of sexual acts, exposure of their naked bodies and live sexual relations (strip-tease, erotic shows), or disclosure of images in magazines, films, photos, videos or websites 13 .
The occurrence of a work accident can also reveal exposure to early work. It derives from unnatural causes comprising accidents and violence (Chapter XX of the International Statistical Classification of Diseases and Related Health Problems 10 V01 to Y98), related to the tasks of the function (typical) or the journey between home and work (from path), causing bodily injury or functional disturbance that may lead to temporary or permanent loss or reduction of the ability to work and death 14 . It is associated with child labor if it affects workers under 18 years of age on the date of its occurrence 15 .
In the transfer of microdata from the DATASUS website, files in the dBase File Compacted (*.dbc) format from SINAN were used, dating from 2017 to 2021, in Brazil, referring to the complaints of "interpersonal/self-inflicted violence" and "occupational accident". The conversion of files for further manipulation was performed in RStudio (version 4.2.1). After converting data into the Comma Separated Values (*.csv) format, three selections were defined, described below.
Initially, selection I-interpersonal/self-inflicted violence-from 2017 to 2021 (child labor in general) was the following: in the "FU of notification" field, the number "13", corresponding to the state of Amazonas, was added; in the "type of violence" field, cases of "child labor" was added; and in the "age" field, the range "5 to 17 years old" was added.
Then, selection II-interpersonal/self-inflicted violence-from 2017 to 2021 (child labor in the form of sexual exploitation) was as follows: in the field "FU of notification", the number "13", referring to the state of Amazonas, as added; in the field "if sexual violence occurred, which type?", cases of "sexual exploitation" was entered; and in the "age" field, the range "5 to 17 years old" was added.
Finally, selection III-work accident-from 2017 to 2021 (child labor in the form of work accident): in the field "FU https://doi.org/10.1590/1980-549720220042 of notification", the number "13", referring to the state of Amazonas, as added; and in the "age" field, the range "5 to 17 years old" was added.
Based on the selections, notification rates were calculated through the number of cases reported in each situation of interest per 100,000 children and adolescents aged 5 to 17 years. The population projection 16 and the distribution of the 62 Amazonian municipalities by mesoregion were established according to the Brazilian Institute of Geography and Statistics (IBGE).
Therefore, four Amazonian mesoregions were considered: • North (six municipalities): Barcelos, Novo Airão, Santa Isabel do Rio Negro, São Gabriel da Cachoeira, Japurá and Maraã; • Southwest (16 municipalities Finally, to calculate the proportion of underreporting of child labor in relation to cases of sexual exploitation by mesoregion in the state of Amazonas between the years of interest, Equation 1 was used: 100 -(Number of child labor cases/Number of sexual exploitation cases)×100 (1) In the second analysis, which corresponds to the proportion of underreporting of child labor in relation to cases of work accidents involving children and adolescents, the following calculation was performed (Equation 2): 100 -(Number of child labor cases/Number of work accident cases involving children and adolescents)×100 (2) In both formulas, cases of children and adolescents between 5 and 17 years of age were included.
The study was not submitted to the Research Ethics Committee, as it uses secondary data in the public domain, which does not allow the identification of cases.

RESULTS
According to SINAN, from 2017 to 2021, 312 cases of child labor were recorded in the state of Amazonas, the sum of all possible sources of information on the condi-tion (Table 1). There were 146 cases of children (46.8%) and 166 cases of adolescents (53.2%), predominantly in the age range from 5 to 13 years old, in situation of child labor and in the form of sexual exploitation.
Most cases involved female children and adolescents (51.6%). In cases of work accidents, however, the records of boys were more common (85.4%). Regarding race/skin color, brown individuals were mostly involved (79.8%).
Most of the occurrences of child labor occurred in the victim's home (51.9%) and on public roads (18.3%). About a third of reports of child labor and child sexual exploitation were repeated situations (34%).
In the analyzed period, there was a gradual reduction in notifications of cases of child labor in all regions of Amazonas (Table 2), except in 2020 (0.48 cases/100,000 children and adolescents). The central region of Amazonas was the mesoregion with the highest rate of notifications (0.46), followed by the south region (0.23). It is noteworthy that the northern mesoregion did not register cases in the period examined. In total, there were 141 notifications of cases of child labor in the form of sexual exploitation of both children and adolescents ( Table 3). The peak notification rate of case occurrences was in 2019. The southern region (6.27) and the southwest region (2.68) of Amazonas had the highest rates of notification. The north mesoregion had the lowest rate (1.49).
Regarding work accidents involving children and adolescents aged 5 to 17 years (Table 4), the highest rates of notification were found in the southwest (10.37) and south (8.40) mesoregions. An upward trend was observed in the first three years, followed by a downward trend in the last two years.
In Amazonas, the northern mesoregion had the lowest rate of child labor notification in all forms. On the other hand, the central mesoregion had the highest rates of child labor notification in general; the southern mesoregion stood out with records of child and adolescent sexual exploitation; and the southwest mesoregion was the one that most identified work-related accident situations involving children and adolescents.
Cases of sexual exploitation of children and adolescents and work-related accidents involving people under the age of 18 must be notified as child labor; however, there is underreporting ( Table 5). The southern and central regions of the state had the highest proportion of underreporting of sexual exploitation as child labor. The southwest and south    mesoregions were the ones that most underreported situations of work accidents involving children and adolescents as child labor. Sexual abuse and exposure to risk situations related to early work are not properly recognized by the victims themselves, guardians or professionals, given the complexity of these phenomena. This indicates that the proportions of underreporting of child labor in the form of sexual exploitation and child labor in the form of work accidents may be underestimated.

DISCUSSION
This study points out that, in the state of Amazonas, at least 9 out of 10 cases of child labor result from sexual exploitation. The occurrence is uneven according to the analyzed mesoregion, and underreporting is evident.
The results found are even more alarming when compared with data from IBGE's Continuous National Household Sample Survey, which highlights, in the state of Amazonas, the existence of 56,601 children and adolescents aged 5 to 17 years in child work situation in 2019. Given that the estimated population in this age group was 937,193 in the same year, the universe of working children and adolescents was equivalent to 6% of the total children and adolescents in the state, above the national average of 4.8% 17 .
The pattern of reported cases points to a greater involvement of female children in situations of child labor in general and in the form of sexual exploitation. This phenomenon has a multicausal aspect, basically related to cultural, socioeconomic and interrelational/psychological contexts, bringing negative consequences for the physical and mental health of children and adolescents, especially girls, and for the health services that receive victims of associated abuse 18 .
As for adolescents between 14 and 17 years of age, work can be encouraged in certain contexts as a tool for social and psychological development, and not seen as harmful, but as a potentiator of learning and skills. It is essential to avoid and prevent work in precarious conditions, the exploitation of child labor and disrespect for the rules and regulations in force 19 .
With regard to race/skin color, in Amazonas, mixed race was predominant in all situations. This finding corroborates the results of a survey on the characteristics of victims of child labor carried out in Divinópolis (MG) 20 .  According to the notifications analyzed, child labor situations occurred mainly in the victims' own homes. In this regard, the invisibility of domestic child labor stems from historical and cultural factors, such as slavery and the custom of bringing girls from rural areas to study in the capital and live in family homes 21 . A study carried out in João Pessoa (PB) revealed a set of activities performed by this group, as they start early to take care of the house, people and food in the absence of adults, who usually are absent at work. The modalities of domestic work "socialization" and "help" give a character of invisibility, not being seen as work. Although these types of work are invisible, the research identified complaints of tiredness, fatigue, overload and impairment of school activities 22 .

Table 5. Percentage of child labor underreporting registered in the Notifiable Diseases Information System, by
There is an expressive percentage of repetition of child labor cases. The reiteration of exposure to family violence is often accompanied by restricted access to school. Children and adolescents are threatened, exposed and witness situations that leave them emotionally vulnerable. They learn that situations of violence are part of life, consider them trivial and can repeat them in their current and future relationships 23 .
SINAN records show a reduction in the number of notifications in all Amazonas mesoregions in the last two years studied compared to the initial years, which may show the impact of the COVID-19 pandemic on the identification and notification of occurrences. In a study carried out in the state of Rio Grande do Sul, there was a 54% decrease in reports of violence against children and adolescents in 2020 compared to the same period in 2019 24 . This result was also pointed out in a study conducted in the state of Santa Catarina, which compared compulsory notifications of violence against children and adolescents and reported a decrease of 55.3% in the period of isolation, which may reflect possible difficulties in accessing protection and assistance institutions 25 .
In a study carried out in Rio de Janeiro on the quality of reports of interpersonal/self-inflicted violence, only nine cases of child labor were recorded from 2009 to 2016 26 . The number of reports of child sexual exploitation in Santa Catarina increased by 662.5% from 2009 to 2019, especially in the last year analyzed 27 . An increase in the number of notifications of work accidents involving children and adolescents was also observed in a study carried out in João Pessoa 28 .
On the other hand, underreporting of child labor is evidenced in this study and reached high levels in all years of analysis. One of the hypotheses for this result is that identifying child labor is a complex task for professionals responsible for the notification, especially when it comes to the worst forms of child labor such as sexual exploitation and exposure to dangerous labor, which can lead to work accidents 29 .
Brazil was part of the group of pilot countries for the implementation of the Child Labor Risk Identification Model (MIRTI), carried out by the ILO-Brazil and the Economic Commission for Latin America and the Caribbean. This tool can generate information to improve the response to such cases and advance in the prevention and eradication of child labor. According to the MIRTI, in terms of risk of child labor, the municipalities in Amazonas are distributed as follows: high (12); medium (35); and low (10). Therefore, the following municipalities are at high risk for child labor: Presidente Figueiredo, São Sebastião do Uatumã, Careiro da Várzea, Manaquiri, Nova Olinda do Norte, Apuí, Manicoré, Tapauá, Pauini, Ipixuna, Juruá and Uarini 30 . This study shows that all these municipalities are geographically inserted in the mesoregions that most underreport.
The study has some limitations, as it relies on secondary data, over which researchers have little control. In addition, there is no information that makes it possible to describe in detail the exact pattern of child labor cases and the factors that may be associated with the occurrences.
The analyses of data referring to periods before and after the declaration of the COVID-19 pandemic by the World Health Organization reinforce the need to intensify the training and qualification of professionals responsible for filling out notification forms. Such actions should especially cover municipalities in the central portion of the state that are at high risk of child labor and located in the mesoregions that most underreport cases.
A flow of identification and treatment of child labor in the municipalities of Amazonas must be established, determining the entry points and including health services, which configure reporting units and make up a reference network for the prevention and eradication of child labor. These actions need to be a priority in the municipalities of the mesoregions with the highest proportions of underreporting.