Abstract

OBJECTIVE:

to measure underreporting of typical fatal occupational injuries (FOI) on the Mortality Information System (SIM) in Belo Horizonte city, Minas Gerais State, Brazil, in 2011.

METHODS:

the capture-recapture method was used to assess all FOI recorded among city residents and/or occurring in the city, using probabilistic linkage of three data sources: SIM, the Notifiable Diseases Information System (SINAN) and the Belo Horizonte Forensic Medicine Institute Autopsy Database.

RESULTS:

the majority of the 54 cases found occurred in males (n=53), with drivers and construction workers accounting for half of them (n=28); the most common causes of death were traffic accidents (n=18) and falls (n=10); 15 deaths were underreported on SIM.

CONCLUSION:

the use of the capture-recapture method contributed to the detection of FOI underreporting.

Keywords:
Information Systems; Occupational Health; Occupational Mortality; Forensic Medicine

Introduction

Deaths due to occupational injuries are a serious Public Health issue in Brazil and all over the world. Occupational injuries can be defined as 'typical', if they happen during the work activity, and as 'commuting', if they happen during a person's commute to or from work.¹1. Ministério da Saúde (BR). Secretaria de Atenção à Saúde. Departamento de Ações Programáticas Estratégicas. Notificação de acidentes do trabalho fatais, graves e com crianças e adolescentes. Brasília: Ministério da Saúde; 2006. (Série A. Normas e Manuais Técnicos); (Saúde do Trabalhador; 2. Protocolos de Complexidade Diferenciada). In 2010, there were around 350 thousand work-related deaths in the world.²2. Hämäläinen P, Leena Saarela K, Takala J. Global trend according to estimated number of occupational accidents and fatal work-related diseases at region and country level. J Safety Res. 2009;40(2):125-39.^,33. International Labour Organization. Safety and health at work: a vision for sustainable prevention. In: Annals of 20th World Congress on Safety and Health at Work: Global Forum for Prevention. 2014 Aug 24-27; Frankfurt. Frankfurt: International Labour Organization; [s.d]. 48p. This is a relevant amount, mainly if we consider the possible underreporting, especially in developing countries.⁴4. Fingerhut M, Driscoll T, Nelson DI, Concha-Barrientos M, Punnet L, Pruss-Ustin A, et.al. Contribution of occupational risk factors to the global burden of disease. SJWEH. Suppl. 2005; 1(1):58-61. According to the International Labour Organization (ILO), also in 2010, around 18 thousand deaths occurred in American low and middle income countries - including Brazil.³3. International Labour Organization. Safety and health at work: a vision for sustainable prevention. In: Annals of 20th World Congress on Safety and Health at Work: Global Forum for Prevention. 2014 Aug 24-27; Frankfurt. Frankfurt: International Labour Organization; [s.d]. 48p.

In Brazil, the workers with formal work contracts under the Consolidation of Labour Laws (CLT) are submitted to the General Social Security Policy. All fatal occupational injuries covered by the Work-related Injury Insurance (SAT) should be reported to the Ministry of Social Security (MPS), through a form called Communication of occupational injury (CAT). In 2010, more than 2,800 occupational deaths were reported to the MPS system.⁵5. Universidade Federal da Bahia, Centro Colaborador em Vigilância dos Acidentes de Trabalho; Ministério da Saúde (BR). Acidentes de trabalho fatais no Brasil 2000-2010: óbitos por acidentes de trabalho caem em todo país. Bol Epidemiol Acid Trab Fatais. 2011 abr;1(1):1-4. Also according to social security data, mortality rates due to work-related injuries fell from 14.9 in 1998 to 7.7 in 2006, per one thousand workers.⁶6. Lippel TH. Mortalidade por acidentes de trabalho, Brasil: análise de séries temporais, 1998 a 2006 [dissertação]. Florianópolis (SC): Universidade Federal de Santa Catarina; 2009. On the other hand, in São Paulo state, there was an increase in the fatality rate of almost six times, raising from 0.18% in 1970 to 1.07% in 2003.⁷7. Santana V, Nobre L, Waldvogel BC. Acidentes de trabalho no Brasil entre 1994 e 2004: uma revisão. Cien Saude Coletiva. 2005 out-dez;10(4):841-55. The increase in fatality rate and the reduction in mortalities may point to underreporting by the MPS.⁷7. Santana V, Nobre L, Waldvogel BC. Acidentes de trabalho no Brasil entre 1994 e 2004: uma revisão. Cien Saude Coletiva. 2005 out-dez;10(4):841-55. Moreover, it is noteworthy that public employees, domestic workers and informal workers are not covered by the SAT. The absence of SAT coverage for these workers restrains the real dimensioning of the epidemiological situation of cases in the country and can impact in the planning of actions focused on its control.⁷7. Santana V, Nobre L, Waldvogel BC. Acidentes de trabalho no Brasil entre 1994 e 2004: uma revisão. Cien Saude Coletiva. 2005 out-dez;10(4):841-55.

8. Gonçalves KR. Análise especial dos acidentes de trabalho assentados no Sistema de Informação de Agravos de Notificação (SINAN) em uma capital brasileira [dissertação]. Belo Horizonte (MG): Universidade Federal de Minas Gerais; 2012.

9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.

10. Correa PRL, Assunção AA. A subnotificação de mortes por acidentes de trabalho: estudo de três bancos de dados. Epidemiol Serv Saude. 2003 dez;12(4):203-12.^-1111. Alvares JK, Pinheiro TMM, Santos AF, Oliveira GL. Avaliação da completude das notificações compulsórias relacionadas ao trabalho registradas por município polo industrial no Brasil, 2007-2011. Rev Bras Epidemiol. 2015 jan-mar;18(1):123-36.

The notification of fatal occupational injuries is part of one of the attributions of the Ministry of Health that uses a specific notification form as source to feed the Information System for Notifiable Diseases (Sinan). When these injuries result in death, they are notified in the Mortality Information System (SIM), by using the Death Declaration (DO), which presents an specific field that relates the death to the occupation.⁹9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.

10. Correa PRL, Assunção AA. A subnotificação de mortes por acidentes de trabalho: estudo de três bancos de dados. Epidemiol Serv Saude. 2003 dez;12(4):203-12.

11. Alvares JK, Pinheiro TMM, Santos AF, Oliveira GL. Avaliação da completude das notificações compulsórias relacionadas ao trabalho registradas por município polo industrial no Brasil, 2007-2011. Rev Bras Epidemiol. 2015 jan-mar;18(1):123-36.^-1212. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5. Both systems are fed by the Municipal Health Departments (SMS), are considered universal and mean great advance in capturing cases not registered by the MPS, although they still present underreporting issues.¹⁰10. Correa PRL, Assunção AA. A subnotificação de mortes por acidentes de trabalho: estudo de três bancos de dados. Epidemiol Serv Saude. 2003 dez;12(4):203-12.

11. Alvares JK, Pinheiro TMM, Santos AF, Oliveira GL. Avaliação da completude das notificações compulsórias relacionadas ao trabalho registradas por município polo industrial no Brasil, 2007-2011. Rev Bras Epidemiol. 2015 jan-mar;18(1):123-36.

12. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.^-1313. Waldvogel BC, Salim CA, coordenadores. Vinculação de bancos de dados de acidentes do trabalho fatais dos Estados de São Paulo e Minas Gerais 2006-2008: relatório técnico - banco de dados. São Paulo: Fundação Sistema Estadual de Análise de Dados, Fundação Jorge Duprat Figueiredo de Segurança e Medicina do Trabalho; 2012.

Although one of SIM's objectives is to fully cover deaths in Brazil, it is almost impossible to determine the number of deaths due to occupational injuries that are not recorded on this system. The routine data recovery, from different information sources, may be a strategy to minimize the problem. Analyzing the relation between the register of serious occupational injuries on Sinan and deaths on SIM, as well as continuously evaluating other information systems in the Workers' Health area (MPS) has been a feasible action, considering that some authors have succeeded in identifying non reported cases by using these means.¹⁰10. Correa PRL, Assunção AA. A subnotificação de mortes por acidentes de trabalho: estudo de três bancos de dados. Epidemiol Serv Saude. 2003 dez;12(4):203-12.^,1212. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.^,1313. Waldvogel BC, Salim CA, coordenadores. Vinculação de bancos de dados de acidentes do trabalho fatais dos Estados de São Paulo e Minas Gerais 2006-2008: relatório técnico - banco de dados. São Paulo: Fundação Sistema Estadual de Análise de Dados, Fundação Jorge Duprat Figueiredo de Segurança e Medicina do Trabalho; 2012.

The search of these absent events from the systems of compulsory notification, conducted through investigation in police sources and verbal autopsy, for example, is another form, although less common, of reducing the underreporting of fatal occupational injuries.¹⁴14. Pepe CCCA. Estratégias para superar a desinformação: um estudo sobre os acidentes de trabalho fatais no Rio de Janeiro [dissertação]. Rio de Janeiro (RJ): Fundação Oswaldo Cruz; 2002.^,1515. Lacerda KM, Fernandes RCP, Nobre LCC, Pena PGL. A (in)visibilidade do acidente de trabalho fatal entre as causas externas: estudo qualitativo. Rev Bras Saude Ocup. 2014 jul-dez;39(130):127-35. An easy and low cost option is to estimate underreported cases using the capture-recapture method. This method consists on searching common and exclusive data on independent information sources, in order to enable the underreporting evaluation in each one of the databases, the total collection of cases and the correction of databases.¹⁶16. Coeli CM, Veras RP, Coutinho ESF. Metodologia de captura-recaptura: uma opção para a vigilância das doenças não transmissíveis na população idosa. Cad Saude Publica. 2000 out-dez;16(4):1071-82.^,1717. Schmid B, Silva NN. Estimação de sub-registro de nascidos vivos pelo método de captura e recaptura. Rev Saude Publica. 2011 dez; 45(6):1088-98.

The objective of this study was to measure underreporting of typical fatal occupational injuries (FOI) on the Mortality Information System (SIM) in Belo Horizonte city, Brazil, in 2011.

Methods

This is a descriptive study using capture-recapture method, in order to comprehend the total FOI reported in Belo Horizonte (BH), whenever this city was the place of the incident and/or place of residence. This method was chosen due to its simplicity and easiness to perform, and it was used in the past used to estimate the population of France and the size of wild animals' populations. This method can also be used in epidemiological researches, because it allows the usage of incomplete information sources through records overlap to measure underreporting of each system.¹⁶16. Coeli CM, Veras RP, Coutinho ESF. Metodologia de captura-recaptura: uma opção para a vigilância das doenças não transmissíveis na população idosa. Cad Saude Publica. 2000 out-dez;16(4):1071-82. The method's requisites were reached for this study:¹⁷17. Schmid B, Silva NN. Estimação de sub-registro de nascidos vivos pelo método de captura e recaptura. Rev Saude Publica. 2011 dez; 45(6):1088-98. closed population (deaths of BH residents and/or that occurred in that city in 2011); unequivocal records (death); each individual presented the same probability of belonging to each sample (equiprobability); in each data source the case was captured independently from the others; and the search was performed in independent data sources (each of them considering a population sample) - SIM, Sinan and the Belo Horizonte Forensic Medicine Institute Autopsy Database (BNIML/BH). The records identifier for pairing the cases was formed based on the combination of at least two of the following attributes: full name, birth date, and mother's name. We considered as case captured the FOI registered on SIM; and as case recaptured, those registered only on Sinan and/or BNIML/BH.

According to the 2010 Demographic Census, in that year, BH had a population of 2,375,151 people and the population density was 7,167. There were 1,590,502 employed people in the city and the human development index was considered high (HDI=0.810), although the city presents areas of agglomeration and extreme poverty.¹⁸18. Instituto Brasileiro de Geografia e Estatística. Censo Demográfico de 2010 [Internet]. Rio de Janeiro: Instituto Brasileiro de Epidemiologia e Estatística; 2010 [citado 2015 nov 09]. Disponível em: Disponível em: http://censo2010.ibge.gov.br/
http://censo2010.ibge.gov.br/... 2011 was the chosen year in order to resume another study.¹²12. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.

The FOI which occurred in the city were processed on SIM and Sinan. On SIM, the cases of BH residents who died in other cities composed the municipal base through a feedback process. The BH Forensic Medicine Institute (IML/BH) conducted autopsy examinations of FOI which occurred in BH and in cities from the metropolitan region, except for Betim-MG. Therefore, the sources used in this study contain information of deaths of residents in BH, deaths occurred in BH and in other cities.

The deaths selection followed the steps below:

To ensure the data integrity, the researchers double-checked the DO and the autopsy reports. To solve doubts concerning the characteristics of the FOI in each one of the sources, the medical reports were used for complementary information. Cases in which FOI information was incorrect or inconclusive were excluded. Commuting occupational injuries were also excluded, because typical injuries are more specific and have less underreports.¹⁹19. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74.

Not all the variables/categories were available on the three studied sources. In cases of information divergence, information from SIM was considered.

The software R was used (version 3.0.1) for data description and to obtain central trend measures.

The research was approved by the Research Ethics Committee of the Municipal Health Department of Belo Horizonte (SMS/BH) - Register No. 0002.0.410.000-11 - and was conducted according to the National Health Council Resolution No. 466, dated 12 December, 2012.

Results

Of the 43 FOI cases recorded on SIM, 4 were excluded (Figure 1): 2 for not being occupational injuries and 2 for being inconclusive about the type of accident (typical or commuting). All the 46 cases recorded on Sinan were considered eligible. Of the 39 cases obtained from the BNIML/BH, 2 were excluded for being inconclusive concerning the characteristic of the occupational injury and 10 for not having occurred in BH or the person not being resident in that city (Figure 2).

After applying the capture-recapture method, the final database was composed of 54 cases of FOI: 25 victims lived in BH and 29 in other cities (Table 1).

Among the 25 deaths of BH residents, having occurred or not in that city, 16 were registered on SIM, 22 on Sinan and 12 on BNIML/BH. Of the 9 cases that were not recorded as FOI on SIM, 5 were captured only through Sinan due to the impossibility of linking the death due to aggression to work on SIM (Table 1). The other 4 cases did not occur in BH: 3 were captured by the BNIML/BH and 1 by Sinan. The underreporting of these cases happened because of the lack of connection between the death and the work by the city of occurrence. On Sinan, 3 FOI were not recorded and were captured by the BNIML/BH. With regard to the BNIML/BH, 13 cases were underreported: 7 of them occurred out of BH and the autopsy did not happen in the capital (6 were exclusively captured by Sinan and 1, by both SIM and Sinan), 5 were related only to aggression (registered only on Sinan) and 1 was considered commuting (notified by both SIM and Sinan) (Table 1).

The 29 FOI of residents of other cities that happened in BH were registered as follows: 23 on SIM, 24 on Sinan and 15 on BNIML/BH. Of the 6 cases not captured by SIM, 3 were only on Sinan, 2 were only on BNIML/BH and 1 was on both Sinan and BNIML/BH. 3 cases could not be registered as FOI on SIM, because they were aggressions related to work (Table 1). Sinan presented 5 underreported cases: 1 was captured only on SIM, 2 only on BNIML/BH and 2 on both SIM and BNIML/BH. There were 14 underreported cases on BNIML/BH: 10 were registered on both SIM and Sinan, 1 only on SIM and 3 only on Sinan. Of these 14 cases, 3 were related to aggression and 7 to traffic injuries.

Of the 54 cases, only one victim was female and 30 had brown skin color. The average age was 41.81 years old (standard deviation [SD]=12.96; coefficient of variation [CV]=0.31, and only 6 individuals had completed elementary school (Table 2). About one third of the individuals were aged between 45 and 54 years and almost half of them were single. Drivers and construction workers accounted for more than half of the victims, and 27 of 31 victims worked with formal contracts under the CLT.

The most common causes of death were traffic accidents (one third of the cases) and falls (one fifth of the cases) (Table 3). Most deaths occurred on the same day of the accident (32 cases) and on urban streets (20 cases). Among the 34 victims who did not die at the accident scene, only one did not receive medical care in a reference hospital of the metropolitan region. The average time between the accident and the death was three days.

There was FOI underreport related to traffic accidents (15/18 cases), aggressions (8/8 cases), drivers deaths (13/14 cases) and customer service workers (9/9 cases) on BNIML/BH. Incompleteness was observed for the following variables: 'marital status' (1 case), 'skin color' (5 cases), 'situation in the job market' (23 cases) and 'place of death' (5 cases).

Discussion

Data available on electronic databases were used, and this would allow the method's application to qualify these information sources, bases of the Workers' Health surveillance in Brazil. The capture-recapture allowed the identification of underreport of 15 of the 54 FOI on SIM and 8 on Sinan. Most of the victims were men, single, brown-skinned, aged 45 to 55 years old, and had low education level. Among the FOI, the most common accidents were: traffic accidents; on urban streets; involving drivers; construction workers; and individuals with formal contracts under CLT. Usually, the event occurred outside the company and the death happened the same day of the accident. On SIM, underreporting occurred for all the homicides and for 25% of deaths of BH residents that happened in other cities.

Underreporting on SIM was mainly related to the current impossibility of this system to notify deaths due to aggression, when these are related to occupational activities, and this is a major failure:⁹9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53. if SIM presented the possibility to include these notifications, in BH, for example, the underreporting would drone in one third (5 of 15 cases). Other underreport causes were (i) difficulties on exchanging information between the cities of residence and of death occurrence and the (ii) major incompleteness of the variable that relates the death to the work in the DO.¹⁹19. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74.^-2020. Alves MMM, Nomellini PF, Pranchevicius MCS. Mortalidade por acidente de trabalho no Estado do Tocantins, Brasil: estudo descritivo, 2000-2010. Epidemiol Serv Saude. 2013 abr-jun;22(2):243-54. These issues could probably be solved with investments on improving the information flow between the cities of residence and of occurrence, training and stimulating the staff responsible for filling the DO and the death investigation teams, and a bigger union between SIM managers and Workers' Health surveillance.²¹21. Smarzaro DC. A informação sobre mortes por causas externas: estudo do preenchimento da causa básica de óbito em um serviço de medicina legal no Espírito Santo. 2000-2002 [dissertação]. Rio de Janeiro (RJ): Fundação Oswaldo Cruz; 2005. The limitation in capturing these deaths may also be connected to problems in the DO flowchart (at registrars and IML, for example), irregular cemeteries (where burials without necessary documents are possible) and flaws on the integration of the municipal health managers.²²22. Figueiroa BQ, Vanderlei LCM, Frias PG, Carvalho PI, Szwarcwald CL. Análise da cobertura do Sistema de Informações sobre Mortalidade em Olinda, Pernambuco, Brasil. Cad Saude Publica. 2013 mar;29(3):475-84.

The use of IML/BH data contributed to a raise of 5 FOI on SIM and Sinan, appearing as an alternative to cases confirmation during routine investigation conducted by the Municipal Health Department (SMS). Police records have already been used - successfully - in the Rio de Janeiro state, in a research on work-related deaths.¹⁴14. Pepe CCCA. Estratégias para superar a desinformação: um estudo sobre os acidentes de trabalho fatais no Rio de Janeiro [dissertação]. Rio de Janeiro (RJ): Fundação Oswaldo Cruz; 2002.^,1515. Lacerda KM, Fernandes RCP, Nobre LCC, Pena PGL. A (in)visibilidade do acidente de trabalho fatal entre as causas externas: estudo qualitativo. Rev Bras Saude Ocup. 2014 jul-dez;39(130):127-35.

In this study, the higher underreport was found on BNIML/BH, concerning traffic accidents and aggressions. This finding reinforces the need to change a wrong but persistent perception in which occupational injuries would only be those that happen inside the company, and this could allow a more effective surveillance through training actions.¹⁴14. Pepe CCCA. Estratégias para superar a desinformação: um estudo sobre os acidentes de trabalho fatais no Rio de Janeiro [dissertação]. Rio de Janeiro (RJ): Fundação Oswaldo Cruz; 2002.^-1515. Lacerda KM, Fernandes RCP, Nobre LCC, Pena PGL. A (in)visibilidade do acidente de trabalho fatal entre as causas externas: estudo qualitativo. Rev Bras Saude Ocup. 2014 jul-dez;39(130):127-35.

For being preventable and avoidable deaths, the number of FOI found shows a vulnerability of work conditions, still not solved by the current laws in the country.¹⁶16. Coeli CM, Veras RP, Coutinho ESF. Metodologia de captura-recaptura: uma opção para a vigilância das doenças não transmissíveis na população idosa. Cad Saude Publica. 2000 out-dez;16(4):1071-82. With regard to the victims profile, other researchers have found similar results to this study: adult men,⁹9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.^,1212. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.^,1919. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74.^,2020. Alves MMM, Nomellini PF, Pranchevicius MCS. Mortalidade por acidente de trabalho no Estado do Tocantins, Brasil: estudo descritivo, 2000-2010. Epidemiol Serv Saude. 2013 abr-jun;22(2):243-54. with lower education level⁹9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.^,1212. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.^,1919. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74. and brown-skinned.¹⁹19. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74. Typical traffic injuries and falls are the main causes of death, related to the occupation of driver and construction worker, respectively. It is known that the relation between the most common accidents and occupations may vary according to the studied region, productive profile, risk exposure, informality on work relations, existence or not of vulnerable groups, and other specific characteristics related to each population.⁸8. Gonçalves KR. Análise especial dos acidentes de trabalho assentados no Sistema de Informação de Agravos de Notificação (SINAN) em uma capital brasileira [dissertação]. Belo Horizonte (MG): Universidade Federal de Minas Gerais; 2012.^,2323. Santana VS, Araújo-Filho JB, Silva M, Albuquerque-Oliveira PR, Barbosa-Branco A, Nobre LCC. Mortalidade, anos potenciais de vida perdidos e incidência de acidentes de trabalho na Bahia, Brasil. Cad Saude Publica. 2007 nov;23(11):2643-52. Local studies are necessary to better understand the profile of each region and to establish specific strategies aiming at handling occupational injuries.²³23. Santana VS, Araújo-Filho JB, Silva M, Albuquerque-Oliveira PR, Barbosa-Branco A, Nobre LCC. Mortalidade, anos potenciais de vida perdidos e incidência de acidentes de trabalho na Bahia, Brasil. Cad Saude Publica. 2007 nov;23(11):2643-52.

Deaths due to aggression and traffic accidents among workers point to the relevance of urban violence in Brazil and the expansion of occupational environment, much beyond the companies' rooms, reaching the streets and the facilities of other people^.99. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.^,1919. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74.^,2424. Lucca SR, Mendes R. Epidemiologia dos acidentes do trabalho fatais em área metropolitana da região sudeste do Brasil, 1979-1989. Rev Saude Publica. 1993 jun; 27(3):168-76.^-2525. Miranda FMDA, Sarquis LMM, Cruz EDA, Kirchhof ALC, Scussiato LA. Caracterização das vítimas e dos acidentes de trabalho fatais. Rev Gauch Enferm. 2012 jun;33(2):45-51. The proportion of deaths due to aggression related to work, similarly to what was described by other authors, shows the severity and complexity of a problem that requires multilateral intervention. In the metropolitan region of Campinas, São Paulo state, a research on data of the National Social Security Institute (INSS) related to the period 1979-1989 related 9.2% of deaths due to occupational injuries to homicides.²⁴24. Lucca SR, Mendes R. Epidemiologia dos acidentes do trabalho fatais em área metropolitana da região sudeste do Brasil, 1979-1989. Rev Saude Publica. 1993 jun; 27(3):168-76. According to INSS and DO data of the 1997-1999 period, this proportion was of 18.8% in São Paulo state; in a sentinel unit in Curitiba, Paraná state, it was 12% (2006 to 2010); and according to Salvador IML, in Bahia state, it was 37.3% in 2004.⁹9. Waldvogel BC. A população trabalhadora paulista e os acidentes do trabalho fatais. Sao Paulo Perspect. 2003 abr-jun; 17(2):42-53.^,1919. Lacerda KM, Fernandes RCP, Nobre LCC. Acidentes de trabalho fatais em Salvador, BA: descrevendo o evento subnotificado e sua relação com a violência urbana. Rev Bras Saude Ocup. 2014 jan-jun;39(129):63-74.^,2525. Miranda FMDA, Sarquis LMM, Cruz EDA, Kirchhof ALC, Scussiato LA. Caracterização das vítimas e dos acidentes de trabalho fatais. Rev Gauch Enferm. 2012 jun;33(2):45-51. Although these studies have adopted different methods, all of them point to a common evidence: the magnitude of homicides related to work in Brazilian big cities. In BH (2011), the proportion of occupational injuries related to homicides was 14.8%, slightly higher than the 13% found to the same city between 2008 and 2010, when only SIM and Sinan data were used.¹²12. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5.

A limitation concerning the results presented in this study is that the conclusions should be cautiously interpreted because the data analyzed were obtained from a specific region, with singular investigation flow, comprising only a one-year period. There are some singularities of the SMS/BH on searching and identifying FOI cases, such as the communications of occupational injuries - CAT - as main source of investigation of cases registered on Sinan, to which can be added the active search on the Internet, in trauma reference hospitals, and in the IML, and the continuous data crossing between SIM and Sinan.¹²12. Drumond EF, Silva JM. Avaliação de estratégia para identificação e mensuração dos acidentes de trabalho fatais. Cien Saude Coletiva. 2013 mai;18(5):1361-5. Another fact to be considered is that, regardless of the IML belonging to the state, there are administrative singularities related to how it works in the different Brazilian states, which can affect the access, by the various SMS, to data related to FOI search. Particularly, the IML/BH is attached to the Civil State Police and has objective flow, allowing and favoring investigation of deaths by the SMS staff. Sinan has showed to be the most important information source to capture cases; however, we observed that most notifications referred to formal workers under the CLT, and this fact limited the understanding of risks related to occupational activities of a great amount of workers not covered by the Work-related Injury Insurance (SAT). To reach all the workers, the sentinel units of notification (Reference Centers in Workers' Health [Cerest]; reference hospitals in urgency and emergency care and/or medium and high complexity care; primary and medium complexity health care accredited as sentinels), must be encouraged to intensify its performance, enriching the registers' correction. In FOI cases, notification is mandatory in these sentinel units, and is to be performed by any health professional, and also by the responsible staff for those establishments.²⁶26. Brasil. Ministério da Saúde. Portaria nº104, de 25 de janeiro de 2011. Define as terminologias adotadas em legislação nacional, conforme o disposto no Regulamento Sanitário Internacional 2005 (RSI 2005), a relação de doenças, agravos e eventos em saúde pública de notificação compulsória em todo o território nacional e estabelece fluxo, critérios, responsabilidades e atribuições aos profissionais e serviços de saúde. Diário Oficial da República Federativa do Brasil, Brasília (DF), 2011 jan 26; Seção 1:37. The inclusion of new notifiers sentinel units, such as the IML, could also contribute hugely to reduce underreporting.⁸8. Gonçalves KR. Análise especial dos acidentes de trabalho assentados no Sistema de Informação de Agravos de Notificação (SINAN) em uma capital brasileira [dissertação]. Belo Horizonte (MG): Universidade Federal de Minas Gerais; 2012. The flaws in reporting may be related to the employers, to the employee and/or the health professional who received the injured worker, for unawareness of the reporting obligation, for the incorrect characterization of cases, or for the fear of legal implications. Finally, the incompleteness/loss of information observed in this study, even after comparing three data sources, is still a challenge to overcome, so that more quality data are available.¹¹11. Alvares JK, Pinheiro TMM, Santos AF, Oliveira GL. Avaliação da completude das notificações compulsórias relacionadas ao trabalho registradas por município polo industrial no Brasil, 2007-2011. Rev Bras Epidemiol. 2015 jan-mar;18(1):123-36.

The information systems will become more reliable as long as their limitations are diagnosed and overcome.¹¹11. Alvares JK, Pinheiro TMM, Santos AF, Oliveira GL. Avaliação da completude das notificações compulsórias relacionadas ao trabalho registradas por município polo industrial no Brasil, 2007-2011. Rev Bras Epidemiol. 2015 jan-mar;18(1):123-36. In this study, the inclusion of one more information source through the capture-recapture method was essential to recover underreported cases on the Mortality Information System - SIM -, besides showing important limitations of this system, such as the current impossibility of relating deaths due to aggression at work and the underreporting of residents deaths by the occurrence city.

The incentive and application of the capture-recapture method in health services that work with epidemiological data would allow to quantify underreporting of the available data sources, through registers overlap. In this study, as the focus was evaluating FOI, the underreport obtained would be more substantial if the commuting accidents were included, because they are probably more underreported. This is a suggestion for future and wider researches on the topic, as a last recommendation in this study.

Acknowledgements

To Dr. João Batista Rodrigues Júnior, responsible for encouraging this paper conduction, for authorizing the study as a Board member of the Belo Horizonte Forensic Medicine Institute - IML/BH -, for all the support, essential to the study conduction.

References

Publication Dates

History