Tuberculosis (TB) is a worldwide health problem. One-third of the world’s population may be infected with Mycobacterium tuberculosis. Brazil is among the 22 countries with the highest incidence of TB, with a rate of 43 cases per 100,000 inhabitants in 2010 [1]. In the world and in Latin America, studies show a high incidence of TB in indigenous populations, exceeding that of the general population [2–7].

Since 1998, TB control has been the responsibility of Brazil’s primary healthcare [8]. The decentralization strategy aims to offer health services related to disease diagnosis and to institute the DOT for all TB patients [9, 10]. In recent years, many efforts have been made to increase the coverage of Primary Health Care (PHC). Although 50% of Brazilians have coverage for primary care [11], the majority of diagnoses of tuberculosis are performed in the emergency department or in secondary and tertiary hospitals [12, 13]. PHC presents the longest time to diagnosis and the lowest proportion of diagnoses [14]. In addition, increased coverage does not result in an improvement in the success rate of treatment [15].

Operational studies are important because they provide information that may help services focus their practice more precisely and efficiently [16]. Although several studies have evaluated health services for tuberculosis control in Brazil [12–14, 17–19], no study has evaluated the decentralization of healthcare in the context of the indigenous population.

Dourados was a priority city for disease control determined by the National TB Control Program and includes the largest indigenous population of the State of Mato Grosso do Sul with a high incidence of TB. The incidence of TB in the urban population of Dourados is 44 cases per 100,000 inhabitants, and the incidence among the indigenous people of this municipality is even higher, at 230 per 100,000 inhabitants [20].

Before 2000, TB treatment for indigenous patients was hospitalization for several months, and the default rate was 20% [21, 22]. Croda et al. [20] retrospectively evaluated the performance of health services for this population and found that with the implementation of the Directly Observed Treatment Strategy (DOTS), there was a considerable reduction in default to treatment of 2%. However, this study has some limitations and potential confounding factors. The secondary database is also used for epidemiological surveillance; therefore, the results found in this study are limited by the absence of important variables related to treatment default, as described in the literature, such as social characteristics, smoking, malnutrition, housing conditions, intravenous drug use, socioeconomic status, and access to health services. Another relevant limitation is the incompleteness of some of the information, such as the number of medical appointments and the time required to establish the diagnosis of tuberculosis.

To clarify these initial findings, a cross-sectional study was proposed to evaluate the performance of two different TB control programs (indigenous and non-indigenous) in Dourados using a standardized tool to evaluate the performance of health services.

One hundred and eighty-six patients with TB were notified during the study period, and 46 patients (24.7%) were excluded for not having a confirmed diagnosis or being in a seclusion regimen. Among the remaining 140 patients, 13 had follow-up losses due to our inability to locate, 11 died before the end of treatment, and six patients refused to participate in the study. Four non-indigenous patients presented treatment default. The cure rate was 90.6% versus 81.5%, mortality rate 9.3% versus 6.7% and default rate of 0% versus 5.2% for indigenous and urban populations, respectively. Thus, 109 patients constituted the sample evaluation of health services (Figure 2).

Regarding losses, no significant differences were found between the two groups. A total of 12 children required legal representatives to answer the PCAT, including 2/52 (3.8%) in the non-indigenous population and 10/57 (17.5%) in the indigenous population. There was no difference between those who self-responded and those who did not. The internal consistency coefficient (Cronbach’s alpha) ranged from 0.71 to 0.80 for the Likert scale variables.

Despite the differences between the indigenous and non-indigenous groups, the time to diagnosis was shown to be unsatisfactory in both groups. Previous studies have also found delayed TB diagnoses [28–32]. As a result of this delay, the risk of transmission and the possibility of death remain high [13]. The evolution to cure was higher among the indigenous populations (90.6 vs 81.5%, the mortality rate was similar among indigenous and urban population (9.3% vs 6.7%). The default rate is 0% in the indigenous versus 5.2% in the urban populations. This findings suggesting the superiority of indigenous tuberculosis control program related to DOT covered [10].

Indigenous families are large and have more people per room than non-indigenous families (2.23 ± 1.69 vs. 0.99 ± 1.67; < 0.01). The size and number of rooms in the home together with aglomeration are factors that influence the risk of TB transmission [33] and can be associated with the high incidence of the diseases in the indigenous populations [20]. The risk of developing TB in households with four or more people is approximately three times higher compared to households with two people or fewer [33]. Recent findings by Welch and Coimbra Jr. [34], analyzing cultural perspectives of tuberculosis transmission and treatment among the Xavante of Mato Grosso State, Brazil support this idea.

In this study, home visits for the treatment of tuberculosis occurred less frequently in the non-indigenous population (4.57 ± 0.89 vs. 1.68 ± 4.1 for non-indigenous people, p <0.01). A study conducted by Croda et al. [20] observed lower rates of DOT coverage for this population when compared with the indigenous population (non-indigenous 35% vs. indigenous 92%, p <0.001), further highlighting the higher default rates (17% vs. 2%; p <0.001). Amaral et al. [35] evaluated the decentralization program to control tuberculosis between 2003 and 2006 and found a worsening in the percentage of default (7.7% to 10.2%) and a significant increase in the percentage of deaths (3.0% to 15.4%). Despite decentralization, primary care does not make frequent home visits, which hinders compliance and the completion of treatment.

The WHO highlights the importance of the organizational and performance dimensions of healthcare services (HS), stating that the problem does not lie in detecting or treating TB, but in how healthcare services are organized to detect and treat cases of TB [36]. This perspective emphasizes the need to expand the diagnosis of tuberculosis and to offering more X-rays and smears in primary healthcare. This approach would provide the opportunity for earlier diagnosis by decreasing the time to diagnosis and the number of late diagnoses in emergency and secondary or tertiary services.

Although the majority of patients sought medical care at the PHC, the current study showed that most TB diagnoses were performed in emergency departments and hospitals. The healthcare model found in the municipality explains this trend because the Mission Hospital specializes in TB diagnosis among the indigenous population, whereas University Hospital primarily serves the non-indigenous population. These findings are similar to several studies conducted in Brazil [12, 13].

Supplying HIV/AIDS exams did not result in satisfactory performance, particularly among the indigenous population. In Brazil, it is recommended that the investigation of HIV should be supplied to 100% of TB patients [10]. This exam is performed only in the Sexually Transmitted Diseases reference service, with no corresponding service in indigenous reserves. The implementation of rapid HIV testing in primary care service would facilitate access to the exam, as observed in other studies with neglected populations [37–39].

The indigenous patients felt that they were not as well understood as the non-indigenous patients when they asked health questions. Indigenous people consider diseases to be individual, a sign of social or world disorder, which contradicts the understanding of occidental biomedicine [40]. Sensitivity to the culture and its conceptions about the health-disease process is essential for improving the doctor-patient relationship [41, 42].

This study has limitations related to possible sample selection bias because of participant attrition related to death and treatment default, which led to the non-completion of the questionnaire and the exclusion of data from the analyses. Communication difficulties with the indigenous community were not reduced even in the presence of an indigenous health worker. This difficulty was due to both the low level of education and the language barriers, which resulted in a lack of understanding of certain terms and phrases as well as difficulty in responding according to the Likert scale. The instrument is extensive and requires response measures. The application of the PCAT for the indigenous population may not be the best alternative, although it is the only validated instrument in Brazil for the evaluation of health services in primary tuberculosis.

Despite these limitations, this study showed that after operating the decentralization process, there were improvements in TB control among the indigenous population, with home visits providing a reduction in the incidence rate. In the non-indigenous population, despite decentralization, there is a lack of access to diagnosis and a need for greater coverage and commitment of health professionals in providing the DOT.

Despite the differences between the indigenous and non-indigenous groups, the time to diagnosis was shown to be unsatisfactory in both groups. In particular, the rapid HIV-AIDS test should be offered through primary healthcare, especially in indigenous communities. For non-indigenous people, improvements in access to diagnosis and treatment are necessary, especially with regard to greater coverage of DOT. Incentives for the strengthening of primary healthcare with the co-responsibility of other services may contribute to effective improvements in TB treatment.