Establishing a cancer cohort in Latin America: The Mexican Teachers Cohort

doi:10.21203/rs.3.rs-4927140/v1

Download PDF

Research Article

Establishing a cancer cohort in Latin America: The Mexican Teachers Cohort

https://doi.org/10.21203/rs.3.rs-4927140/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Purpose

In Latin America, prospective cancer research is restricted by limited access to cancer outcome data. In this study, we demonstrate a strategy for identifying incident breast cancer cases within an established prospective cohort, the Mexican Teachers’ Cohort (MTC), through the use of self-reports, electronic health records (EHRs), administrative databases, and registries.

Methods

The MTC consists of 115,275 female teachers aged 25 years and older who completed a questionnaire on reproductive history, lifestyle, and health between 2006 and 2008. We used self-reported breast cancer, mortality and cancer registries, and EHRs and administrative databases for case ascertainment. Person-time was calculated from the date of baseline questionnaire response to diagnosis, death, or December 31, 2019. We estimated age-specific and age-standardized incidence rates for breast cancer.

Results

We identified 1,313 incident breast cancers; 94% were confirmed through registries and/or databases, and 6% directly by contacting participants. The crude breast cancer incidence was 102 per 100,000 person-years, with the highest incidence in women aged 65–69 (185 per 100,000 person-years). The age-standardized incidence was 77 per 100,000. Over one-third of cases (n = 476) occurred in women younger than 50 years.

Conclusion

The efficient identification of participants with incident breast cancer in the MTC demonstrates that prospective cancer cohorts can be successfully established in Latin America.

Cancer cohort

Latin America

observational

breast cancer

prospective

Most well-established cancer cohorts include individuals from high-income countries in North America and Europe.[1] Furthermore, within USA-based cohorts, the inclusion of Hispanics, the fastest-growing demographic group in the USA, is low.[2] Low population diversity in existing cancer cohorts affects our understanding of exogenous and endogenous exposures in relation to cancer. While data pooling efforts of cohorts in Latin America will likely advance cancer research in the region,[3] the low coverage of population-based cancer registries (PBCRs) remains an important challenge.[4] However, a recent effort in Brazil exemplified the usefulness of combining different data sources to ascertain cancer outcomes.[5] We aimed to illustrate an efficient strategy that relies on self-reports and large medical and administrative databases to identify incident breast cancer cases in a large prospective cohort in Mexico.

The Mexican Teachers’ Cohort (MTC)

It was established between 2006 and 2008 when 115,275 women agreed to participate in the study by providing information on their reproductive history, lifestyle, and health outcomes.[6] Participants were public education teachers (aged 25 and older) working in a twelve-state area in Mexico. The response for the first 2011–2014 follow-up cycle was 83%. For 2015–2021, which was affected by the COVID-19 pandemic, the response was 63%. A new follow-up cycle is underway. For this analysis, we excluded 730 women who reported breast cancer at baseline or with a breast cancer diagnosis prior to the date of completion of the baseline questionnaire. The National Institute of Public Health’s IRB approved the study (#CI-1645).

Self-reported breast cancer

At baseline and follow-up cycles, participants were asked, using paper and electronic questionnaires and phone interviews, if they had ever been diagnosed with breast cancer, the year of diagnosis, and whether they received treatment. Women were contacted via a mailed questionnaire or a telephone call to confirm the diagnosis (by providing treatment details and pathology report and/or hospitalization discharge note when available) and consent for medical record review and access to tumor tissue.

Mortality and cancer registries for breast cancer identification

As previously described[6], we used two national death certificate-based mortality registries to identify and confirm breast cancer cases using the International Classification of Diseases 10th Revision (ICD-10) C50 code in any of the six causes of death. Deaths reported by the employer or next-of-kin were crosslinked to registry records between January 1, 2006, and December 31, 2021, using a robust linkage algorithm developed for Mexico.[7] We also crosslinked all cohort participants known to be alive. Most women who died of breast cancer deaths (81.9%) had previously reported their diagnosis. Study participants were also crosslinked to a network of PBCRs on 13,517 persons with cancer between 2017 and 2020.[8]

Electronic health records and administrative databases for breast cancer identification

Table 1 describes all data sources used to identify breast cancer cases. Close to 80% (n = 91,007) of MTC participants had medical insurance and health coverage by one institution (ISSSTE; Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado or Institute of Social Security and Services for Civil Servants). We obtained data from two of ISSSTE’s health information systems: a medical and financial transaction data system and a drug supply system. From the first, we used six datasets: medical visits, hospitalization discharge, in-hospital procedures, and mammography results datasets and chemotherapy and radiation oncology registries. From the second, we used an outpatient prescription database Our structured approach to ascertaining cancer cases using large-scale electronic health data led to the development of an algorithm employing various identification strategies..[9] We assessed the performance of our algorithm with a random sample of 100 women with breast cancer and 100 women without breast cancer who were patients at an ISSSTE tertiary care center in Mexico City. The positive predictive value of the algorithm was 97% (95% CI 77, 100).[10]

We obtained access to a temporary medical leave registry (2007–2019) for the second-largest health provider for MTC participants (IMSS; Instituto Mexicano del Seguro Social, or Mexican Institute for Social Security), which covers 11.9% of the MTC (n = 13,696). The registry included ICD-10 codes and dates for the medical leave. We identified all observations with a C50 code. We also obtained electronic health record (EHR) information from two local institutions that provide care to MTC participants: ISSSTECali (Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado de Baja California) n = 2,066 or 1.8% of the MTC; and ISSTECh (Instituto de Seguridad Social de los Trabajadores del Estado de Chiapas) n = 3,753 or 3.3%. Staff personnel from ISSSTECali identified breast cancer cases from electronic medical records (2006–2021). For women insured by ISSTECh, we obtained information from an oncology treatment registry (2006–2020).

Finally, while all MTC participants have social security, some women may have chosen to receive care from a non-social security hospital. Between 2007 and 2019, breast cancer treatment in Mexico for people not covered by social security health insurance schemes was financed through Seguro Popular, a voluntary health insurance scheme. We used yearly reimbursement data for 2007 and 2016 (268,703 breast cancer reimbursement claims).

Breast cancer incidence

We considered a breast cancer case confirmed when the case was identified in the mortality or cancer registries or any of the data sources presented in Table 1. For cases identified exclusively through self-reports in follow-up questionnaires, we required additional clinical information from the participants and/or pathology reports to confirm the case. We calculated person-time from the date of completion of the baseline questionnaire to the date of diagnosis, death, or December 31, 2019, whichever occurred first. We estimated women contributed person-time to different age groups for age-specific incidence as follow-up elapsed. We estimated age-standardized breast cancer incidence using the direct method with the world population as the standard population.[11]

After a median follow-up of 11.2 years (1,287,406 person-years), we confirmed 1,313 women with breast cancer in the MTC between 2006 and 2019. There were 709 self-reports of breast cancer. Among these, 2.8% (n = 20) were reclassified as being cancer other than breast, and 9.6% (n = 68) were benign breast lesions. There were 157 women with suspected breast cancer whose diagnosis has not been confirmed to date. These women were not counted as cases in our estimates.

Most participants with confirmed breast cancer were found in databases and/or registries (94.0%; n = 1,235). The remaining 6.0% (n = 78) were confirmed by a mailed questionnaire or a telephone call. Among them, IMSS provided treatment for close to one-third (n = 22) of these cases; 15.4% (n = 12) reported receiving care at ISSSTE, which were not found using our breast cancer identification algorithm; the remaining 56.4% of cases were treated by private (n = 23) or public local providers (n = 21) with whom it was not possible to access databases. The crude incidence of breast cancer was 102 per 100,000 person-years (Table 2). Incidence increased steadily with age. The highest incidence was observed in women aged 65–69 (185 per 100,000 person-years). The age-standardized incidence rate was 77 per 100,000. We identified 476 women (36.3% of confirmed cases) with early-onset breast cancer (i.e., < 50 years).

In a large prospective cohort of women living in Mexico, we effectively identified participants with incident breast cancer leveraging EHRs and administrative data. Our age-standardized breast cancer incidence estimate of 77 per 100,000 is almost identical to Globocan’s (Global Cancer Observatory) estimate for Mexican women 25 years and older (78 per 100,000).[12] The corresponding estimate from a PBCR in Yucatan in southern Mexico was much lower (49 per 100,000).[13] Differences with this estimate may be due to the inclusion of women younger than 25 years from a state with a lower breast cancer burden relative to other regions in Mexico.[14]

In Latin America, breast cancer is diagnosed at earlier ages relative to high-income countries. [15] Breast cancer in young women may be etiologically distinct from later-onset disease, and there is an increasing concern about our limited understanding of early-onset breast cancer. In the MTC, almost 40% of breast cancer cases occurred in women younger than 50 years of age. Thus, the MTC could provide novel insights into early-onset breast cancer.

Our results show that even when PBCR coverage is limited, cancer outcome assessment can be performed effectively using EHRs and administrative data. Some breast cancers in MTC participants may have been missed due to coding errors, missing data, or breast cancer ascertainment algorithms. Also, cancer care could have been provided in facilities not captured in our data sources for some participants, underscoring the importance of active follow-up. This misclassification in outcome ascertainment will likely result in conservative estimates with sufficient precision for most risk factor analyses to be informative. As the MTC matures, increasing the number and quality of data sources while strengthening follow-up based on self-reported questionnaires should be emphasized.

Prospective cancer cohorts can be successfully implemented in limited resource settings. The MTC is now the largest cancer cohort in Latin America and is ideally positioned to contribute to bridging the knowledge gaps in cancer research and addressing population disparities. Allocating resources to cancer research infrastructure in diverse populations may result in impactful advancement of cancer research.

Funding

This work was supported by the American Institute for Cancer Research (Grant number 635420).

Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Author contributions

Conceptualization; Liliana Gómez-Flores-Ramos, Mauricio Hernandez-Avila, Isabelle Romieu, Ruy López-Ridaura, Martín Lajous. Data curation; Marion Brochier, Adrian Cortes, Hugo Sánchez-Blas, Alberto Castellanos. Formal analysis; Liliana Gómez-Flores-Ramos, Jocelyn Jaen, Martín Lajous. Funding acquisition; Martín Lajous, Isabelle Romieu, Ruy López-Ridaura. Investigation; Liliana Gómez-Flores-Ramos, Marion Brochier, Dalia Stern, Ruy López-Ridaura, Andrés Catzin-Kuhlmann and Martín Lajous. Methodology; Liliana Gómez-Flores-Ramos, Marion Brochier, Adrian Cortes, Dalia Stern, Antonio García-Anaya, Jocelyn Jaen, Mauricio Hernández-Avila, Juan Eugenio Hernández-Ávila, Angélica Pon, Olga Ontiveros, Pabel Miranda, Patricia Pérez, Alberto Castellanos, Mildred Chávez, Alejandro Mohar, Martín Lajous. Project administration; Marion Brochier. Validation; Liliana Gómez-Flores-Ramos, Marion Brochier, Adrian Cortes, Pabel Miranda, Alberto Castellanos, Martín Lajous. Visualization; Liliana Gómez-Flores-Ramos, Marion Brochier. Writing - original draft; Liliana Gómez-Flores-Ramos, Marion Brochier, Martín Lajous. Writing - review & editing; Liliana Gómez-Flores-Ramos, Marion Brochier, Alejandro Mohar, Martín Lajous. All authors reviewed and approved the final version of the article.

Data availability

The datasets analyzed during the current study are not publicly available due to data sharing policies by participating healthcare organizations and deidentification concerns but are available from the corresponding author on reasonable request.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Research, research Ethics and Biosecurity Committees at the National Institute of Public Health of Mexico (#CI-1645).

Consent to participate

Participants received the baseline questionnaire with a letter of invitation detailing study objectives. Response to the questionnaire was considered consent.

Acknowledgments: We thank the teachers participating in the study. Professor Víctor Sastré and our liaisons for state Ministries of Public Education have provided significant support with their effort and time. ISSSTE, IMSS, ISSSTECali, and ISSSTECH have provided necessary institutional support. Finally, we thank Lourdes Santamaría, Laura Quezada, Rocío Santoyo, Nohemí Figueroa, Maribel Colorado, Hilda Vázquez, Esperanza Piña, Roberto García, Enrique Alvarado, Raúl Torres, and Adriana Rodríguez who have worked with passion and commitment on the project for over 17 years. This work is dedicated to the loving memory of Dr. Aura Erazo Valle-Solís, a breast cancer research and care leader at ISSSTE. This work was supported by the American Institute for Cancer Research (Grant number 635420).

NCI Cohort Consortium National Cancer Institute
Elena JW, Gallicchio L, Pottinger CA et al (2023) Cancer Epidemiology in Hispanic Populations: An Analysis of Funded Observational Research at the National Cancer Institute. Cancer epidemiology, biomarkers & prevention: a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 32:653–658
Carrillo-Larco RM, Di Cesare M, Hambleton IR et al (2020) Cohort Profile: The Cohorts Consortium of Latin America and the Caribbean (CC-LAC). Int J Epidemiol 49:1437–g
Piñeros M, Abriata MG, de Vries E et al (2021) Progress, challenges and ways forward supporting cancer surveillance in Latin America. Int J Cancer 149:12–20
Moreira AD, Teles de Menezes S, Júnior MM, da Silva Vicente JT, Camelo LV, Barreto SM (2023) Investigation of cancer incidence in ELSA-Brasil. Cancer Epidemiol 87:102467
Lajous M, Ortiz-Panozo E, Monge A et al (2017) Cohort Profile: The Mexican Teachers' Cohort (MTC). Int J Epidemiol 46:e10
Quezada-Sánchez AD, Espín-Arellano I, Morales-Carmona E et al (2022) Implementation and validation of a probabilistic linkage method for population databases without identification variables. Heliyon 8:e12311
Mohar A, Paredes A, Brau-Figueroa H et al (2023) Implementation of a population-based cancer registry network in Mexico 2017–2020. 66:104–112 Salud Pública de México
Earles A, Liu L, Bustamante R et al (2018) Structured Approach for Evaluating Strategies for Cancer Ascertainment Using Large-Scale Electronic Health Record Data. JCO Clin cancer Inf 2:1–12
Castellanos-Medina J, Miranda-Aguirre P, González Rodríguez K et al (2023) Uso de registros médicos electrónicos para estimar la prevalencia de cáncer de mama en el ISSSTE: Resultados iniciales. Congreso de Investigación en Salud Pública. Cuernavaca, Mexico
Nations U (2011) World Population Prospects: The 2010 Revision, Volume I: Comprehensive Tables. ST/ESA/SER. .A/313 ed
(2024) Global Cancer Observatory: Cancer Today. Lyon, France: International Agency for Research on Cancer
Leal YA, Torres J, Gamboa R et al (2022) Cancer Incidence in Merida, Mexico 2015–2018: First Report from the Population-based Cancer Registry. Arch Med Res 53:859–866
(2019) Global Burden of Disease Study 2019 (GBD 2019). Global Burden of Disease Collaborative Network
Zahed H, Feng X, Sheikh M et al (2024) Age at diagnosis for lung, colon, breast and prostate cancers: An international comparative study. Int J Cancer 154:28–40

Table 1

Database sources of breast cancer cases identification and confirmation
	Database source	Description	Period	Date of diagnosis	MTC participants covered
Registries	Cancer registry	Population-based local cancer registries	2017–2020	Pathology report	12.2% of participants. Covers cities in seven states with MTC participants.
Registries	Mortality registry	National mortality registry	2006–2021	Midpoint between the date of the last available questionnaire response and the date of death.	100%
Electronic health records and administrative databases	IMSS	Temporary medical leave registry	2007–2019	Date of first disability claim	11.9%
	ISSSTE	Medical and financial transaction data registry and an oncologic drug supply registry	2006–2019	Date of first ICD-10 C50 code	78.9%
	Seguro Popular	Reimbursement database for cancer diagnosis and treatment of uninsured individuals	2007–2016	Date of first breast cancer claim	Women who sought treatment in facilities outside social security institutions. Uncertain coverage.
Electronic medical records	ISSSTECali	Electronic medical records	2006–2021	Pathology report	1.8%
Electronic medical records	ISSTECh	Electronic medical records	2006–2020	Pathology report	3.3%

*Multiple cases have been confirmed by more than one source, and the more precise date is considered for person-year estimations. IMSS: Instituto Mexicano del Seguro Social, or Mexican Institute for Social Security; ISSSTE: Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado or Institute of Social Security and Services for Civil Servants; ISSSTECali: Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado de Baja California; ISSTECh: Instituto de Seguridad Social de los Trabajadores del Estado de Chiapas.

Table 2

Age-specific and overall breast cancer incidence in 1,313 women from the Mexican Teachers Cohort, 2006–2019
Age group	Cases	Person-years	Crude incidence	Age-standardized incidence
25–34 years	6	58,596	10.2
35–39 years	38	126,487	30.0
40–44 years	121	194,657	62.2
45–49 years	311	291,822	106.6
50–54 years	390	314,694	123.9
55–59 years	273	193,330	141.2
60–64 years	119	71,986	165.3
65–69 years	45	24,291	185.3
70–74 years	8	7,504	106.6
≥ 75 years	2	2,038	98.1
Overall	1,313	1,285,405	102.1	77.0

No competing interests reported.

Download PDF

Editorial decision: Revision requested
20 Sep, 2024
Reviews received at journal
19 Sep, 2024
Reviewers agreed at journal
11 Sep, 2024
Reviews received at journal
09 Sep, 2024
Reviewers agreed at journal
02 Sep, 2024
Reviewers agreed at journal
29 Aug, 2024
Reviewers agreed at journal
28 Aug, 2024
Reviewers invited by journal
28 Aug, 2024
Editor assigned by journal
17 Aug, 2024
Submission checks completed at journal
17 Aug, 2024
First submitted to journal
16 Aug, 2024

You are reading this latest preprint version

Establishing a cancer cohort in Latin America: The Mexican Teachers Cohort

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Introduction

Materials and Methods

The Mexican Teachers’ Cohort (MTC)

Self-reported breast cancer

Mortality and cancer registries for breast cancer identification

Electronic health records and administrative databases for breast cancer identification

Breast cancer incidence

Results

Conclusion

Declarations

References

Tables

Additional Declarations

Status:

Version 1