Abstract:

The COVID-19 pandemic has significantly impacted healthcare systems worldwide, especially on the management of chronic diseases such as cancer. This study explores the effects of COVID-19 on cancer mortality trends in Brazil, Chile, and Peru. The monthly age-standardized mortality rates in different places of death (hospital/clinic or home) were estimated using vital statistics and death certificate databases. An interrupted time series analysis was performed for each country, using the date of lockdown implementation as the intervention point. Overall cancer mortality rates reduced after the implementation of pandemic restrictions, with a significant decrease in Brazil. In total, 75.3%, 55.4%, and 45.7% of deaths in Brazil, Peru, and Chile, respectively, occurred in hospitals. After lockdowns were implemented, at-home deaths increased in all countries, and in-hospital deaths correspondingly decreased only in Chile. Our results suggest that COVID-19 has significantly affected rates of cancer mortality and place of death in Latin America.

Keywords:
Neoplasms; COVID-19; Cause of Death

Resumo:

A pandemia de COVID-19 impactou significativamente os sistemas de saúde ao redor do mundo, especialmente no manejo de doenças crônicas, como o câncer. Este estudo explora os efeitos da COVID-19 nas tendências de mortalidade por câncer no Brasil, Chile e Peru. As taxas de mortalidade mensais padronizadas por idade em diferentes locais de morte (hospital/clínica ou domicílio) foram estimadas usando estatísticas vitais e bancos de dados de atestados de óbito. Uma análise de série temporal interrompida foi realizada para cada país, tendo como ponto de intervenção a data de implementação do lockdown. As taxas gerais de mortalidade por câncer reduziram após a implementação das restrições, com uma queda significativa no Brasil. No total, 75,3%, 55,4% e 45,7% dos óbitos no Brasil, Peru e Chile, respectivamente, ocorreram em hospitais. Depois da implementação dos lockdowns, as mortes em domicílio aumentaram em todos os países, e as mortes hospitalares diminuíram de forma correspondente apenas no Chile. Nossos resultados sugerem que a COVID-19 afetou significativamente as taxas de mortalidade por câncer e o local de morte na América Latina.

Palavras-chave:
Neoplasias; COVID-19; Causas de Morte

Resumen:

La pandemia de COVID-19 impactó significativamente los sistemas de salud de todo el mundo, sobre todo en el manejo de enfermedades crónicas, como el cáncer. Este estudio explora los efectos de la COVID-19 en las tendencias de mortalidad por cáncer en Brasil, Chile y Perú. Las tasas de mortalidad mensuales estandarizadas por edad en diferentes locales de fallecimiento (hospital/clínica o domicilio) se estimaron utilizando estadísticas vitales y base de datos de certificados de defunción. Se realizó un análisis de serie temporal interrumpida para cada país, teniendo como punto de intervención la fecha de implementación del lockdown. Las tasas generales de mortalidad por cáncer redujeron tras la implementación de las restricciones, con una disminución significativa en Brasil. En total, el 75,3% de los óbitos ocurrieron en hospitales en Brasil, el 55,4% en Perú y el 45,7% en Chile. Tras la implementación del lockdown, las muertes domiciliarias aumentaron en todos los países, y las muertes hospitalarias solo redujeron de forma correspondiente en Chile. Nuestros resultados sugieren que la COVID-19 afectó significativamente las tasas de mortalidad por cáncer y el local del fallecimiento en América Latina.

Palabras-clave:
Neoplasias; COVID-19; Causas de Muerte

Introduction

The COVID-19 pandemic, caused by SARS-CoV-2, has affected health systems worldwide, prompting questions about the impact of mobility restriction measures and hospital overcrowding on chronic disease management, including cancer. After noticing the swift spread of different variants and the high mortality rates in countries affected early by the virus ¹1. Saglietto A, D'Ascenzo F, Zoccai GB, De Ferrari GM. COVID-19 in Europe: the Italian lesson. Lancet 2020; 395:1110-1.^,22. Abdullatif M, Rahim F. Cross-country comparison of case fatality rates of COVID-19/SARS-CoV-2. Osong Public Health Res Perspect 2020; 11:74-80., many countries quickly took action to prevent similar scenarios from happening in their own territories. Governments worldwide enforced mobility restrictions, implemented in significantly different ways depending on the country ³3. Observatorio para la Contención de COVID-19 en América Latina, Universidad de Miami. Índice de adopción de políticas públicas. http://observcovid.miami.edu/americas/?lang=es (accessed on 24/Jan/2023).
http://observcovid.miami.edu/americas/?l... .

Latin America was one of the most affected regions in the world with highly heterogeneous COVID-19 responses and fatality counts ⁴4. Schaal A, Ardavin JA. COVID-19 in Latin America and the Caribbean: an overview of government responses to the crisis. Updated 11 Novermber 2020. https://www.oecd.org/coronavirus/policy-responses/covid-19-in-latin-america-and-the-caribbean-an-overview-of-government-responses-to-the-crisis-0a2dee41/ (accessed on 24/Jan/2023).
https://www.oecd.org/coronavirus/policy-... . But, in addition to COVID-19 mortality, the impact of the pandemic can be seen in other disease indicators ⁵5. Roth GA, Vaduganathan M, Mensah GA. Impact of the COVID-19 pandemic on cardiovascular health in 2020: JACC state-of-the-art review. J Am Coll Cardiol 2022; 80:631-40.^,66. Fedeli U, Casotto V, Schievano E, Bonora E, Zoppini G. Diabetes as a cause of death across different COVID-19 epidemic waves. Diabetes Res Clin Pract 2022; 190:109984.. For example, in cases where COVID-19 has been described as a more significant risk factor for death, such as immunity-debilitating diseases or cancer treatment and therapy.

The question about the intersection of COVID-19 and cancer has been raised several times during the pandemic. Even though extensive research has already been produced ⁷7. Cuadrado C, Vidal F, Pacheco J, Flores-Alvarado S. Cancer care access in Chile's vulnerable populations during the COVID-19 pandemic. Am J Public Health 2022; 112(S6):S591-601.^,88. Maringe C, Spicer J, Morris M, Purushotham A, Nolte E, Sullivan R, et al. The impact of the COVID-19 pandemic on cancer deaths due to delays in diagnosis in England, UK: a national, population-based, modelling study. Lancet Oncol 2020; 21:1023-34.^,99. Han X, Hu X, Zhao J, Jemal A, Yabroff KR. Identification of deaths caused by cancer and COVID-19 in the US during March to December 2020. JAMA Oncol 2022; 104:19-21.^,1010. Rottoli M, Gori A, Pellino G, Flacco ME, Martellucci C, Spinelli A, et al. Colorectal cancer stage at diagnosis before vs during the COVID-19 pandemic in Italy. JAMA Netw Open 2022; 5:e2243119.^,1111. Fernandes GA, Feriani D, França e Silva ILA, Mendonça e Silva DR, Arantes PE, Canteras JS, et al. Differences in mortality of cancer patients with COVID-19 in a Brazilian cancer center. Semin Oncol 2021; 48:171-80.^,1212. Jazieh AR, Akbulut H, Curigliano G, Rogado A, Alsharm AA, Razis ED, et al. Impact of the COVID-19 pandemic on cancer care: a global collaborative study. JCO Glob Oncol 2020; (6):1428-38., evidence for this topic is still scarce in Latin America. Studies such as the one by Vásquez Rosas et al. ¹³13. Vázquez Rosas T, Cazap E, Delgado L, Ismael J, Bejarano S, Castro C, et al. Social distancing and economic crisis during COVID-19 pandemic reduced cancer control in Latin America and will result in increased late-stage diagnoses and expense. JCO Glob Oncol 2021; (7):694-703. reported a decrease in cancer screening and delays in cancer treatments in several Latin American countries. This is further supported by Cuadrado et al. ⁷7. Cuadrado C, Vidal F, Pacheco J, Flores-Alvarado S. Cancer care access in Chile's vulnerable populations during the COVID-19 pandemic. Am J Public Health 2022; 112(S6):S591-601., who also reported this situation in Chile. Additionally, in a Colombian cohort of oncologic patients, a higher mortality risk was observed among cancer patients with an active disease ¹⁴14. Ospina AV, Bruges R, Mantilla W, Triana I, Ramos P, Aruachan S, et al. Impact of COVID-19 infection on patients with cancer: experience in a Latin American country: the ACHOCC-19 Study. Oncologist 2021; 26:e1761-73..

The pandemic also influenced the general population’s place of death. Changes in healthcare-seeking behaviors, due to fear or to the unavailability of services may have affected where people died. Edwards & Wohl ¹⁵15. Edwards JK, Wohl DA. Dying at home due to coronavirus disease 2019. Open Forum Infect Dis 2022; 9:ofac439. compared the probability of dying at home before (2014-2019) and after the pandemic (2020-2021) in North Carolina (United States). They pointed out a 23% increase in all-cause deaths at home and also observed that this risk varied depending on race/ethnicity: by the end of 2020, at-home deaths had increased by 65% for Hispanics.

The extent to which the place of death for people with cancer varied among Latin American countries and how these deaths were affected by the COVID-19 pandemic is unknown. Data collection efforts in this region have yielded widely available and good quality vital statistics, especially data on deaths, which creates an opportunity to explore the effects of the COVID-19 pandemic on different aspects of death in several countries. In this study, we aimed to evaluate cancer mortality trends during the COVID-19 pandemic in Brazil, Chile, and Peru, comparing at-home and in-hospital deaths.

Methods

We identified the number of deaths from July 2019 to July 2021 using vital statistics and death certificate databases from Brazil (Mortality Information System - SIM, acronym in Portuguese), Chile (Department of Statistics and Health Information - DEIS, acronym in Spanish), and Peru (National Information System of Deaths - SINADEF, acronym in Spanish). Our analysis included people with malignant neoplasms as underlying causes of death. We calculated the monthly age-standardized mortality rates (ASMR) in each country and reported places of death (hospital/clinic or home). Since these were no major differences between sexes, we chose to report overall mortality in our study. However, the Supplementary Material 1 (https://cadernos.ensp.fiocruz.br/static//arquivo/suppl-1-e00057423_5787.pdf) shows the analyses by sex.

We conducted an interrupted time series for each country using the date of implementation of lockdowns or when strict mobility measures took place (March 2020 for the three countries) as the intervention point. We then fitted a linear model, considering a variable for the time of the interventions, a dummy variable for pre- and post-intervention periods, and an interaction term between them. To assess the changes according to the place of death, we included a variable indicating whether deaths happened at home or in hospitals and included interaction terms of this variable with the time of the interventions and with the indicator of the pre- and post-intervention periods. Confidence intervals were obtained parametrically with the maximum likelihood estimation of the standard error of the coefficients from the regression.

As sensitivity analyses, we first tested a lag effect, changing our intervention date from March 2020 to April 2020 to explore if the effect is only observed at the intervention or if it is delayed in time by an induction period. Then, we used a lead control, changing the intervention date to December 2019, to verify that the effect started at the intervention and not before.

Results

During the observed period, 3,753,804 deaths with reported place of death occurred. Of these, 13.1% and 21.8% deaths were due to cancer and COVID-19, respectively. Table 1 shows that, from 2020 to 2021, 30,128 (1%) and 2,425 (0.5%) people in Brazil and Peru, respectively, had both cancer and COVID-19 on their death certificates. This information was not available for Chile. In all countries, most recorded deaths were of men and adults over 40 years of age. The highest proportion of cancer deaths was observed in Chile (20.2%). Additionally, 32.8% of Peru’s deaths in 2020-2021 were caused by COVID-19, while in Brazil and Chile 20.2% and 20.6% of deaths were caused by the virus, respectively. In total, 75.3%, 55.4%, and 45.7% of deaths in Brazil, Peru, and Chile, respectively, happened in hospitals. In all countries, men had a higher proportion of deaths (55.7%, 53.6%, and 58.3% in Brazil, Peru, and Chile, respectively). In Peru, 61.1% of people who died in hospitals were men, whereas the percentages were 53% and 55.9% in Brazil and Chile, respectively. Death certificates that reported both COVID-19 and cancer as causes of death were most frequent for in-hospital deaths, both in Brazil (1.3% were reported as in-hospital deaths and 0.1% as at-home deaths) and Peru (1% referred to in-hospital deaths and 0.1% to at-home).

Figure 1 shows the interrupted time series for the overall cancer mortality. After pandemic restrictions were implemented, the overall cancer ASMR was immediately reduced, with -14.9 (95% confidence interval - 95%CI: -25.2; -4.8) deaths and -5.3 (95%CI: -23.4; 12.8) deaths per million people in Brazil and Chile, respectively. In both countries, especially Chile, trends before the pandemic pointed toward a reduction in cancer mortality. In Peru, however, the trend for cancer mortality was increasing during the pre-intervention period, with 9.3 (95%CI: -5.5; 24.1) more deaths per million people - and this may have occurred due to strategies that were implemented to improve death registration (see Discussion). Therefore, our immediate change estimate must be interpreted cautiously, as it is probably misestimating the real change.

Figure 2 shows the interrupted time series for cancer mortality by place of death. We observed three different pre-intervention scenarios: in Brazil, in-hospital cancer mortality was higher, while in Chile, at-home cancer deaths prevailed. Before the pandemic, Peru’s death registrations seem to have been distributed evenly between hospitals and homes, but as aforementioned, the high slope could reflect changes in the completeness of death registrations.

Figure 2 also shows that after the implementation of mobility restrictions to prevent COVID-19, including lockdowns, a similar pattern occurred in all countries: at-home cancer deaths increased, and cancer deaths in hospitals decreased. Brazil, Chile, and Peru had 17.9 (95%CI: 11.5; 24.3), 34.7 (95%CI: 19.2; 50.2), and 63.3 (95%CI: 37.8; 88.9) more at-home deaths per million people, respectively. Conversely, hospital mortality rates decreased by -33.1 (95%CI: -46.0; -20.2), -38.0 (95%CI: -49.5; -26.4), and -55.3 (95%CI: -75.7; -34.9) deaths per million people.

In Brazil, there were -50.9 (95%CI: -64.9; -37.0) less in-hospital cancer deaths per million people than at-home cancer deaths. This difference was -72.7 (95%CI: -91.4; -53.9) and -118.6 (95%CI: -150.0; -86.8) less deaths per million people in hospitals compared to homes, for Chile and Peru respectively.

The sensitivity analyses we conducted supported our findings. When the date of the intervention was moved forward, the effect we found was sustained, as would be expected in a dynamic situation such as a pandemic. Moving the intervention date backwards did not change mortality rates by place of death. For Peru, however, the lead control analyses showed an increase in mortality even before the pandemic, supporting our cautious approach when interpreting immediate changes in overall mortality. These findings, available and briefly discussed in the Supplementary Material 2 (https://cadernos.ensp.fiocruz.br/static//arquivo/suppl-2-e00057423_1368.pdf), concur with the policies in force in Peru to improve death registration.

Discussion

Our results highlight the drastic changes that occurred in 2020 and 2021 due to the COVID-19 pandemic, impacting death rates, particularly cancer deaths, in Brazil, Peru and Chile. In the three countries, a distinct shift in the reported place of cancer deaths emerged during this period, being characterized by an increase in deaths at home and a simultaneous decrease in deaths in hospitals. Chile presented the most balanced shift, where the rise in at-home fatalities was almost equal to the decrease of in-hospital deaths.

This paper raises important questions about the impact of the pandemic on cancer mortality and on the healthcare systems of Latin American countries. Further research is needed to investigate the reasons for the shift in the reported at-home and in-hospital deaths, and to better understand the effects of the pandemic on cancer mortality. It is also important to unravel the potential implications of this shift regarding access to health care and health outcomes for different subpopulations.

We propose two main possible explanations for the observed changes in cancer mortality during the COVID-19 pandemic. First, public health measures such as mobility restrictions changed access and care-seeking behaviors, delaying hospital presentation and changing preferences about the place of death for cancer patients. Secondly, COVID-19 deaths may have presented as a competing event for cancer death, which could be the specific explanation for the unbalanced shift of cancer deaths observed in Brazil.

COVID-19 as a competing event, in turn, can be due to COVID-19 causing death per se in cancer patients and reducing cancer mortality (factual decrease). Alternatively, because even in cases where the person died of cancer complications, if they had a positive COVID-19 test, and some of the surveillance guidelines would identify them as COVID-19 deaths, they would not count as cancer deaths in official statistics (registration decrease). The extent of the latter situation is unknown and will vary by country orders and pandemic control regulations in place.

Our research was limited by data availability. Since the death registration database does not contain clinical information, we could not ascertain the exact cancer stage patients were experiencing when they died and whether they were undergoing treatments. Furthermore, we could not assess individual preferences regarding place of death, which may have influenced the different mortality rates at home and in hospitals. For example, López-Valcárcel et al. ¹⁶16. López-Valcárcel BG, Pinilla J, Barber P. Dying at home for terminal cancer patients: differences by level of education and municipality of residence in Spain. Gac Sanit 2019; 33:568-74. reported that European patients with cancer preferred to die at home. We do not know if this is also the case in Latin America. Notably, our data showed higher at-home cancer mortality in Chile before the pandemic, which aligns with the country’s policy of universal palliative care coverage for cancer patients, implemented in 2005 ¹⁷17. Pastrana T, Eisenchlas J, Centeno C, De Lima L. Status of palliative care in Latin America: looking through the Latin America Atlas of Palliative Care. Curr Opin Support Palliat Care 2013; 7:411-6.^,1818. Bastías G, Valdivia G. Reforma de salud en Chile; el plan AUGE o régimen de Garantías Explícitas en Salud (GES). Su origen y evolución. Bol Esc Med 2007; 32:50-5..

Our study may also have measurement errors, particularly in our evaluation of data from Peru, due to the country’s historically low death registration coverage ¹⁹19. Mathers CD, Fat DM, Inoue M, Rao C, Lopez AD. Counting the dead and what they died from: an assessment of the global status of cause of death data. Bull World Health Organ 2005; 83:171-7.. However, due to Peru’s recent significant improvement in registration processes ²⁰20. Vargas-Herrera J, Miranda Monzón J, Lopez Wong L, Miki Ohno J. La cobertura de muertes con certificación médica en el Perú, 2012-2019. An Fac Med 2022; 83:123-9., our results may have overestimated the changes in mortality rates after the implementation of mobility restrictions in the country, possibly due to the underestimation of mortality rates before these measures. This was supported by the lead control, which showed an increasing mortality when we moved the intervention date to December 2019. The efforts to improve death registration in Peru, through the use of an electronic reporting web platform and many other tools, was recently challenged by failures in the security features, which allowed for data entry of people who had not died ²¹21. Dominguez Loeda G. The enigma of being dead in Peru. La Prensa Latina 2022; 19 oct. https://www.laprensalatina.com/the-enigma-of-being-dead-in-peru/.
https://www.laprensalatina.com/the-enigm... . The SINADEF is currently under review, and its future is uncertain ²²22. Ministerio de Salud. Información de fallecidos del Sistema Informático Nacional de Defunciones - SINADEF. https://www.datosabiertos.gob.pe/dataset/información-de-fallecidos-del-sistema-informático-nacional-de-defunciones-sinadef-ministerio (accessed on 20/Dec/2022).
https://www.datosabiertos.gob.pe/dataset... ^,2323. Vargas-Herrera J, López Reyes L. ¿Quo vadis SINADEF? Antes, durante y después de la pandemia por COVID-19. An Fac Med 2022; 83:267-70..

This study strengths include the use of high-quality whole population data from Chile and Brazil, and moderate-quality data from Peru ²⁴24. Seitz K, Deliens L, Cohen J, Cardozo EA, Tripodoro VA, Iwamoto Marcucci FC, et al. Feasibility of using death certificates for studying place of death in Latin America. Rev Panam Salud Pública 2021; 45:e149.. The information about place of death presented in death certificates is complete and accurate, providing high-quality data for our analyses ²⁴24. Seitz K, Deliens L, Cohen J, Cardozo EA, Tripodoro VA, Iwamoto Marcucci FC, et al. Feasibility of using death certificates for studying place of death in Latin America. Rev Panam Salud Pública 2021; 45:e149.^,2525. Mikkelsen L, Phillips DE, Abouzahr C, Setel PW, De Savigny D, Lozano R, et al. A global assessment of civil registration and vital statistics systems: monitoring data quality and progress. Lancet 2015; 386:1395-406.. Furthermore, the coverage of death registrations in Brazil and Chile is close to 100% ²⁶26. World Health Organization. Civil registration coverage of cause-of-death (%). https://www.who.int/data/gho/data/indicators/indicator-details/GHO/civil-registration-coverage-of-cause-of-death-(-) (accessed on 28/Jun/2023).
https://www.who.int/data/gho/data/indica... . Also, over 95% of completeness of cause of death in both countries ²⁷27. World Health Organization. Completeness of cause-of-death data (%). https://www.who.int/data/gho/data/indicators/indicator-details/GHO/completeness-of-cause-of-death-data-(-) (accessed on 28/Jun/2023).
https://www.who.int/data/gho/data/indica... contribute to the robustness of our findings. Although the effects of COVID-19 on various systems may have delayed death registration processes, leading to the underestimation of mortality at the onset of the pandemic, we did not anticipate the backlog to change our results, as this problem was likely resolved in subsequent periods.

Our analysis considered the underlying cause of death per official statistics. We found that less than 1% of death reports from Brazil and Peru mentioned both COVID-19 and cancer as causes of fatality. These deaths, mostly registered as in hospitals, support our competing event hypothesis. However, it is uncertain if these are factual or due to registration discrepancies. Since the “any mention” approach could cause researchers to overcount deaths, thereby altering the total number of deaths ²⁸28. Bishop K, Balogun S, Eynstone-Hinkins J, Moran L, Martin M, Banks E, et al. Analysis of multiple causes of death: a review of methods and practices. medRxiv 2022; 3 aug. https://www.medrxiv.org/content/10.1101/2022.08.01.22278086v1.
https://www.medrxiv.org/content/10.1101/... , we plan to conduct further research using alternate methods to explore this intersection.

Furthermore, our study underscores the crucial role of accurate information on causes of death, which is not only important in the current crisis, but also in the management of future emergencies, as it aids the effective tracking, surveillance, and long-term impact monitoring of different fatality causes.

Our study is the first to address this question in Latin America, allowing for a unique comparison of three countries with different health care systems and approaches to the COVID-19 pandemic, quantifying a foreseeable consequence of the pandemic, and assessing an unaddressed part of the intersection between COVID-19 and cancer. We hope this work helps public health practitioners, end-of-life care professionals and palliative care experts prepare for future emergencies. Understanding these findings can help adapt and allocate resources to meet the evolving needs of health care even beyond the pandemic.

Acknowledgments

D. Durán received financial support for her doctoral studies as a Kuok Fellow from the Rossy Cancer Network (McGill University, Canada; 2022). D. Durán and R. Calderon-Anyosa received doctoral training fellowships from the Quebec Research Fund (Fonds de Recherche du Québec - Santé, Canada; FRQ-S 2022-2023).

References

Publication Dates

History