Abstract

This study aimed to identify the sociodemographic profiles of women diagnosed as breast cancer in the city of Belo Horizonte and to investigate its association with interval between diagnosis and treatment. A cross-sectional study from hospital records of 715 patients undergoing treatment between 2010 and 2013. Cluster analysis was used to delineate the profiles from the variables: age, color of the skin, education and cost of treatment. The association between profiles and intervals was investigated using multinomial logistic regression. Five profiles were identified: A (white skin color, years of schooling >15 and treatment through private healthcare systems); B (white skin color, years of schooling = 11 and treatment through the Unified National Health System (SUS); C and D (brown skin color, years of schooling = 11 and < 8 respectively, and SUS); E (black skin color, years of schooling < 8, and SUS). Profiles B, C, D and E were associated with increased diagnosis-to-treatment intervals regardless of cancer staging upon diagnosis; and profile E had 37- fold higher chances of interval > 91 days (OR: 37.26; 95% CI:11.91-116.56). Breast cancer patients with social vulnerability profiles wait longer for treatment even after overcoming barriers to access oncology units.

Key words
Breast câncer; Interval to treatment; Social vulnerability; Health profile

Introduction

Malignant tumors stand out among chronic non-communicable diseases given their high incidence, mortality and treatment costs¹1. Schmidt MI, Duncan BB, Azevedo e Silva G, Menezes AM, Monteiro CA, Barreto SM. Doenças crônicas não transmissíveis no Brasil: carga e desafios atuais. Lancet 2011; 6736(11)60135-60139.. Breast cancer is the most frequent type of cancer in women, and ranks first as cause of cancer-related deaths in developing countries, and second in developed countries²2. World Health Organization (WHO). GLOBOCAN: estimated cancer incidence, mortality and prevalence worldwide in 2012. Geneva: IARC; 2014.. Worldwide, over one million women are diagnosed with cancer every year, and 40% of cases will progress to death³3. Coughlin SS, Ekwueme DU. Breast cancer as a global health concern. Cancer epidemiol 2009; 33(5):315-318.. While mortality has actually dropped in high income countries, increased breast cancer incidence and mortality rates have been documented in countries such as Brazil, Colombia and Venezuela²2. World Health Organization (WHO). GLOBOCAN: estimated cancer incidence, mortality and prevalence worldwide in 2012. Geneva: IARC; 2014.. Availability of and access to diagnostic and therapeutic technology partly explain these differences⁴4. Unger‐Saldaña K, Miranda A, Zarco‐Espinosa G, Mainero‐Ratchelous F, Bargalló‐Rocha E, Lázaro‐León JM. Health system delay and its effect on clinical stage of breast cancer: Multicenter study. Cancer 2015; 121(13):2198-2206..

The estimated incidence of breast cancer among Brazilian women in 2016 was 56.2 per 100,000. Breast cancer incidence tends to be higher in Brazilian capital cities⁵5. Instituto Nacional de Câncer (INCA). Coordenação de Prevenção e Vigilância. 2016. Rio de Janeiro. [acessado 2016 Out 15]. Disponível em: http://www.inca.gov.br/estimativa/2016
http://www.inca.gov.br/estimativa/2016... and is similar to that reported in developed countries; however, age-adjusted mortality rates are higher¹1. Schmidt MI, Duncan BB, Azevedo e Silva G, Menezes AM, Monteiro CA, Barreto SM. Doenças crônicas não transmissíveis no Brasil: carga e desafios atuais. Lancet 2011; 6736(11)60135-60139.. Despite higher five-year survival rates in the country over the last two decades (78% to 87%), under-reporting of severe cases may underestimate incidence and overestimate survival⁶6. Allemani C, Weir HK, Carreira H, Harewood R, Spika D, Wang XS, Bannon F, Ahn JV, Johnson CJ, Bonaventure A, Marcos-Gragera R, Stiller C, Azevedo e Silva G, Chen WQ, Ogunbiyi OJ, Rachet B, Soeberg MJ, You H, Matsuda T, Bielska-Lasota M, Storm H, Tucker TC, Coleman MP; CONCORD Working Group. Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2). Lancet 2015; 385(9972):977-1010..

Delays in making diagnosis and starting treatment have been associated with worse prognosis and decreased survival⁵5. Instituto Nacional de Câncer (INCA). Coordenação de Prevenção e Vigilância. 2016. Rio de Janeiro. [acessado 2016 Out 15]. Disponível em: http://www.inca.gov.br/estimativa/2016
http://www.inca.gov.br/estimativa/2016... . Delays between disease suspicion and first appointment with a cancer specialist are often associated with patient´s characteristics, such as old age, low level of education, lack of information about the disease, lack of health insurance coverage and lack of financial resources to afford medical services⁷7. Jassem J, Ozmen V, Bacanu F, Drobniene M, Eglitis J, Lakshmaiah KC, Kahan Z, Mardiak J, Pieńkowski T, Semiglazova T, Stamatovic L, Timcheva C, Vasovic S, Vrbanec D, Zaborek P. Delays in diagnosis and treatment of breast cancer: a multinational analysis. Eur J Public Health 2014; 24(5):761-767.^–99. Yoo TK, Han W, Moon HG, Kim J, Lee JW, Kim MK, Lee JW, Kim MK, Lee E, Kim J, Noh DY. Delay of Treatment Initiation Does Not Adversely Affect Survival Outcome in Breast Cancer. Cancer Res Treat 2016; 48(3):962-969.. On the other hand, delays in the intervals between appointment, diagnosis and treatment are often related to the healthcare context¹⁰10. Welle D, Vedsted P, Rubin G, Walter FM, Emery J, Scott S, Campbell C, Andersen RS, Hamilton W, Olesen F, Rose P, Nafees S, van Rijswijk E, Hiom S, Muth C, Beyer M, Neal RD. The Aarhus statement: improving design and reporting of studies on early cancer diagnosis. Br J Cancer 2012; 106(7):1262-1267..

The interval between diagnosis and treatment initiation is of particular concern in several countries. A study carried out by the Organization for Economic Co-operation and Development (OECD) on healthcare system administration, between 2001 and 2004, recommended reducing this interval to a minimum ranging from seven to 30 days¹¹11. Organização para a Cooperação e Desenvolvimento Económico (OECD). Focus on health. Cancer care. Assuring quality to improve survival. Paris: OECD/European Comission; 2013.. A review study showed that intervals of up to 60 days between confirmation of diagnosis and initiation of treatment, particularly in the initial stages of cancer, have no impact on disease-free survival or overall survival¹⁰10. Welle D, Vedsted P, Rubin G, Walter FM, Emery J, Scott S, Campbell C, Andersen RS, Hamilton W, Olesen F, Rose P, Nafees S, van Rijswijk E, Hiom S, Muth C, Beyer M, Neal RD. The Aarhus statement: improving design and reporting of studies on early cancer diagnosis. Br J Cancer 2012; 106(7):1262-1267.. In 2014, in an effort to reduce this interval, the Brazilian Ministry of Health determined that cancer treatment must be initiated within 60 days of diagnosis¹²12. Brasil. Portaria nº 1.220, de 3 de junho de 2014. Dispõe sobre a aplicação da Lei nº 12.732, de 22 de novembro de 2012, que versa a respeito do primeiro tratamento do paciente com neoplasia maligna comprovada, no âmbito do Sistema Único de Saúde (SUS). Diário Oficial da União 2014; 3 jun..

Belo Horizonte, like other Brazilian southeastern capital cities, has one of the highest incidence rates of breast cancer in the country. The estimated number of new cases in 2016 was 1020 (75.6 per 100,000)⁵5. Instituto Nacional de Câncer (INCA). Coordenação de Prevenção e Vigilância. 2016. Rio de Janeiro. [acessado 2016 Out 15]. Disponível em: http://www.inca.gov.br/estimativa/2016
http://www.inca.gov.br/estimativa/2016... . The healthcare system has a duty to provide early diagnosis and timely access to treatment for all women without distinction¹²12. Brasil. Portaria nº 1.220, de 3 de junho de 2014. Dispõe sobre a aplicação da Lei nº 12.732, de 22 de novembro de 2012, que versa a respeito do primeiro tratamento do paciente com neoplasia maligna comprovada, no âmbito do Sistema Único de Saúde (SUS). Diário Oficial da União 2014; 3 jun.. However, inequities in access to and utilization of healthcare services are observed in Brazil¹³13. Coelho IB. Democracy without reform and nineteen years of National Health System in Brazil. Cien Saude Colet 2010; 15(1):171-183.^,1414. Renk DV, Barros F, Domingues MR, Gonzalez MC, Sclowitz ML, Caputo EL, Gomes LM. Equidade no acesso ao rastreamento mamográfico do câncer de mama com intervenção de mamógrafo móvel no sul do Rio Grande do Sul, Brasil. Cad Saúde Pública 2014; 30(1):88-96..

Socioeconomic and demographic characteristics are known be important determinants of individual behavior when seeking and using health services⁸8. Rezende MCR, Koch HA, Figueiredo JA, Thuler LCS. Causas do retardo na confirmação diagnóstica de lesões mamárias em mulheres atendidas em um centro de referência do Sistema Único de Saúde no Rio de Janeiro. Rev. Bras. Ginecol. Obstet. 2009; 31(2):75-81.^,1515. Berger F, Doussau A, Gautier C, Gros F, Asselain B, Reyal F. Impact du statut socioéconomique sur la gravité du diagnostic initial de cancer du sein. Revue d’Épidémiologie et de Santé Publique 2012; 60(1):19-29., and have been associated with delayed diagnosis of several types of cancer, including breast cancer¹⁶16. Tervonen HE, Walton R, Roder D, You H, Morrell S, Baker D, Aranda S. Socio-demographic disadvantage and distant summary stage of cancer at diagnosis – A population-based study in New South Wales. Cancer epidemiol 2016; 40:87-94.^–1919. Piñeros M, Sánchez R, Perry F, García AO, Campo R, Cendales R. Demoras en el diagnóstico y tratamiento de mujeres con cáncer de mama en Bogotá, Colombia. Salud Pública Mex 2011; 53(6):478-485.. However, Brazilian studies investigating the relation between individual characteristics and diagnosis-to-treatment interval delays in patients enrolled in cancer treatment unit programs are scarce²⁰20. Medeiros GC, Bergmann A, Aguiar SSD, Thuler LCS. Análise dos determinantes que influenciam o tempo para o início do tratamento de mulheres com câncer de mama no Brasil. Cad Saúde Pública 2015; 31(6):1269-1282..

Therefore, this study has two objectives: (1) to identify the sociodemographic profiles of women undergoing breast cancer treatment in oncology units in Belo Horizonte, Minas Gerais, and (2) to investigate potential associations between the patient profiles and intervals between diagnosis and initiation of treatment, regardless of cancer staging. The hypothesis being tested is that greater social vulnerability profiles are associated with increased time to initiate treatment.

Method

A cross-sectional study involving a population of women with confirmed diagnosis of primary breast cancer and classified as C50 (International Classification of Diseases, version 10), regardless of age. All patients lived in Belo Horizonte and had received their first treatment (surgery, chemotherapy/hormone therapy or radiation therapy) in ten municipal oncology units, from 2010 to 2013. Of the 10 units included in the study, five were devoted exclusively for SUS patients, two exclusively for patients covered by private health insurance or private patients, and three were mixed (accessible to SUS and privately insured patients).

This study was based on sociodemographic, clinical and treatment data extracted from the hospital-based cancer registry Sis-RHC/INCA (Sistema de Informação de Registros Hospitalares de Câncer)²¹21. Instituto Nacional de Câncer (INCA). Registros Hospitalares de Câncer: planejamento e gestão. 2ᵃ ed. Rio de Janeiro: INCA; 2010.. Registration of all cancer cases in HCRs (hospital-based cancer registries) has been made compulsory by the Ministry of Health and must be complied with by all high complexity oncology assistance units and centers (UNACONs and CACONs)²²22. Brasil. Ministério da Saúde (MS). Portaria nº 741, de 19 de dezembro de 2005. Regulamenta o credenciamento de centros de alta complexidade em oncologia, unidades hospitalares de radiologia, hematologia e quimioterapia. Diário Oficial da União 2005; 23 dez.. Despite issues regarding quality and completeness of records, the system is an important tool to plan actions regarding cancer surveillance, control and treatment in the country²³23. Pinto IV, Ramos DN, Costa MDCED, Ferreira CBT, Rebelo MS. Consistência e completude dos dados dos registros hospitalares de câncer no Brasil. Cad. Saúde Coletiva 2012; 20(1):113-120..

The inclusion of a postal code (ZIP code) in HCRs as of 2010 allowed the correct identification of patients’ residence and served as a reference for determination of the study period.

A total of 1,405 records of women with a primary diagnosis of cancer and undergoing initial treatment were found. Indigenous and yellow women (n = 7) were present in small numbers and were therefore excluded; other exclusion criteria included cases with ZIP codes from different municipalities (n = 30) and patients with duplicate records (n = 31). Of the remaining 1,337 records, only those with complete data regarding the four variables selected for profile definition (n = 715) were retained; records with information gaps (“no information” - digit 9 or “missing data”) were excluded. Selected variables did not differ significantly between women that remained in the study and those in the initial sample.

Breast cancer patient profiling was based on the following variables: “age” (continuous); self-reported “race/skin color” as white, black or brown; “schooling level” as < 8 years, 8 years, 11 years or ≥ 15 years; “marital status” as single, married/de facto relationship, widow or divorced; “treatment financing” -SUS or health insurance/private coverage.

Hypothesis verification was based on the dependent variable “diagnosis-to-treatment interval”, defined as the number of days between diagnosis (histopathological confirmation) and initiation of cancer treatment (≤ 60 days, 61 to 90 days or ≥ 91 days).

“Cancer staging at diagnosis” was considered a potential intervening variable with respect to “diagnosis-to-treatment interval”. Cancer staging was determined according to the TNM Classification of Malignant Tumors based on tumor size (T), presence and location of lymph nodes (N) and metastases (M)²⁴24. Instituto Nacional de Câncer (INCA). TNM: classificação de tumores malignos. 6ᵃ ed. Rio de Janeiro: INCA; 2004.. In this study, cancers were staged as “ in situ” and I, II, III or IV, according to TNM category combinations.

Categorical variables were described by absolute and relative frequencies. Mean, median and standard deviation were calculated for the variable “age”.

Profiles were delineated using cluster analysis (i.e., grouping of a set of cases/objects by similarity)²⁵25. Figueiredo Filho DB, Silva Júnior JA, Rocha EC. Classificando regimes políticos utilizando análise de conglomerados. Opinião Pública 2012; 18(1):109-128.. The sociodemographic variables describing different and internally homogeneous groups of women were included in the model, as follows: “age”, “race/skin color”, “schooling level” and “treatment financing”. The variable “marital status” had no (zero) significance for profile prediction and was therefore excluded. The two-step cluster method (SPSS^® 19.0; Statistical Package for Social Science for Windows, Inc., USA), indicated for procedures involving large databases or databases comprising continuous and categorical variables, was used. The clustering models make the following assumptions: independent variables; normally distributed continuous variables; and ordinal or multinomial categorical variables. However, the procedure is quite robust to assumption violations²⁶26. Bacher J, Knut W, Melanie V. SPSS TwoStep Cluster-a first evaluation. Berlin: Lehrstuhl für Soziologie; 2004.^,2727. Chan YH. Biostatistics 304. Cluster analysis. Singapore Med J 2005; 46(4):153-159.. The procedure is based on a series of agglomerative partitionings. First, pre-clusters corresponding to individual cases or small groups are formed; pre-clusters are then regrouped to yield final subprofiles according to an optimal number of clusters. The optimal number of clusters was determined using the Bayesian Information Criterium (BIC), as well as the log-likelihood distance measure (default program features).

Sociodemographic profiles were described and compared with respect to the variables “diagnosis-to-treatment interval” and “cancer staging at diagnosis”, via analysis of differences in proportions (Pearson's chi square test or Fisher's exact test with Bonferroni correction). The level of significance was set at 5% (p < 0.05).

Multinomial logistic regression analysis was used to verify the power of associations between patient profiles and the variable “diagnosis-to-treatment interval”, regardless of cancer staging upon diagnosis (level of significance, p< 0.05). The category time to treatment ≤60 days was used as a reference for regression analysis, since it corresponds to the maximum waiting time to initiate treatment established by the Ministry of Health¹²12. Brasil. Portaria nº 1.220, de 3 de junho de 2014. Dispõe sobre a aplicação da Lei nº 12.732, de 22 de novembro de 2012, que versa a respeito do primeiro tratamento do paciente com neoplasia maligna comprovada, no âmbito do Sistema Único de Saúde (SUS). Diário Oficial da União 2014; 3 jun..

This study is part of the research project Mulheres com câncer de mama em Belo Horizonte: perfil, trajetória e representações sobre o cuidado [Women with breast cancer in Belo Horizonte: profile, trajectory and representations about care], approved by the Research Ethics Committee of the Universidade Federal de Minas Gerais.

Results

Study participants are described in Table 1. Mean age was 57 years; women with brown skin color, less than 8 years of education and married prevailed in the sample. In almost 75% of cases, cancer treatment was financed by SUS. Cancer stage upon diagnosis corresponded to 0, I or II in 53.7% of cases. Intervals of up to 60 days between diagnosis and treatment initiation were documented in 54.3% of cases, with mean and median intervals of 67.8 and 55 days, respectively.

Seven out of 715 records were classified as outliers and excluded from the analysis. The most significant variables for patient profile prediction (two-step clustering procedure) were “race/skin color” and “schooling level”, followed by “treatment financing”; age was less important. Five different profiles were identified (Table 2).

Mean age varied little within clusters and was consistent with the age group with higher prevalence of this type of cancer (50 to 59 years)⁵5. Instituto Nacional de Câncer (INCA). Coordenação de Prevenção e Vigilância. 2016. Rio de Janeiro. [acessado 2016 Out 15]. Disponível em: http://www.inca.gov.br/estimativa/2016
http://www.inca.gov.br/estimativa/2016... . The following profiles were identified: (A) predominantly white women (68.8%) with higher levels of education (42.9%), treatment financed by health insurance/private (100%) and mean age of 56 years; (B) white women (100%) with up to 11 years of education (54.9%), treatment financed by SUS (100%) and mean age of 59 years; (C) predominantly brown women (100%), schooling level up to 11 years (53.3%), treatment predominantly financed by SUS (100%), mean age of 52 years; and (D) predominantly brown women (72.2%), < 8 years of education, treatment predominantly financed by SUS (95.3%) and mean age of 55 years. Finally, profile (E) black women, mostly with < 8 years of education (65.3%), treatment financed by SUS (100%) and mean age of 59 years (Table 2).

Most women with profile A or B (86.5% and 53.7%, respectively) were treated within ≤ 60 days of diagnosis, compared to less than half of profile C, D or E women (43.6%, 43.4% and 36.1% respectively) initiated treatment within this time frame; and 41.7% of profile E women began treatment ≥ 91 days after diagnosis (p < 0.05). Significantly higher rates of stage III and IV cancer at diagnosis were documented in profile C, D and E women (48.1%, 45.3% and 50.0% respectively), while early cancer stages (in situ and I) were more common in profile A women (44.4%). Stage II cancer was diagnosed in 39.2% of profile B women (Table 3).

Profile B, C, D and E women had higher chances of initiating treatment within 61 to 90 days or > 91 days after diagnosis compared to profile A women. Associations between profiles C, D and E and 61 to 90 days intervals persisted following adjustment for cancer staging at diagnosis. All of these profiles were associated with > 91 day intervals; chances of initiating treatment within this time frame were up to 37 times higher in profile E compared to profile A (OR: 37.26; CI95%:11.91-116.56) (Table 4).

Discussion

Cluster analysis allowed identifying five different groups of women with an apparent continuum between race/skin color and schooling level, i.e., white women with higher levels of education and black women with low levels of education at opposite ends of the spectrum. Also, treatment financed by health insurance/private care set profile A apart from profiles B to E.

Multivariate analysis results sustain the initial hypothesis that the interval between breast cancer diagnosis and initiation of treatment is longer for women with more vulnerable social characteristics, regardless of disease staging.

There is a clear consensus in the literature that the shorter the interval between diagnosis and treatment, the better the prognosis and the longer the survival of the patient. Prompt intervention is paramount for treatment efficacy in more advanced stages of the disease, or for patient comfort in cases where treatment is palliative²⁸28. Instituto Nacional de Câncer José Alencar Gomes da Silva (INCA). Controle do Câncer de Mama, Documento de Consenso. Rio de Janeiro: INCA; 2004.. Cancer staging had little impact on diagnosis-to-treatment intervals in this study. However, social characteristics associated with different profiles were so robust that interval differences persisted following adjustment for cancer staging.

Still, it should be noted that other potential confounding factors, such as health-related behaviors and obesity were not included in the analysis (missing data).

The findings of this study could be compared to those of a retrospective cohort including 137,593 women seen at SUS facilities from 2000 to 2011, and listed in Sis-RHC. In that study, diagnosis-to-treatment intervals of up to 60 days were documented in more than 50% of cases, and delays were more common among nonwhite women with less than eight years of education and (different from this study) early stage cancer²⁰20. Medeiros GC, Bergmann A, Aguiar SSD, Thuler LCS. Análise dos determinantes que influenciam o tempo para o início do tratamento de mulheres com câncer de mama no Brasil. Cad Saúde Pública 2015; 31(6):1269-1282..

Social inequalities – or the inequalities that, when associated with individual characteristics such as schooling level, income and race, among others, place some groups at disadvantage compared to others²⁹29. Barata RB. Como e por que as desigualdades sociais fazem mal à saúde. Rio de Janeiro: Editora Fiocruz; 2009. – may translate not only into worse health conditions, but also into inequalities in access to and utilization of services. The use of healthcare services is a complex resulting from interactions between several factors, including socioeconomic, demographic, cultural and psychic characteristics, health-related needs, characteristics of services and of healthcare professionals and geographic and social access availability, among others³⁰30. Travassos C, Martins M. Uma revisão sobre os conceitos de acesso e utilização de serviços de saúde A review of concepts in health services access and utilization. Cad Saúde Pública 2004; 20(Supl. 2):S190-S198.. These factors may have different impacts on access to health depending on the type of care (prevention, cure or rehabilitation), service (inpatient or outpatient) and level of complexity (primary, specialized or high complexity) involved³¹31. Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995; 36(1):1-10.. That said, differences in diagnosis-to-treatment intervals between the five profiles described in this study may reflect inequalities in utilization of highly complex oncologic services, resulting from interactions between the following conditions: lower levels of education, brown or black race/skin color and lack of access to health insurance/private appointment.

Differences in breast cancer diagnosis and treatment within the public and private healthcare systems were reported, and demonstrated advantages of the private system³²32. Liedke PE, Finkelstein DM, Szymonifka J, Barrios CH, Chavarri-Guerra Y, Bines J, et al. Outcomes of breast cancer in Brazil related to health care coverage: a retrospective cohort study. Cancer Epidemiology Biomarkers & Prevention 2014; 23(1):126-133.^–3434. Guerra MR, Silva GA, Nogueira MC, Leite ICG, Oliveira RDVCD, Cintra JRD, Bustamante-Teixeira MT. Breast cancer survival and health iniquities. Cad Saúde Pública 2015; 31(8):1673-1684.. In a multicenter study, Liedke et al. (2013) observed higher rates (p < 0.001) of advanced disease upon diagnosis and shorter survival in stage III and IV cancers (p < 0.002 and p < 0.008, respectively) in SUS patients compared to privately insured patients³²32. Liedke PE, Finkelstein DM, Szymonifka J, Barrios CH, Chavarri-Guerra Y, Bines J, et al. Outcomes of breast cancer in Brazil related to health care coverage: a retrospective cohort study. Cancer Epidemiology Biomarkers & Prevention 2014; 23(1):126-133.. Aside from expected differences between women seen by health insurance/private networks and SUS facilities, this study suggests a synergistic effect between brown or black race/skin color and low levels of education leading to poorer outcomes among women receiving care at SUS. Despite the not precise classification of the variable “race/skin color” in this study, the complex relation between race as a social construct and socioeconomic conditions has been demonstrated and is thought to be associated with poor health outcomes³⁵35. Travassos C, Williams DR. The concept and measurement of race and their relationship to public health: a review focused on Brazil and the United States. Cad Saúde Pública 2004; 20(3):660-678.^,3636. Chor D. Health inequalities in Brazil: race matters. Cad Saúde Pública 2013; 29(7):1272-1275..

Marmot (2005) suggested that ethnic inequalities in health are largely a reflex of socioeconomic factors, such income and level of education³⁷37. Marmot M. Social determinants of health inequalities. Lancet 2005; 365(9464):1099-1104.. Moreover, there is evidence to show that racial harassment and discrimination experiences, alongside the perception of living in a discriminative society, contribute to inequalities in health³⁸38. Nazroo JY, Williams DR. The social determination of ethnic/racial inequalities in health. In: Marmot M, Wilkinson RG, editors. Social determinants of health. Oxford: Oxford University Press; 2005. p. 238-266.. The issue was addressed by Travassos and Bahia (2011) in an article suggesting that affirmative policies in Brazil reinforce subgroup identities (racial, gender and others), promoting stigma and shifting the focus away from the true causes of discrimination to the institutional domain; discrimination would be “fundamentally derived from relationships between healthcare professionals and patients”³⁹39. Travassos C, Bahia L. Qual é a agenda para o combate à discriminação no SUS? Cad Saúde Pública 2011; 27(2):204-205..

Different from the situation in the United States, skin color related iniquities are not commonly investigated in health literature, in spite of expressive social inequality between black and white citizens³⁵35. Travassos C, Williams DR. The concept and measurement of race and their relationship to public health: a review focused on Brazil and the United States. Cad Saúde Pública 2004; 20(3):660-678.^,4040. Araújo EM, Costa MCN, Noronha CV, Hogan VK, Vines AI, Araújo TM. Desigualdades em saúde e raça/cor da pele: revisão da literatura do Brasil e dos Estados Unidos (1996-2005). Saúde Coletiva 2010; 7(40):116-121.. The relevance of the topic has been widely demonstrated. Discrimination in access to prenatal care and delivery based on education level and skin color, in both private and public health care services has been reported⁴¹41. Leal MC, Gama SG, Cunha CB. Desigualdades raciais, sociodemográficas e na assistência ao pré-natal e ao parto, 1999-2001. Rev Saúde Pública 2005; 39(1):100-107.. Black women are also more likely to present with advanced stages of breast cancer at diagnosis⁸8. Rezende MCR, Koch HA, Figueiredo JA, Thuler LCS. Causas do retardo na confirmação diagnóstica de lesões mamárias em mulheres atendidas em um centro de referência do Sistema Único de Saúde no Rio de Janeiro. Rev. Bras. Ginecol. Obstet. 2009; 31(2):75-81.; moreover, studies investigating diagnosis-to-treatment intervals pointed to associations between non-white skin color and treatment delay in Brazil²⁰20. Medeiros GC, Bergmann A, Aguiar SSD, Thuler LCS. Análise dos determinantes que influenciam o tempo para o início do tratamento de mulheres com câncer de mama no Brasil. Cad Saúde Pública 2015; 31(6):1269-1282..

Several authors have investigated the mechanisms through which social inequalities may affect health³⁴34. Guerra MR, Silva GA, Nogueira MC, Leite ICG, Oliveira RDVCD, Cintra JRD, Bustamante-Teixeira MT. Breast cancer survival and health iniquities. Cad Saúde Pública 2015; 31(8):1673-1684.^,3737. Marmot M. Social determinants of health inequalities. Lancet 2005; 365(9464):1099-1104.^,4242. Santos JAF. Classe Social e Desigualdade de Saúde no Brasil. Revista Brasileira de Ciências Sociais 2011; 26(75):28-55.. However, little is known about the impact of such inequalities on healthcare services utilization once access has been gained. Therefore, further investigation of the impact of social inequalities on the healthcare trajectory of breast cancer patients, particularly of those seen at the public health system, is warranted for deeper understanding of the results of this study.

This study has some limitations that should be mentioned. The use of an administrative database with a high percentage of incomplete records limits the applicability of our results to the study population, which does not reflect the total number of women undergoing breast cancer treatment in Belo Horizonte during the study period. It is worth mentioning the characteristics of patients included in the final analysis and those in the initial sample were not statistically different. However, the quality of data entry in medical registries of Sis-RHC must be improved, given the potential contribution of such databases for cancer surveillance and organization and planning of oncology services.

This is a cross-sectional study; nevertheless, the possibility of reverse causality (i.e., the disease having impacted sociodemographic profiles) is highly unlikely, since only women in the in the early stages of cancer treatment were included in the sample. Data concerning patient income were missing; therefore, “treatment financing” was used as a proxy for individual socioeconomic status, in light of the association between private healthcare services utilization, higher levels of education, formal employment and personal assets in Brazil⁴³43. Viacava F, Souza-Júnior PRB, Szwarcwald CL. Coverage of the Brazilian population 18 years and older by private health plans: an analysis of data from the World Health Survey. Cad Saúde Pública 2005; 21(Supl. 1):S119-128..

For the most part, diagnosis-to-treatment intervals in this sample fell within the timeframe set by the Brazilian Ministry of Health (up to 60 days). Still, this study revealed interval differences placing women with more vulnerable social profiles at disadvantage even after the barriers to accessing oncology treatment units are overcome.

References

Publication Dates

History