Artigos

Conteúdo (article):

Vulnerable populations and tuberculosis treatment outcomes in Brazil

Abstract
Objective: To assess the association between being a prisoner or homeless and treatment failure in cases of tuberculosis diagnosed in Brazil in 2015.
Methods: We examined cases of tuberculosis in prisoners and the homeless in Brazil in 2015 reported to the national notifiable diseases information system using descriptive analysis and logistic regression.
Results: There were 82,056 cases of tuberculosis in 2015. Of these, 7,462 (10.3%) were prisoners and 2,782 (3.9%) were homeless. The rate of treatment success in prisoners was 78.6%, while the rate of failure in the homeless was 63.2%. Being a prisoner was a protective factor against treatment failure (adjusted odds ratio 0.68 – 95% CI 0.63-0.73), while being homeless was a risk factor for treatment failure (adjusted odds ratio 2.38 95% CI 2.17-2.61).
Conclusions: Treatment success and failure rates differed between prisoners and the homeless. Our findings reinforce the need for public health policies tailored to the specific needs of these groups implemented in conjunction with social services and public security agencies in order to have a significant impact on TB incidence.
Keywords: Tuberculosis; Vulnerable Populations, Prisons; the Homeless.

Introduction
Tuberculosis (TB) is the leading causes of death from an infectious disease worldwide (1). It is known that poverty is a cause of TB and that the disease is also a cause of poverty, with this vicious circle being played out at an individual, household and community level (1).
Studies point to a direct association between socioeconomic factors and the occurrence of the disease, both at the individual and collective level, emphasizing that TB is intimately related to an individual’s living conditions and social environment (2).
Specific groups are especially vulnerable to the disease (3), with data showing that the risk of falling ill with TB is 28 and 56 times greater among prisoners and the homeless, respectively, than in the general population.
The literature recognizes a number of factors that may influence the outcome of the treatment of tuberculosis, including: sex, age, level of education, form of TB, and associated problems such as AIDS, diabetes, mental illness, alcoholism, smoking, and drug use (4). However, there is a lack of information on the factors affecting disease outcome among prisoners and the homeless in Brazil.
Given the complexity of the social determinants of TB, pillar 2 of Brazil’s Post-2015 Tuberculosis Control Strategy recommends the implementation of social protection policies directed at patients and universal access to healthcare (5), meaning that it is essential to understand illness behavior in these vulnerable groups.
In view of the above, the objective of this study was to assess the association between being a prisoner or homeless and treatment failure in cases of tuberculosis diagnosed in Brazil in 2015.

Methods
The study population consisted of cases of tuberculosis diagnosed in Brazil in 2015 reported to Brazil’s national notifiable diseases information system (SINAN in Portuguese). Individuals aged under 15 years were excluded because disease behavior differs in the age group.
For the purposes of this study, vulnerable populations were defined as prisoners or homeless persons, based on the information in the item “special situations” on the disease notification form. This item was added to the form in the latest revision at the end of 2014.
Case outcomes were defined based on the categories of the SINAN’s tuberculosis follow-up form, as follows: cure, abandonment of treatment (loss to follow-up), death from TB, death from other causes, transfer, multidrug-resistant tuberculosis (MDR-TB), change in scheme, and failure and primary abandonment (primary loss to follow-up). For the regression analysis, these categories were regrouped into treatment success (cure) or failure (all other categories).
Other potential confounding factors were included based on the literature, as follows: sex, age, race/skin color, level of education, region of residence, type of TB case, form of tuberculosis, associated illnesses and problems (AIDS, diabetes mellitus, mental illnesses, alcoholism, smoking, and drug use), examinations (x-ray, sputum smear microscopy and culture), and directly observed therapy. It is important to note that with regard to level of education, the category “not applicable” is used for patients aged under seven years.
The data used in the study were obtained from the SINAN database updated in 2017. The SINAN is a national information system designed to support the collection and processing of data on diseases and public health problems and events, including TB. The database provides valuable inputs to help in planning and disease prevention, assessment and control, and serves as an important epidemiological surveillance tool (6).
The cases were presented according to sociodemographic and clinical characteristics and grouped into the following population types: total population (all notified cases), prisoners and nonprisoners, and the homeless and nonhomeless. For each variable, the chi-squared test or Fisher\'s exact test were used to compare proportions across the populations and control groups.
We used two logistic regression models to separately assess the association between being a prisoner or being homeless and treatment failure. For this purpose, we performed multiple imputation of missing data following the steps suggested by Harrell (7). First, we examined the pattern of the unregistered data, observing whether the missing data of a given variable were related to the filled in data for another variable in the model - “missing at random” (MAR). For this purpose, we used the “predictive mean matching” (PMM) method, which replaces the missing data for a variable with the actual value from a donor observation (predicted mean) (7). The following variables were included in the imputation: race/skin color, level of education, type of TB case, form of TB, associated problems (AIDS, diabetes mellitus, mental illnesses, alcoholism, smoking, and drug use), x-ray examination, sputum smear microscopy and culture, directly observed therapy, and prisoners and the homeless. Ten datasets were generated for the imputation using the variables included in the final models as auxiliary variables.
To obtain an estimation of the association between being a prisoner or homeless and treatment failure we adjusted the logistic regression models including all a priori-selected potential confounders. In the final logistic regression models based on the imputed data, the hypothesis was tested using the Wald test and Wald intervals for 95% confidence intervals (95% CI).
The statistical analyses were performed using the R program 3.4.1 and the Hmisc and rms packages were used for imputation.
The study was approved by the Oswaldo Cruz Foundation’s Sérgio Arouca National School of Public Health Research Ethics Committee (approval code number 1.866.469, 14/12/2016).

Results
There were 84,405 reported cases of tuberculosis in Brazil in 2015. A total of 2,349 (2.8%) individuals aged under 15 years were excluded, resulting in a final study sample of 82,056 cases. Of these, 7,462 (10.3%) were prisoners and 2,782 (3.9%) were homeless persons.
The sociodemographic characteristics of prisoners and the homeless were similar, with the majority being male (95.6% and 79.9%, respectively), aged between 15 and 39 years (85.3% and 53.1%, respectively), brown (44.1% and 44.1%, respectively), having between 5 and 8 years of schooling (34.7% and 27.6%, respectively), and living in Brazil’s Southeast Region (53.6% and 54.5%, respectively) (Table 1).
The comparison of the clinical characteristics of the populations revealed that prisoners showed more new cases (75.2% vs 56.0%), less reenrollment after loss to follow-up (11.9% vs 33.7%), lower prevalence of AIDS (7.6% vs 23.8%), diabetes (1.4% vs 3.1%), mental illnesses (1.2% vs 6.7%), alcoholism (10.5% vs 52.5%), smoking (21.0% vs 42.2%), and drug use (19.2% vs 53.0%), and higher prevalence of directly observed therapy (42.0% vs 33.2%). With regard to case outcome, the homeless showed a lower frequency of cure (34.7% vs 67.9%) and higher frequency of loss to follow-up (33.7% vs 7.6%), death from TB (5.9% vs 1.0%), and death from other causes (5.7% vs 1.6%) (Table 2).
The data show that the rate of treatment success was slightly higher in prisoners (78.6%) than in the total population (70.4%), while in the homeless the rate of treatment failure (63.2%) was considerably higher than the rate of success (36.8%) (Graph 1).
It is worth highlighting that this study encompassed all types of TB cases. In this regard, a comparison between all types of TB cases and only new cases by population group showed that the rate of cure (treatment success) was slightly higher in new cases across all populations (total population, prisoners and the homeless).
Table 3 shows the results of the final separate models for prisoners and the homeless. The findings show that being a prisoner was a protective factor against treatment failure (adjusted OR 0.68 95% CI 0.63-0.73), while being homeless was a risk factor for treatment failure (adjusted OR 2.38 95% CI 2.17-2.61), even after adjustment.

Discussion
Prisoners and the homeless accounted for 10.3% and 3.9%, respectively, of all notified cases of tuberculosis in Brazil in 2015. The sociodemographic characteristics of the prisoners and the homeless in our sample are similar to those of diagnosed cases in these populations reported by previous studies (8,9), reinforcing the association between falling ill with tuberculosis and individual and socioenvironmental factors (2,10).
It is worth highlighting that the rate of reenrollment after loss to follow-up was considerably higher in the homeless than in the nonhomeless (33.75% vs 8.5%). Factors such as drinking every day, injecting and non-injecting drug use, and unemployment contribute significantly to this problem (11). The prevalence of associated problems was also higher in the homeless than in the other populations, highlighting a common characteristic among this group (12,13) that directly influences TB treatment. The prevalence of associated problems in prisoners was similar to that of the total population.
The proportion of cases receiving directly observed therapy (DOT) was higher in prisoners than in nonprisoners (42.6% vs 34.2%). These findings are similar to those reported by a study in the United States that showed that inmates with TB were more likely to receive DOT than noninmates (65.0% vs 41.0%) (14). Although the proportion of directly observed therapy was higher among prisoners than in the other groups, it is important to note that the percentage is substantially lower than the 100% recommended by the Ministry of Health (15).
Although both prisoners and the homeless are vulnerable groups and TB incidence was greater in these groups than in the general population (15), our findings show that case outcome differed between these populations.
Data reported to the US national TB surveillance system from 1993 through 2003 showed that inmates with TB were more likely to have at least one TB risk factor compared with noninmates (60.1% vs 42.0%) and less likely to complete treatment (76.8% vs 89.4%) (14). However, the proportion of inmates who showed a favorable outcome was similar to the rate found for the prisoners in the present study (78.6%). A study in El Salvador covering the period 2009 to 2013 reported that treatment success rates in prisons were over 95% (16).
Although we were unable to find other studies in Brazil that showed an association between being a prisoner and favorable case outcomes (protective factor against treatment failure) (OR 0.68, 95% CI 0.63-0.73), we believe that this finding may be associated with the benefits of receiving DOT. A study in Brazil showed that receiving DOT led to a 25% reduction in unfavorable TB treatment outcomes (17). Our findings show that receiving DOT was associated with treatment success in both prisoners and the homeless.
A number of factors may hamper the continuation of treatment among the homeless, including: food insecurity, associated health problems, alcoholism and drug use, low self-esteem, difficulty recognizing symptoms, and the fact that living conditions make it difficult for homeless people to take medications regularly due theft or loss of medicines (18).
Our findings show that the rate of treatment failure among the homeless was 63.2% and that the adjusted odds ratio for unfavorable outcomes (failure) was 2.38 (95% CI 2.17-2.61). Similar results were found by other studies: a study in São Paulo (9) reported a 57.3% failure rate and an OR of 4.96 (95% CI 4.27-5.76); in London(19), homelessness was associated with multidrug resistance (OR 2.1), poor adherence (OR 2.5), and loss to follow-up (OR 3.8); and in Nicaragua (20), being homeless was associated with loss to follow-up (OR 3.08, 95% CI 1.57-6.49).
A cohort study conducted in the United States during the period 1994 to 2010 showed that the homeless were twice as likely to not complete treatment (21). Most of the homeless TB patients were young male adults and the sample showed a high prevalence of excessive alcohol use, drug use and HIV infection (21), as observed in the present study.
It is important to emphasize that issues related to difficulty of access to health facilities such as lack of flexibility of opening hours aggravate this problem (18).
This study has some limitations. First, the “special situations” item on the SINAN notification form was only introduced in 2014, meaning that the number of homeless persons in 2015 may be underestimated as health professionals were getting used to the new form during the study period. In addition, both the homeless and prisoners are socially stigmatized groups, which means that these persons may have been ashamed to admit that they belong to these groups at the time of notification.
Although the National Tuberculosis Program uses tools designed to control the quality of SINAN data, there was a large amount of incomplete data for certain variables, meaning it was necessary to use multiple imputation. In this regard, the comparisons of the analyses using the imputed data and the analysis using complete data showed similar values. It is important to note that the SINAN is a well-structured national information system and the advantages of using this system include the low cost of data collection and broad coverage of data (6).
Considering the need for a social inclusion policy that addresses the promotion of the human rights of prisoners, in 2003, the ministries of health and justice introduced the National Penitentiary System Health Plan (PNSSP in Portuguese), which reoriented health care for this population. A decade later, the Plan was reviewed and updated, resulting in the introduction of the National Policy for Comprehensive Health Care for Prisoners in the Prison System (PNAISP in Portuguese) in January 2014. The primary objective of the PNAISP is to guarantee prisoners access to comprehensive health care under Brazil’s public health system, the Sistema Único de Saúde (SUS) or Unified Health System. The PNAISP provides that prison system health services should be part of the SUS’s Health Care Network, confirming primary care within the prison system as the front door of the system and organizer of health actions and services throughout the network (22,23). These policies have contributed significantly to the improvement of health care in the prison system and to the diagnosis, treatment and control of tuberculosis among prisoners.
It is important to highlight that the main factors related to tuberculosis in prisoners and the homeless are social and institutional, resulting from social inequality and barriers to access to health services. Actions that address only biological aspects of the disease are therefore insufficient for infection control.
Cash transfer programs have made an effective contribution to reducing TB incidence (24) and curing the disease (25). Research has also shown the benefits of DOT, suggesting that that this strategy should be expanded, especially among the homeless. In 2011, the National Primary Care Policy (PNAB in Portuguese) introduced the Consultório na Rua or Street Clinics, aimed at improving access to health services for the homeless by providing timely comprehensive care to this group. Street Clinics have multiprofessional teams that provide mobile care in partnership with conventional primary care teams (18), playing an important role in promoting DOT among the homeless.
Given the complexity of the context of tuberculosis in prisoners and the homeless revealed by this study, public health policies need to be tailored to the specific needs of these groups and implemented in conjunction with social services and public security agencies in order to have a significant impact on TB incidence.

References
1. World Health Organization (WHO). Global Tuberculosis Report 2016. 2016.
2. San Pedro A, Oliveira RM de. Tuberculosis and socioeconomic indicators: systematic review of the literature. Rev Panam Salud Pública. 2013 Apr;33(4):294–301.
3. Populações Vulneráveis - Tuberculose [Internet]. Portal da Saúde – Ministério da Saúde – www.saude.gov.br. [cited 2017 Jul 25]. Available from: http://portalsaude.saude.gov.br/index.php/o-ministerio/principal/leia-mais-o-ministerio/743-secretaria-svs/vigilancia-de-a-a-z/tuberculose/l2-tuberculose/11941-viajantes-tuberculose
4. Maciel EL, Reis-Santos B. Determinants of tuberculosis in Brazil: from conceptual framework to practical application. Rev Panam Salud Publica Pan Am J Public Health. 2015 Jul;38(1):28–34.
5. Maciel ELN. Estratégias da agenda pós-2015 para o controle da tuberculose no Brasil: desafios e oportunidades. Epidemiol E Serviços Saúde. 2016 Jun;25(2):423–6.
6. Malhão TA, Oliveira GP de, Codennoti S, Moherdaui F. Avaliação da completitude do Sistema de Informação de Agravos de Notificação da Tuberculose, Brasil, 2001-2006. Epidemiol E Serviços Saúde. 2010 Sep;19(3):245–56.
7. Harrell FE. Regression Modeling Strategies [Internet]. Cham: Springer International Publishing; 2015 [cited 2017 Sep 28]. (Springer Series in Statistics). Available from: http://link.springer.com/10.1007/978-3-319-19425-7
8. Ribeiro Macedo L, Reis-Santos B, Riley LW, Maciel EL. Treatment outcomes of tuberculosis patients in Brazilian prisons: a polytomous regression analysis. Int J Tuberc Lung Dis Off J Int Union Tuberc Lung Dis. 2013 Nov;17(11):1427–34.
9. Ranzani OT, Carvalho CRR, Waldman EA, Rodrigues LC. The impact of being homeless on the unsuccessful outcome of treatment of pulmonary TB in São Paulo State, Brazil. BMC Med [Internet]. 2016 Mar 23 [cited 2017 Jul 24];14. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4804546/
10. Hargreaves JR, Boccia D, Evans CA, Adato M, Petticrew M, Porter JDH. The Social Determinants of Tuberculosis: From Evidence to Action. Am J Public Health. 2011 Apr;101(4):654–62.
11. Feske ML, Teeter LD, Musser JM, Graviss EA. Counting the Homeless: A Previously Incalculable Tuberculosis Risk and Its Social Determinants. Am J Public Health. 2013 May;103(5):839–48.
12. Halpern SC, Scherer JN, Roglio V, Faller S, Sordi A, Ornell F, et al. Clinical and social vulnerabilities in crack users according to housing status: a multicenter study in six Brazilian state capitals. Cad Saúde Pública [Internet]. 2017 [cited 2017 Oct 31];33(6). Available from: http://www.scielo.br/scielo.php?script=sci_abstract&pid=S0102-311X2017000605002&lng=en&nrm=iso&tlng=en
13. Prevalence and vulnerability of homeless people to HIV infection in São Paulo, Brazil [Internet]. [cited 2017 Oct 31]. Available from: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0034-89102012000400012
14. MacNeil JR, Lobato MN, Moore M. An unanswered health disparity: tuberculosis among correctional inmates, 1993 through 2003. Am J Public Health. 2005 Oct;95(10):1800–5.
15. Manual de recomendações para o controle da tuberculose no Brasil. Brasília/DF: Ministério da Saúde: Secretaria de Vigilância em Saúde: Departamento de Vigilância das Doenças Transmissíveis; 2019. 364 p.
16. Ayala G, Garay J, Aragon M, Decroo T, Zachariah R, Ayala G, et al. Trends in tuberculosis notification and treatment outcomes in prisons: a country-wide assessment in El Salvador from 2009–2014. Rev Panam Salud Pública. 2016 Jan;39(1):38–43.
17. Reis-Santos B, Pellacani-Posses I, Macedo LR, Golub JE, Riley LW, Maciel EL. Directly observed therapy of tuberculosis in Brazil: associated determinants and impact on treatment outcome. Int J Tuberc Lung Dis Off J Int Union Tuberc Lung Dis. 2015 Oct;19(10):1188–93.
18. Manual sobre o cuidado à saúde junto a população em situação de rua. Brasília/DF: Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Atenção Básica; 2012. 98 p. (Série A. Normas e Manuais Técnicos).
19. Story A, Murad S, Roberts W, Verheyen M, Hayward AC. Tuberculosis in London: the importance of homelessness, problem drug use and prison. Thorax. 2007 Aug 8;62(8):667.
20. Soza Pineda NI, Pereira SM, Barreto ML. [Dropout from tuberculosis treatment in Nicaragua: the results of a comparative study]. Rev Panam Salud Publica Pan Am J Public Health. 2005 Apr;17(4):271–8.
21. Bamrah S, Woodruff RSY, Powell K, Ghosh S, Kammerer JS, Haddad MB. Tuberculosis among the homeless, United States, 1994–2010. Int J Tuberc Lung Dis. 17(11):1414–9.
22. Ministério da Saúde. Plano Nacional de Saúde no Sistema Penitenciário (PNSSP). Brasília/DF: Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Ações Programáticas Estratégicas. Área Técnica de Saúde no Sistema Penitenciário; 2004. 62 p.
23. Ministério da Saúde. Legislação em Saúde no Sistema Prisional. Brasília/DF: Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Ações Programáticas Estratégicas. Coordenação de Saúde no Sistema Prisional; 2014. 93 p.
24. Nery JS, Rodrigues LC, Rasella D, Aquino R, Barreira D, Torrens AW, et al. Effect of Brazil’s conditional cash transfer programme on tuberculosis incidence. Int J Tuberc Lung Dis Off J Int Union Tuberc Lung Dis. 2017 Jul 1;21(7):790–6.
25. Torrens AW, Rasella D, Boccia D, Maciel ELN, Nery JS, Olson ZD, et al. Effectiveness of a conditional cash transfer programme on TB cure rate: a retrospective cohort study in Brazil. Trans R Soc Trop Med Hyg. 2016 Mar;110(3):199–206.