Artigos

Introduction
Ovarian cancer ranks as the seventh most common neoplasm and the eighth leading cause of cancer death among women worldwide1. In 2020, the incidence rate in high-income countries was 1.22 times higher than that estimated for middle- and low-income countries (7.1 vs. 5.8 new cases per 100,000 women)1. A similar pattern is observed in Brazil, where the high-income South region exhibits an incidence rate 1.91 times higher than the low-income North region (6.90 vs. 3.61 new cases per 100,000 women)2.
The magnitude and temporal trend of ovarian cancer incidence and mortality are influenced by the prevalence of risk and protective factors in the female population. The main factors include mutations in the Brca1 and Brca2 genes, reproductive factors (nulliparity, lactation, use of oral contraceptives, tubal ligation surgery, and oophorectomy), changes in habits and lifestyle (increased consumption of fats and sugars, low consumption of fruits and vegetables, smoking, overweight, and obesity), and occupational exposure to asbestos3-7. The impact of these factors on the carcinogenesis process depends on the histological type of the tumor, with a stronger association seen with epithelial ovarian cancer, which accounts for over 90% of ovarian cancers3-7.
Among the factors associated with ovarian cancer, reproductive factors stand out, accounting for a high population-attributable fraction3-7. In South Korea, nulliparity, never having breastfed, not undergoing tubal ligation, and never having used oral contraceptives account for 86% of the population-attributable risk4. Given that ovarian cancer lacks a screening test with high sensitivity and reliability, the factors associated with its incidence interact with inequities in access to health services (including oncological care networks and therapeutic innovations), thereby increasing the risk of death from ovarian cancer8-10.
Disparities in reproductive behaviors, access to sexual and reproductive health policies, and oncology care significantly contribute to the variations in incidence and mortality rates of ovarian cancer between low- and middle-income countries (LMICs) and high-income countries (HICs). It is important to highlight that within HICs, access to timely treatment for ovarian cancer is not uniformly distributed. Specifically, black women, immigrants, and uninsured individuals often experience less timely access to treatment. Such delays can substantially reduce their survival chances11--14.
These factors are distributed unevenly among countries due to varying levels of socioeconomic development, different intensities and speeds of the industrialization process, urbanization, and changes in women's reproductive behavior9,14,15. In high-income countries, studies have shown higher mortality rates for this neoplasm in women who experienced the two World Wars (cohorts from 1900 to 1920), followed by a reduction in women born from the 1940s onwards. These women experienced a post-war increase in fertility rates, expanded use of oral contraceptives13-20, and the effects of therapeutic innovations in chemotherapy and surgery (debulking), factors that contributed to the reduction of incidence and mortality9,14,17-24. In developing countries, an inverse profile is observed, with lower mortality rates in older generation women, followed by an increase in rates in younger cohorts20,23,25.
In this regard, a study conducted in Brazil highlighted disparities in the cohort effect on ovarian cancer mortality between the South and Northeast regions: women born in the 1950s onwards showed a reduced risk of death in the South and an increased risk in the Northeast25. It is worth noting that the South has an older age structure, better socioeconomic development, and greater access to health services and oncological care networks compare to the Northeast2,26,27. However, it's important to emphasize that there are differences among the states that make up these regions of Brazil.
Given this background, the present study aimed to analyze the age, period, and cohort (APC) effect on ovarian cancer mortality in the states of the South and Northeast regions of Brazil between 1980 and 2019. Thus, it sought to address the following research question: Are there differences in the APC effect between and within the states of the South and Northeast regions of Brazil?
The age effect represents biological changes that occur in individuals according to their life cycle15-17. At a conjunctural level, the age structure of the population influences its epidemiological profile. The period effect encompasses conjunctural changes that act across all age groups (such as health policies, therapeutic and diagnostic innovations) 15-17. The cohort effect is driven by sociocultural and political changes that impact a specific generation and bring about changes of varying magnitudes in the incidence and mortality rates across successive age groups. It also reflects the interaction between the effects of age and period as a result of accumulated exposure over a 15-17.
Methods
Study Design and location
This was an ecological study on the temporal trend of the APC effect on ovarian cancer deaths in women residing in the states of the South and Northeast regions of Brazil Between 1980 and 2019, following the recommendations of the Guidelines for Accurate and Transparent Health Estimates Reporting: The GATHER Statement28.
Brazil is politically and administratively divided into 27 federative units (26 states and the Federal District). These federative units are organized into five regions (North, Northeat, Midwest, Southeast and South).The South region consists of three states (Paraná, Rio Grande do Sul, and Santa Catarina) and the Northeast of nine (Alagoas, Bahia, Ceará, Maranhão, Paraíba, Pernambuco, Piauí, Rio Grande do Norte, and Sergipe). There are significant differences in socioeconomic, demographic and health indicators between the states of the South and Northeast, with the former showing. The former have a higher HDIM, lower vacancy rate, and greater coverage of gynecological preventive care and mammography. Furthermore, they show higher rates of mortality and incidence of cancer associated with the Westernization of habits and lifestyle (breast) and lower rates of cervical cancer (associated with infection), infant mortality and maternal mortality ratio (Supplementary Material Table 1).
Study Variables and Data Sources
Mortality data were obtained from the Mortality Information System (SIM/DATASUS) of the Ministry of Health. During the analyzed period, the ninth (ICD-9) and tenth (ICD-10) revisions of the International Statistical Classification of Diseases and Related Health Problems were in effect. Specifically, codes 183 from ICD-9 and C56 (ICD 56.0-56.9) from ICD-10 were utilized. We opted to collect data from death records of women starting from the age groups of 30 to 34 years due to the excessive number of zeros in younger age groups. Population data were sourced from the Brazilian Institute of Geography and Statistics (IBGE), utilizing the population censuses of 1980, 1991, 2000, and 2010. For inter-census years, estimates provided by the IBGE were employed. Given that this data is openly accessible and does not identify subjects, the present study was exempt from evaluation by an Ethics Research Committee.
Due to disparities in the quality and coverage of death records from SIM/DATASUS 25,29-31, we corrected the ovarian cancer death records for data quality and underreporting according to the following steps: Correction for quality of information: (i): proportional redistribution of 50% of deaths classified as ill-defined cause among defined natural causes by year, age group and state; (ii): proportional redistribution by year, age group and state of records classified as incomplete diagnosis of general cancer; (iii) the sum of the values obtained in the previous steps was added to the ovary cancer deaths registered in SIM/DATASUS; Correction for of information (ill-defined causes and incomplete cancer diagnosis) and coverage (underreporting): we multiplied the results of step iii by the correction factors obtained from adjusted synthetic extinct generations (SEG-adj) method by year, age group and state 25,29-31.
Statistical analysis
The APC effects were estimated using regression models with a Poisson distribution for the number of deaths observed in each age group i and period j (?ij). In this study, we grouped age brackets and periods into five-year intervals, resulting in I = 11 age groups, J = 8 periods, and K = 18 birth cohorts (from 1900 to 1999).
There is a linear relationship between the three time effects (age, period, and cohort). This means that we can determine any of the factors based on information from the other two, i.e. Age=Period-Cohort; Cohort=Period-Age; Period= Age+Cohort.This perfect linear relationship between the temporal factors prevents the estimation of the complete model (age-period-cohort) and is known as the non-identifiability problem11,12. Many methodologies have been proposed to correct this limitation, but there is no consensus in the literature on the best methodological strategy to use. Therefore, this study adopted the methodology most frequently recommended by authors who have established comparisons between classic statistical methods, i.e. estimable functions15,12,32-34.
The effects were additively related to the logarithm of the expected mortality rate (E(rij)), in accordance with Holford's proposal15,16. In this way, we have:
ln(E[r_ij ])=ln(?_ij/N_ij )=?+?_i+?_j+?_k+?_ij
where E[r_ij] denotes the expected rate, ?_ij represents the observed number of deaths, and N_ij is the population at risk of death at age i and period j. The parameter ? symbolizes the average effect, ?_i represents the effect of age group i, ?_ij the effect of period j, and ?_k the effect of cohort k and ?_ij is the random error at age (i) and period (j).
The APC effect was estimated using estimable functions, as proposed by Holford 11, in the Epi library (https://CRAN.R-project.org/package=Epi) of the R software (https://www.R-project.org/)32. The estimable functions are confined to the analysis of linear combinations and curvatures of the temporal terms (APC). The curvatures are estimable functions of the parameters and remain constant regardless of the parameterization used. The linear trend of the effects is divided into two components: the linear effect of age and the drift effect (linear effect of period and cohort) 14,29,30. The first drift term represents the linear trend of the logarithm of age-specific rates over time, equal to the sum of the period and cohort slopes. (?L + ?L), being ?L and ?L period and cohort linear trends, in that order. The second drift term represents longitudinal trend of age, equal to the sum of age and period slope (?L + ?L), being ?L and ?L linear trends of age and period, respectively 11,33,34.
The period 1995-1999 was chosen as a reference because it predates the implementation of the National Policy for Oncological Care in Brazil. The reference cohort was 1945-1949, as the generation from 1950 onward experienced the process of urbanization and industrialization that led to significant changes in risk factors and protective measures for ovarian cancer. Model fitting was assessed by the Akaike Information Criterion (AIC) and a significance level of 5%11,33,34. The results are presented as specific average rates by age group (age effect) and relative risk (period and cohort effect) and their respective 95% confidence intervals.
Results
States in the South showed better quality and coverage of death records compared to those in the Northeast (Table 1). We observed a progressive increase in ovarian cancer mortality rates per 100,000 women over the five-year periods in the Northeast states, with a decrease in the 2015-2019 period compared to the 2010-2014 period in the Southern states (Table 2).
In all locations, we observed a positive gradient in mortality rates starting from the age group 40 to 44 years, with the South region and its states showing coefficients of greater magnitude. Mortality rates by birth cohort and age group in the Northeast region showed a progressive increase up to the 1965 to 1969 generation (Figure 1), with similar findings in its states (data not shown). In contrast, in the South region, we observed a reduction in mortality coefficients starting from the 1935 to 1939 cohort (Figure 1), with similar results in its states (data not shown).
After estimating the APC model using estimable functions, the complete model best fit the data for the South region and its states, as well as for the Northeast region's states of Alagoas, Bahia, Maranhão, Paraíba, and Piauí. In Ceará, Pernambuco, Rio Grande do Norte, and Sergipe the best fitting model was the age-cohort model, as it presented the lowest AIC value. (Supplementary Material Table 2).
The age effect, adjusted for period and cohort, showed a progressive increase in mortality rates as age advanced in all the regions and states under study (Figure 2 and Table 3). As for the period effect, it is the reduction in the risk of death in the last two five-year periods (RR2010-2014 0.94 CI95% 0.90-0.99; RR2015-2019 0.90 CI95% 0.86-0.95) in the South region and in the last period in its states compared to the reference period (1995 to 1999): Paraná (RR=0.89 CI95% 0.82-0.97), Rio Grande do Sul (RR=0.92 CI95% 0.87-0.97) e Santa Catarina (RR=0.87 CI95% 0.78-0.98) (Figure 2 and Table 3). In the Northeast region, there was increase for all periods in comparison to the 1995 to 1999 period, in the five-year periods of the 2000s there was a progressive increase in the risk of death, ranging from 1.02 to 1.11 (2000 to 2004 vs. 2015 to 2019, p< 0.001) (Figure 3), results that are similar to those in Maranhão, Paraíba, and Piauí (Table 3). In the state of Rio Grande do Norte, there was a significant increase in the periods 1985 to 1990 (RR185-1989=1.14, p<0.001; RR1990-1994=1.05, p<0.001) and 2015 to 2019 (RR=1.25 ,p<0.001) (Table 3).
In the Northeast region, we observed a reduction in the risk of death in older cohorts compared to the reference generation (1945 to 1949), followed by an increase in younger cohorts, the risk of death varying from 0.31 (CI95% 0.28-0.34) in the 1900 to 1904 cohort to 1.54 (CI95% 1.37-1.73) in the 1985 to 1989 generation. Similar results were observed in most of its states, except in Piauí and Sergipe, where the increase in risk was not statistically significant (Figure 2 and Table 3).
In Rio Grande do Sul, there was a reduction in the risk of death in all cohorts compared to the 1945 to 1949 generation, the relative risk varied from 0.60 to 0.89 between the generations from 1900 to 1924, followed by a non-significant increase in the cohorts from 1930-1934 to 1940-1944, with a progressive reduction in the risk of death (p<0.001) in the cohorts from 1955-1959 to 1985-1989 (RR1955-1959 0.94 CI95% 0.89-0.99; RR1985-1989 0.66; CI95% 0.54-0.82). In Santa Catarina and the South region, the reduction in cohorts from 1955 to 1959 onwards was not statistically significant (RR>1, p>0.05). In Paraná, there was a significant increase in risk in the 1960 to 1969 cohort (RR1960-1964 .1.11, CI95%1.03-1.23; RR1965-19691.12; CI95%1.03-1.26); for the other generations, the increase was not statistically significant (RR>1, p>0.05). (Figure 2 and Table 3).
Discussion
We observed disparities in the quality and coverage of death records between the states of the South and Northeast of Brazil. Women residing in the Northeast have a lower likelihood of having their cause of death accurately recorded or their death reported in the mortality information system compared to those residing in the South due to inequities in access to health services experienced by the Northeastern states26,27,35,36. After adjusting the death records, we identified differences between the states in the magnitude of mortality rates and the risk of death according to the period and cohort effects.
States in the South showed higher mortality rates compared to those in the Northeast, results that were expected, as regions with a higher Human Development Index tend to exhibit higher incidence and mortality rates for cancers associated with Westernized lifestyle habits. In contrast, in areas with greater socioeconomic vulnerability, cancers associated with infections (such as cervical, stomach, liver, among others) prevail 1,37. In the South of Brazil, breast and colorectal cancers are the main types affecting women, excluding non-melanoma skin cancer, while in the Northeast, breast and cervical cancers stand out 2,38. In the states investigated here, the magnitude of the mortality rates was lower than those in high-income countries1. However, while these countries showed a declining temporal trend in rates from the late 1990s onwards 9,10,20,22, our study highlights a progressive increase in mortality rates throughout the 2000s in the South and Northeast regions of Brazil.
The incidence and mortality rates of cancer are closely associated with various sociodemographic determinants, such as migration status, race/ethnicity, marital status, and access to healthcare services1-4,11-14. Research conducted in Hong Kong involving short-stay migrants, long-stay migrants, and natives has revealed a marked increase in cancer-related mortality among short-stay migrants compared to long-stay migrants and the local Hong Kong population11. These observations highlight significant disparities in exposure to risk factors, availability of protective measures, and accessibility to preventive interventions and treatment options for cancer 11.
In relation race/ethnicity disparities, a systematic review with meta-analysis highlighted an 18% increase in the risk of mortality among Black patients compared to White patients (RR 1.18, 95% CI 1.11–1.26)39. Additionally, a 10% increase in the risk of death was demonstrated among women with unfavorable socioeconomic conditions compared to those in better conditions (RR 1.10, 95% CI 1.03–1.18), as well as a lower likelihood of receiving adequate treatment (RR = 0.85, 95% CI = 0.77–0.94)39. In relation to marital status, women living without a partner (single, widowed and divorced) had a worse prognosis when compared to married women13. Likewise, there was a higher risk of death in women without health insurance (RR 1.23 CI 1.05 to 1.44)13, there is also evidence of a lower probability of receiving adequate treatment in women treated in small hospitals compared to treatment in a large hospital (RR = 0.70, 95% CI = 0.58 to 0.85)39.
In relation racial/ethnic disparities concerning ovarian cancer, a systematic review with meta-analysis revealed an 18% increased risk of mortality among black patients compared to white patients (relative risk [RR] 1.18, 95% confidence interval [CI] 1.11–1.26) 39. Additionally, the review identified a 10% increased risk of death among women in unfavorable socioeconomic conditions compared to those in better conditions (RR 1.10, 95% CI 1.03–1.18), coupled with a lower likelihood of these women receiving adequate treatment (RR 0.85, 95% CI 0.77–0.94)39.Regarding marital status, the data showed that women without a partner (single, widowed, or divorced) had a worse prognosis compared to married women13. Moreover, there was a higher risk of death among women lacking health insurance (RR 1.23, 95% CI 1.05–1.44) 13. The review also noted a lower probability of receiving adequate treatment for women treated in smaller hospitals as opposed to those treated in larger hospitals (RR 0.70, 95% CI 0.58–0.85)39. These findings underscore the importance social determinants on health outcomes in ovarian cancer.
These factors for are not evenly distributed across ages, periods, and cohorts. Therefore, to understand the temporal behavior of this disease, it was necessary to analyze the effect of each of these temporal factors in the selected states 17-25.
The age effect displayed a similar profile across all studied locations, with a positive gradient as age progressed. Elderly women are more likely to be diagnosed with advanced-stage ovarian cancer because they access gynecological care services less frequently after menopause. Furthermore, the higher prevalence of comorbidities in these women can lead to complications, limiting treatment options 9,10,13,24.
After estimating the APC models (Age-Period-Cohort), we observed a reduction in the risk of death in the last five years (2015-2019) in the southern states and an inverse profile in Maranhão, Paraíba, and Piauí, with an increase in the risk of death in the decade de 2000. As for the cohort effect, there were also differences in the South and Northeast regions. In the southern region, the state of Rio Grande do Sul showed a decreased risk of death in younger cohorts. Meanwhile, there was an increased risk of death for women born after 1955 to 1959 in the Northeastern states.
The differences observed in the period effect may correlate with inequalities in access to health services, which are concentrated in the capitals and major metropolises of the South and Southeast of Brazil26,27,35,36,39. Thus, women residing in Northeast regions are more likely to receive timely diagnosis and treatment, increasing their survival compared to women living in the Northeast 26,27,35,37,39.
The results of the period effect observed in the Northeast are similar to those in Brazil's Central-West and northern regions25. These likely relate to the interaction between increased exposure to risk factors (changes in reproductive behavior, increased tobacco consumption, changes in dietary patterns, rising prevalence of obesity) and expanded access to health services. The former factors contributed to increased disease incidence40-44, while the latter facilitated greater access to disease diagnosis (improving death certification). However, the increase in diagnosis was not accompanied by a reduction in mortality due to the inequities in access to oncological treatment in these locations26,27,35,36,40.
The process of colonization of Brazil has generated regional inequities, with productive activities and complex economic structures centralized in the Southern and Southeastern regions26,27. These inequalities are also reflected in unequal health conditions among different populations, differences in the level of exposure to risk and protective factors, causes of illness and death, and differential access to resources and services available in the healthcare system26,27. The organization of the Unified Health System (Sistema Único de Saúde - SUS) itself reproduces these inequalities, with medium and high complexity services and equipment concentrated in the capitals and metropolitan areas, notably those located in the Center-South axis of the country. This pattern of distribution and fragmentation of healthcare provision results in geographic inequities in access26,27.
In Rio Grande do Sul, we observed a progressive reduction in risk in younger cohorts, while the opposite occurred in the Northeast and most of its states. The results from Rio Grande do Sul are similar to those presented by the United Kingdom, United States, Canada, and Western European countries17-23. At the same time, the Northeastern states show a profile similar to Spain and Greece20.
The lower risk of death from ovarian cancer in older generations in the South region, Santa Catarina, Rio Grande do Sul, and all Northeastern states is attributable to the fact that these cohorts were more exposed to the main protective factors for this neoplasm due to the high fertility rate and shorter intervals between pregnancies of women born in the first four decades of the 20th century 41-43. Moreover, this generation did not experience the nutritional transition, which, due to urbanization and industrialization, led to significant changes in dietary habits and the body composition of Brazilian women 8,43,46.
The reduction in the risk of death starting from the 1955 to 1959 cohort in the South region, Rio Grande do Sul, and Santa Catarina (non-significant decrease) may be correlated with the high prevalence of oral contraceptive use42-44 combined with increased access to general health services and the oncological care network 47,48. This promotes early diagnosis, timely treatment, and reduced risk of death, even in women more exposed to the risk factors of this neoplasm36,40,46,47. Protective factors to which women residing in the Northeast were not exposed. Changes in their reproductive behavior and Westernization of habits and lifestyle41,47-49 were not accompanied by improvements in access to sexual and reproductive health and oncological care, contributing to the increased likelihood of death in women born from the 1960s onwards36,40,46,47.
Changes in reproductive behavior are challenging to reverse, as they are the result of industrialization, urbanization, and sociocultural shifts that began to challenge the role of women in patriarchal society36,40. Therefore, prevention measures should be directed at modifiable risk factors such as smoking, diet, overweight, and obesity48,49,50-54. It's essential to facilitate access to natural foods by reducing prices and promoting family farming, state-financed urban gardens, and controlling the advertising of ultra-processed foods52-54. This should be coupled with expanded access to health services and diagnostic tests, prioritizing women at higher risk for disease development and areas with significant health inequities 2,25.
It is essential to highlight the two main limitations of the present study. The first concerns the problem of non-identifiability of the complete model due to the linearity between the three temporal factors, which allows the obtaining of countless solutions for maximum likelihood models, with different estimates for their parameters, providing the exact prediction for any combination of the effects, but making it impossible to estimate the complete model32-34. However, in the present work, we estimate the APC models using the estimable functions proposed by Holford (1983)33,34, and there is consensus in the literature that among the models estimated by classical statistics, the most appropriate method for correcting this counterpoint is the estimable functions32-34. The second limitation concerns the quality of SIM/DATASUS death records regarding the quality and underreporting of SIM/DATASUS death records29-31. These limitations may interfere with the magnitude and temporal trend of ovarian cancer mortality. In this study, we sought to correct this counterpoint through the proportional redistribution of ill-defined causes, incomplete cancer diagnosis, and correcting underreporting through appropriate demographic techniques29-31.
Despite these limitations, this research's main contribution is that it analyzed the temporal trend in ovarian cancer mortality with records corrected for quality and coverage in Brazilian regions with different fertility patterns, access to health services, and levels of socioeconomic development. Represents a contribution to the Strategic Action Plan for Coping with Chronic Noncommunicable Diseases in Brazil (2022-2030), specifically in the mortality surveillance component, providing evidence for the planning and evaluating health policies for the disease as mentioned earlier in these places.
Conclusion
The analysis of temporal effects according to APC using estimable functions revealed differences between the states of the South and Northeast of Brazil. Despite the higher mortality in the southern states, these showed a decreased risk of death in the last five-year period and a decreased risk in the younger cohorts (Rio Grande do Sul). There was an increased risk in the last five-year period and the younger cohorts in the Northeast. These results correlate with the different paces of demographic and epidemiological transition experienced by these two regions, intrinsically linked to socioeconomic conditions and access to healthcare services.
Financing
The study was partially funded by the Brazilian Coordination for the Improvement of Higher Education Personnel (Capes) – Funding Code 001. Karina Cardoso thanks the National Council for Scientific and Technological Development (CNPq) for the productivity grant (306652/2022-6).
Data Availability
The databases used in this study, as well as the R code for the APC effect analyses performed using the Epi library of the software, are available in the Zenodo repository. Meira K. Ovarian cancer mortality in the states of Northeast and South Brazil (1980-2019): effect of age-period and cohort. Zenodo; 2024 (https://zenodo.org/records/13307646)
References
Sung H, Ferlay J, Siegel RL, Laversanne M, Jemal A, Bray F. Global Cancer Statistics. 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209-249.
Brasil, Ministério da Saúde, Instituto Nacional de Câncer José Alencar Gomes da Silva. Estimativa 2020: Incidência de Câncer no Brasil. Rio de Janeiro, RJ: INCA; 2022.
Reid BM, Permuth JB, Sellers TA. Epidemiology of ovarian cancer: a review. Cancer Biol Med. 2017;14:9-32.
Park B, Park S, Shin HR, Shin A, Yeo Y, Choi JY, Jung KW, Kim BG, Kim YM, Noh DY, Ahn SH, Kim JW, Kang S, Kim JH, Kim TJ, Kang D, Yoo KY, Park SK. Population attributable risks of modifiable reproductive factors for breasts and ovarian cancers in Korea. BMC Cancer. 2016;16:5.
Penny M. Ovarian Cancer: An Overview. Radiol Technol. 2020;91:561-575.
Huang X, Wang X, Shang J, Lin Y, Yang Y, Song Y, Yu S. Association between dietary fiber intake and risk of ovarian cancer: a meta-analysis of observational studies. J Int Med Res. 2018;46:3995-4005
Avgerinos KI, Spyrou N, Mantzoros CS, Dalamaga M. Obesity and cancer risk: Emerging biological mechanisms and perspectives. Metabolism. 2019;92:121-135.
Stewart C, Ralyea C, Lockwood S. Ovarian Cancer: An Integrated Review. Semin Oncol Nurs. 2019;35:151-156.
Sopik V, Iqbal J, Rosen B, Narod SA. Why have ovarian cancer mortality rates declined? Part II. Case-fatality. Gynecol Oncol. 2015;138:750-756.
Sopik V, Iqbal J, Rosen B, Narod SA. Why have ovarian cancer mortality rates declined? Part III. Prospects for the future. Gynecol Oncol. 2015;138:757-761
Zhao Y, Zhuang Z, Yang L, He D. Age-period-cohort analysis and projection of cancer mortality in Hong Kong, 1998-2030. BMJ Open. 2023 Oct 11;13(10):e072751.
Mei S, Chelmow D, Gecsi K, Barkley J, Barrows E, Brooks R, Huber-Keener K, Jeudy M, O'Hara JS, Burke W. Health Disparities in Ovarian Cancer: Report From the Ovarian Cancer Evidence Review Conference. Obstet Gynecol. 2023 Jul 1;142(1):196-210.
Gardner AB, Sanders BE, Mann AK, Liao CI, Eskander RN, Kapp DS, Chan JK. Relationship status and other demographic influences on survival in patients with ovarian cancer. Int J Gynecol Cancer. 2020 Dec;30(12):1922-1927.
Villanueva C, Chang J, Ziogas A, Bristow RE, Vieira VM. Ovarian cancer in California: Guideline adherence, survival, and the impact of geographic location, 1996-2014. Cancer Epidemiol. 2020 Dec;69:101825.
Holford TR. The estimation of age, period and cohort effects for vital rates. Biometrics. 1983;39:311-324
Holford TR. Understanding the effects of age, period, and cohort on incidence and mortality rates. Annu Rev Public Health. 1991;12:425-457
Silva IdS, Swerdlow AJ. Recent trends in incidence of and mortality from breast, ovarian and endometrial cancers in England and Wales and their relation to changing fertility and oral contraceptive use. Br J Cancer. 1995;72:485-492.
Zhang J, Ugnat AM, Clarke K, Mao Y. Ovarian cancer histology-specific incidence trends in Canada 1969-1993: age-period-cohort analyses. Br J Cancer. 1999;81:152-158
Wang B, Liu SZ, Zhang RS, Chen WQ, Sun XB. Time trends of ovarian cancer in China. Asian Pac J Cancer Prev. 2014;15:191-1933.
González-Diego P, López-Abente G, Pollán M, Ruiz M. Time trends in ovarian cancer mortality in Europe (1955–1993). Effect of age, birth cohort and period of death. Eur J Cancer. 2000;36:1816-1824.
Tamakoshi K, Kondo T, Yatsuya H, Hori Y, Kikkawa F, Toyoshima H. Trends in the mortality (1950-1997) and incidence (1975-1993) of malignant ovarian neoplasm among Japanese women: analyses by age, time, and birth cohort. Gynecol Oncol. 2001;83:64-71
Hirabayashi Y, Marugame T. Comparison of time trends in ovary cancer mortality (1990-2006) in the world, from the WHO Mortality Database. Jpn J Clin Oncol. 2009;39:860-861
Zhang Y, Luo G, Li M, Guo P, Xiao Y, Ji H, Hao Y. Global patterns and trends in ovarian cancer incidence: age, period, and birth cohort analysis. BMC Cancer. 2019;22:984
Cabasag CJ, Arnold M, Butler J, Inoue M, Trabert B, Webb PM, Bray F, Soerjomataram I. The influence of birth cohort and calendar period on global trends in ovarian cancer incidence. Int J Cancer. 2020;146:749-758.
Meira KC, dos Santos J, da Silva CMFP, Ferreira AA, Guimarães RM, Simões TC. Effects of age-period and cohort on mortality due to ovarian cancer in Brazil and its regions. Cad Saúde Pública. 2019;35:e00087018.
Travassos C, de Oliveira EXG, Viacava F. Desigualdades geográficas e sociais no acesso aos serviços de saúde no Brasil: 1998 e 2003. Ciência Saúde Coletiva. 2006;11:975-986.
Barreto ML. Desigualdades em Saúde: Uma perspectiva global. Ciência Saúde Coletiva. 2017;22:2097-2108.
Stevens GA, et al. Guidelines for Accurate and Transparent Health Estimates Reporting: the GATHER statement. Lancet. 2016;388:e19-23.
Meira KC, dos Santos GWS, dos Santos J, Guimarães RM, Bezerra de Souza DL, Cani Ribeiro GP, Dantas ESO, Leite de Carvalho JB, Jomar RT. Analysis of the effects of the age-period-birth cohort on cervical cancer mortality in the Brazilian Northeast. PLOS ONE. 2020;15:e0226258.
Meira KC, Magnago C, Mendonça AB, Duarte SFS, Freitas PHO, Santos J, Bezerra de Souza DL, Simões TC. Inequalities in Temporal Effects on Cervical Cancer Mortality in States in Different Geographic Regions of Brazil: An Ecological Study. Int J Environ Res Public Health. 2022;19:5591.
Hill K, You D, Choi Y. Death Distribution Methods for Estimating Adult Mortality: sensitivity analysis with simulated data errors Demographic Research, 2009. 21: 235:25.
Carstensen B. Age-period-cohort models for the Lexis diagram. Statist Med. 2006;26:3018-3045.
Holford TR. Approaches to fitting age-period-cohort models with unequal intervals. Stat Med. 2006;25:977-993
Robertson C, Boyle P. Age period-cohort analysis of chronic disease rates I: Modelling approach. Stat Med. 1998;17:1305-1323
Oliveira EXG, Melo ECP, Pinheiro RS, Noronha CP, Carvalho MS. Access to cancer care: Mapping hospital admissions and high-complexity outpatient care flows. The case of breast cancer. Cad Saúde Pública. 2011;27:317-326.
Simões TC, Meira KC, dos Santos J, Câmara DCPC. Prevalence of chronic diseases and access to health services in Brazil: evidence of three household surveys. Cien Saude Colet. 2021;26:3991-4006.
Bray F, Jemal A, Grey N, Ferlay J, Forman D. Global cancer transitions according to the Human Development Index (2008-2030): a population-based study. Lancet Oncol. 2012;13:790-801
Guimarães RM, Muzi CD, Teixeira MP, Pinheiro SS. Cancer’s transition in Brazil and strategical decision-making in women's public health policies. R Pol Públ. 2016;20:33-50.
Karanth S, Fowler ME, Mao X, Wilson LE, Huang B, Pisu M, Potosky A, Tucker T, Akinyemiju T. Race, Socioeconomic Status, and Health-Care Access Disparities in Ovarian Cancer Treatment and Mortality: Systematic Review and Meta-Analysis. JNCI Cancer Spectr. 2019 Oct 9;3(4):pkz084.
Viacava F, Bellido JG. Health, access to services, and sources of payment, according to household surveys. Cien Saúde Colet. 2016;21:351-370.
Alves JE, Cavenaghi S. Transições urbanas e da fecundidade e mudanças dos arranjos familiares no Brasil. Cadernos de Estudos Sociais. 2012;27.
Martine G. Brazil's fertility decline 1965-95: a fresh look at key factors. Popul Dev Rev. 1996;22:47-75.
Perpétuo IH, Wong LLR. Desigualdade socioeconômica na utilização de métodos anticoncepcionais no Brasil: uma análise comparativa com base nas PNDS 1996 e 2006. In: Pesquisa Nacional de Demografia e Saúde da Criança e da Mulher - PNDS 2006: dimensões do processo reprodutivo e da saúde da criança/Ministério da Saúde, Centro Brasileiro de Análise e Planejamento. Brasília: Ministério da Saúde, 2009.
Farias MR, Leite SN, Tavares NU, Oliveira MA, Arrais PS, Bertoldi AD, Pizzol TD, Luiza VL, Ramos LR, Mengue SS. Use of and access to oral and injectable contraceptives in Brazil. Rev Saude Publica. 2016 Dec;50(suppl 2):14s.
Vasconcelos AMN, Gomes MMF. Transição demográfica: a experiência brasileira. Epidemiol Serv Saúde. 2012;21:539-548.
Poirier AE, Ruan Y, Hebert LA, Grevers X, Walter SD, Villeneuve PJ, Brenner DR, Friedenreich CM; ComPARe Study Team. Estimates of the current and future burden of cancer attributable to low fruit and vegetable consumption in Canada. Prev Med. 2019 May;122:20-30
Brasil, Ministério da Saúde, Secretaria de Vigilância em Saúde, Departamento de Análise em Saúde e Vigilância de Doenças Não Transmissíveis. Vigitel Brasil 2006-2021: vigilância de fatores de risco e proteção para doenças crônicas por inquérito telefônico: estimativas sobre frequência e distribuição sociodemográfica do estado nutricional e consumo alimentar nas capitais dos 26 estados brasileiros e no Distrito Federal entre 2006 e 2021: estado nutricional e consumo alimentar. Brasília: Ministério da Saúde; 2022.
Kodaira K, Abe FC, Galvão TF, Silva MT. Time-trend in excess weight in Brazilian adults: A systematic review and meta-analysis. PLoS ONE. 2020;16:e0257755.
Ferreira APS, Szwarcwald CL, Damacena GN, Souza Júnior PRB. Increasing trends in obesity prevalence from 2013 to 2019 and associated factors in Brazil. Rev Bras Epidemiol. 2021 Dec 10;24(suppl 2):e210009.
Brasil, Ministério da Saúde, Instituto Nacional de Câncer José Alencar Gomes da Silva. A situação do câncer de mama no Brasil: síntese de dados dos sistemas de informação. Instituto Nacional de Câncer José Alencar Gomes da Silva. Rio de Janeiro: INCA, 2019.
Albuquerque MV, Viana ALD, Lima LD, Ferreira MP, Fusaro ER, Iozzi FL. Regional health inequalities: changes observed in Brazil from 2000-2016. Ciência & Saúde Coletiva. 2017;22:1055-1064.
Baker P, Machado P, Santos T, Sievert K, Backholer K, Hadjikakou M, Russell C, Huse O, Bell C, Scrinis G, Worsley A, Friel S, Lawrence M. Ultra-processed foods and the nutrition transition: Global, regional and national trends, food systems transformations and political economy drivers. Obes Rev. 2020 Dec;21(12):e13126.
Burlandy L. Construction of the food and nutrition security policy in Brazil: strategies and challenges in the promotion of intersectorality at the federal government level. Cien Saude Colet. 2009;14:851-860.
Popkin BM, Reardon T. Obesity and the food system transformation in Latin America. Obesity Reviews. 2018;19:1028-1064.