The Negative Impact of the New Coronavirus Pandemic on the Trend of Breast Biopsies and Their Direct Costs: Interrupted Time Series Analysis

Abstract

Objectives

The overload of health services because of the COVID-19 pandemic has reduced the population’s access to treatment and prevention of other diseases. This study aimed to identify whether there was a change in the trend of breast biopsies and their direct costs during the COVID-19 pandemic in a developing country’s public and universal healthcare system.

Methods

This was an ecological time series study of mammograms and breast biopsies in women aged 30 years or older from an open-access data set of the Public Health System of Brazil from 2017 to July 2021.

Results

In 2020, there was a reduction of −40.9% in mammograms and −7.9% in breast biopsies compared with the prepandemic period. From 2017 to 2020, there was an increase in the breast biopsies ratio per mammogram (1.37%-2.55%), the percentage of Breast Imaging-Reporting and Data System (BI-RADS) IV and V mammograms (0.79%-1.14%), and the annual direct cost of breast biopsies (Brazilian Real 3 477 410.00 to Brazilian Real 7 334 910.00). In the time series, the negative impact of the pandemic was lower on BI-RADS IV to V mammograms than on BI-RADS 0 to III. There was an association between the trend of breast biopsies and BI-RADS IV to V mammography.

Conclusions

The COVID-19 pandemic harmed the increasing trend of breast biopsies, their total direct costs, BI-RADS 0 to III and IV to V mammograms, observed in the prepandemic period. Furthermore, there was a tendency to screen women at a higher risk for breast cancer during the pandemic.

Introduction

Breast cancer represents 11.7% of all malignant neoplasms worldwide, with an incidence of 2.3 million cases annually.¹ In Brazil, the estimated annual incidence of malignant neoplasms is 625 000 cases, and breast cancer is one of the most frequent (66 000 new cases per year and 29.7% of malignant neoplasms in women).² About 15.8% of total cancer spending in Brazil is for breast cancer.³ Furthermore, the cost of treating advanced breast cancer is essential and can be 3 or 4 times higher than when the disease is diagnosed early.^{4 , 5} Therefore, delay in diagnosis may increase the incidence of advanced disease and the direct and indirect costs of breast cancer treatment.

Brazil’s public healthcare system is universal and financed exclusively by the government. Over 70% of the Brazilian population depends on the country’s public healthcare system to prevent and treat diseases.^{6 , 7} According to the recommendation of the Brazilian Ministry of Health, the target population for breast cancer screening is women aged 50 to 69 years, with biennial bilateral mammography. Yet, women under the age of 50 or 70 years or older also undergo mammography according to their risk for the disease after medical evaluation.⁸ The radiological classification of mammograms Breast Imaging-Reporting and Data System (BI-RADS) defines the indication for breast biopsy. In other words, if the mammogram is BI-RADS IV (suspicion of malignancy) or BI-RADS V (high suspicion of malignancy), there is an indication of breast biopsy to confirm or rule out a cancer diagnosis.⁸

The World Health Organization declared the novel coronavirus (COVID-19) a pandemic in March 2020.^{9 , 10} The exponential increase in people infected with COVID-19 had overwhelmed healthcare systems in many countries.^{9 , 10} Consequently, there was a harmful interference in the population’s access to treatment and prevention of other diseases, including breast cancer.¹¹ To date, there is a lack of studies evaluating the trend of breast biopsies in the early diagnosis of breast cancer since the beginning of the COVID-19 pandemic in a developing country’s public and universal healthcare systems. This information is essential to support planning actions to tackle breast cancer efficiently when healthcare systems are overloaded by an emerging disease (for example, a pandemic). This study aimed to identify whether there was a change in the trend of breast biopsies and their direct costs since the beginning of the COVID-19 pandemic in a developing country’s public and universal healthcare systems.

Methods

This was an ecological time series study using open data from the Brazilian Information System Department of the Public Health System (DATASUS [in Portuguese]) from January 2017 to July 2021.¹² The study area is Brazil, the fifth most populous country in the world, with an estimated population of 213.3 million people in 2021.¹³ A breast biopsy is relatively common in the age group of fewer than 50 years or 70 years or older; therefore, mammograms and breast biopsies in women aged 30 years or older were included in this study.¹⁴

The sociodemographic data of the Brazilian population are from the DATASUS website, obtained from 2017 to 2021, available from projection studies of the last census carried out in the country (in 2010).¹² Bilateral screening mammograms (performed every 2 years) and diagnostic mammograms (performed at 6 and 12 months if the radiological classification is BI-RADS III) during the period of this study are from the Outpatient Information System (SIASUS [in Portuguese]), a subsystem of DATASUS.¹² The radiological classification of mammograms (BI-RADS) is from the Cancer Information System (SISCAN [in Portuguese]), another subsystem of DATASUS.¹² Finally, breast biopsies (nodule excision, needle aspiration punctures, or thick needle breast puncture) and their direct costs in local currency (Brazilian Real [BRL]) during the period of this study also are from SIASUS.¹² All tests and medical procedures performed in the public Health System in Brazil (including mammograms and breast biopsies) are organized in preestablished codes and values throughout the national territory. These values are readjusted only when authorized by the government and do not suffer direct variation with inflation or exchange policies.^{6 , 7 , 12}

For the trend analysis of the dependent variables (breast biopsies, their direct costs, and the BI-RADS radiological classification) in the total and interrupted time series, we used the Joinpoint Regression Program, version 4.9.0.0 of 2021.15, 16, 17, 18 This program tests the different segments of the time series (using the Poisson Regression Model), defining those that best represent the trend of the dependent variable over time. Once this model is determined, the program calculates the monthly percent change (MPC) or the annual percent change (APC) and the statistical significance of each segment and the total time series. Therefore, the trend increases if the MPC or APC is a positive value and significantly different from zero (considering the 95% confidence intervals and 5% significance level). In contrast, if the value is negative with P < .05, the trend decreases. Finally, if there is no statistically significant difference, the trend of the variable is stationary.15, 16, 17, 18 In this study, we parameterized for analysis up to 3 inflection points (or up to 4 segments) and preferentially used the month as the independent variable to reach a more significant number of points observed along the time series to improve the accuracy of the segments defined by the Joinpoint Regression Program.15, 16, 17, 18

In this study, the prepandemic period was from January 2017 to February 2020, and the pandemic period was from March 2020 (the month of the declaration of community transmission of COVID-19 throughout the national territory) to July 2021. We obtained data from October 30, 2021, to November 30, 2021. We used Microsoft Excel® (Microsoft, Redmond, WA) for Office 365 MSO to tabulate sociodemographic data by descriptive statistics (mean, SD, median, and percentages) and the Minitab® (Minitab Inc., Pensilvânia, EUA) 19.2020.1 to build time series plots. This study exclusively used a public database and was conducted by the relevant research guidelines/regulations of research. The Ethics Committee of the Health Sciences Sector of the Federal University of Paraná approved this research (CAAE 51438521.2.0000.0102).

Results

From 2017 to 2021, there was an increase in the number of Brazilian women from 105.2 to 108.2 million, with an increasing trend (APC 0.7%; P < .001). Considering the target population of breast cancer screening (50 to 69 years old), there was an increase from 19.5 to 21.6 million women in the same period, with an increasing trend (APC 2.5%; P < .001).

The average annual mammograms performed on women aged 30 years or older in the prepandemic period (2017 to 2019) was 4.3 million. In 2020 (the beginning of the COVID-19 pandemic), there were 2.5 million mammograms performed. In other words, in 2020, there was a 40.9% reduction in mammograms compared with the annual average of tests carried out in the prepandemic period. In contrast, there was an increase in the percentage of BI-RADS IV to V mammograms, from 0.79% in 2017 to 1.14% in 2020, and a reduction in BI-RADS 0 to III mammograms, from 99.21% to 98.86% in 2020. The average annual number of breast biopsies in the prepandemic period was 70705. In 2020, there were 65 529 breast biopsies performed. In other words, in 2020, there was a 7.9% reduction in breast biopsies compared with the average annual number of prepandemic procedures. The breast biopsies ratio per mammogram increased from 1.37% in 2017 to 2.55% in 2020. The yearly direct cost of prepandemic mammograms reduced from BRL 195 948 550.00 (2017) to BRL 185 149 022.00 (2019) and was BRL 110 448 627.00 in 2020. In contrast, the annual direct cost of prepandemic breast biopsies increased from BRL 3 477 410.00 (2017) to BRL 8 615 490.00 (2019) and was BRL 7 334 910.00 in 2020.

Table 1 and Figure 1 show that the trend of mammograms with radiological classification BI-RADS 0 to III and IV to V was stationary in the total time series. In the interrupted time series, the increasing trend of BI-RADS 0 to III mammograms in the prepandemic months (MPC 0.8%; P < .001) was lower than in BI-RADS mammograms IV to V (MPC 1.5%; P < .001). At the beginning of the pandemic, the decreasing trend of BI-RADS 0 to III mammograms (MPC −43.7%; P = .028; duration 3 months) was greater than in BI-RADS IV to V mammograms (MPC −22.7%; P = .009; duration 4 months). The recovery period in the number of monthly mammograms (increasing trend) was greater in BI-RADS 0 to III (MPC 29.6%; P < .001; duration 6 months) than in BI-RADS IV to V (MPC 17.9%; P = .005; duration 5 months). Subsequently, the trend was stationary in the BI-RADS 0 to III and BI-RADS IV to V mammograms until July 2021.

Table 1.

Trend of mammograms radiological classification (BI-RADS), breast biopsies performed, and their direct costs from January 2017 to July 2021 of Brazil’s public and universal healthcare system.

Interrupted time series		Mammograms BI-RADS 0-III	Mammograms BI-RADS IV-V	Breast biopsies	Direct costs of breast biopsies
Segment 1	Lower endpoint	January 2017	January 2017	January 2017	January 2017
	Upper endpoint	February 2020	January 2020	February 2020	December 2019
	MPC	0.8∗	1.5∗	1.1∗	3.6∗
	95% CI	1.2-4.6	1.2-1.9	0.9-1.4	2.3-5.0
	P value	< .001	< .001	< .001	< .001
	Trend	Increasing	Increasing	Increasing	Increasing
Segment 2	Lower endpoint	March 2020	February 2020	March 2020	January 2020
	Upper endpoint	May 2020	May 2020	May 2020	March 2020
	MPC	−43.7∗	−22.7∗	−23.0∗	−37.5
	95% CI	−66.2 to −6.3	−36.0 to −6.7	−38.7 to −3.3	−91.0 to 332.6
	P value	.028	.009	.025	.627
	Trend	Decreasing	Decreasing	Decreasing	Stationary
Segment 3	Lower endpoint	June 2020	June 2020	May 2020	April 2020
	Upper endpoint	November 2020	October 2020	September 2020	August 2020
	MPC	29.6∗	17.9∗	17.0∗	24.2
	95% CI	16.6-44.0	5.3-32.0	30.9-2.8	−11.0 to 73.2
	P value	< .001	.005	.007	.197
	Trend	Increasing	Increasing	Increasing	Stationary
Segment 4	Lower endpoint	December 2020	November 2020	October 2020	September 2020
	Upper endpoint	July 2021	July 2021	July 2021	July 2021
	MPC	−3.7	−0.4	0.3	0.6
	95% CI	−7.2 to 0.1	−3.1 to 2.3	−1.1 to 1.8	−6.0 to 7.6
	P value	.055	.746	.643	.86
	Trend	Stationary	Stationary	Stationary	Stationary
Total Time Series	MPC	−0.3	0.3	0.4∗	1.4∗
	95% CI	−0.8 to 0.1	−0.0 to 0.7	0.1-0.6	0.6-2.2
	P value	.171	.069	< .004	.001
	Trend	Stationary	Stationary	Increasing	Increasing

Note. March 2020 indicates the COVID-19 pandemic declaration month.

BI-RADS indicates Breast Imaging-Reporting and Data System; CI, confidence interval; MPC, monthly percentage change.

^∗MPC is significantly different from zero (P < .05), with an increasing trend (if positive) or decreasing trend (if negative).

Figure 1.

The monthly trend of radiological classification of mammograms BI-RADS 0 to III and BI-RADS IV to V (A) and the monthly trend of breast biopsies and their direct costs (B) from January 2017 to July 2021 of Brazil’s public and universal healthcare system.

BI-RADS, Breast Imaging-Reporting and Data System.

In the total time series, there was an increasing trend of breast biopsies (MPC 0.4%; P < .004) lower than their direct costs (MPC 1.4%; P = .001), as shown in Table 1 and Figure 1. In the interrupted time series, the increasing trend of breast biopsies in the prepandemic months (MPC 1.1%; P < .001) was lower than the growth in their direct costs (MPC 3.6%; P < .001). At the beginning of the pandemic, there was a decreasing trend of breast biopsies (MPC −23.0%: P = .025; duration 3 months), followed by 5 months with an increasing trend (MPC 17.0%; P = .007), and another period of the stationary trend until July 2021, similar to that observed in BI-RADS IV to V mammograms. The direct costs of breast biopsies showed a stationary trend from the pandemic’s beginning until July 2021.

Discussion

Our study demonstrated that in a developing country’s public and universal healthcare systems, the COVID-19 pandemic negatively affected the increasing trend of breast biopsies, their direct costs, and BI-RADS 0 to III and IV to V mammograms observed in the prepandemic period. In addition, there was a 5-fold reduction in mammograms compared with breast biopsies, with an increase in breast biopsies ratio per mammogram in the COVID-19 pandemic. Furthermore, the negative impact on the trend of BI-RADS IV to V mammograms was less compared with that on the trend of BI-RADS 0 to III. Finally, in the interrupted time series, there was an association between the trend of breast biopsies and BI-RADS IV to V mammograms.

Many countries also suffered the negative impact of the pandemic on the early diagnosis of breast cancer, with the interruption or reduction of screening ranging from 1 to 6 months.^{5 , 11} One study estimated that an interruption in breast cancer screening at 3 and 6 months could increase 5-year mortality from 6.3% to 22.3% and advanced-stage disease from 30% to 109%, respectively.¹⁹ In our study, there was no interruption, but mammograms and breast biopsies were reduced at the beginning of the pandemic, lasting 3 to 4 months. In contrast, the subsequent recovery period was not enough to maintain the same trend observed in the prepandemic period, both in mammograms and breast biopsies. Another study in a highly developed country demonstrated a similar reduction period in mammograms and breast biopsies. Nevertheless, unlike our study, the subsequent recovery period was sufficient to maintain the same trend observed in the prepandemic period.²⁰ Therefore, despite the relatively short period of the decreasing trend of mammograms and breast biopsies in our study, the lack of a subsequent period of efficient recovery could mean a more prolonged negative impact of the COVID-19 pandemic on the early diagnosis of breast cancer.

Breast biopsy depends on radiological mammography classification. In 2018, about 1.6% of women screened with mammography in Brazil underwent breast biopsies.⁸ Our study demonstrated an increase in the breast biopsies rate per mammogram in 2020, greater than the expected rate (2.5%). Furthermore, the impact of the pandemic on BI-RADS IV to V mammograms was less compared with that on BI-RADS 0 to III mammograms. Therefore, our study’s results show a greater tendency to screen women at a higher risk for breast cancer during the COVID-19 pandemic, justifying a more significant number of BI-RADS IV to V mammograms and more breast biopsies performed. A similar result was observed in another study that identified a minor reduction in breast biopsies compared with mammograms during the COVID-19 pandemic.²⁰ The increased breast biopsies ratio per mammogram could have reduced the impact of failure to diagnose breast cancer early during the pandemic. New time series studies are essential to clarify this hypothesis and to evaluate the incidence of and mortality due to the disease over time.

Regarding the costs of breast biopsies, in November 2017, the government authorized a 100% increase in the amount paid for breast biopsies.²¹ After this period, there was stability in the amount paid for the breast biopsy until the end of the study.¹² This could explain a greater MPC in direct breast biopsy costs in the prepandemic period and the subsequent period of stability, consistent with the number of breast biopsies performed.

Our study reinforces the importance of developing systemic, strategic planning to ensure the continuity of promotion, prevention, and healthcare for the population, even during a pandemic. The financial resources available for public health are limited and must be managed appropriately to ensure the highest quality and efficient care for the greatest number of people. Promoting early breast cancer diagnosis can reduce the costs of treating advanced disease and mortality.⁴ The investment of 11.4 million US dollars in prevention strategies can result in savings of over 100 million US dollar in cancer treatment costs.⁴ In contrast, in Brazil, there is a growing trend in breast cancer mortality and direct costs with the treatment of advanced diseases.5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22 Therefore, it is essential to promote health policies to expand access to screening and breast biopsies to all regions of the country (including the most remote ones), increase service providers, periodically update the amounts paid for tests and medical procedures, guarantee an association between the screening trend and the growth of the target population, promote campaigns throughout the year to make the people aware of the importance of early breast cancer diagnosis, actively seek the target population for screening, manage the quality of services provided (for example, radiological classification of mammograms and indications for breast biopsies), and to unify data from public and private healthcare systems in the country for better management of health indicators for the population as a whole.

Our study has some limitations. The epidemiological data from the Brazilian population in DATASUS are based on projection studies from the last census (2010). All tests, medical procedures, and treatments performed and funded in the Brazilian public healthcare system are available on DATASUS. Despite the importance of this database, delays, absences, or errors in data recording may occur. Furthermore, tests and medical procedures performed in the Brazilian private healthcare system are unavailable on DATASUS. Studies of the trend of breast biopsies and their direct costs in the private healthcare system and different age groups are essential to compare with our research results. Finally, BI-RADS radiological classification information is unavailable on all mammograms performed and registered on DATASUS.

Conclusions

In a developing country's public and universal healthcare system, the COVID-19 pandemic harmed the increasing trend of breast biopsies, their total direct costs, and BI-RADS 0 to III and IV to V mammograms observed in the prepandemic. In addition, there has been a more significant trend in screening women at higher risk for breast cancer during the COVID-19 pandemic.

Article and Author Information

Author Contributions:Concept and design: Hyeda, da Costa, Kowalsky

Acquisition of data: Hyeda, da Costa

Analysis and interpretation of data: Hyeda, da Costa, Kowalsky

Drafting of the manuscript: Hyeda, da Costa, Kowalsky

Critical revision of the paper for important intellectual content: Hyeda, da Costa, Kowalsky

Statistical analysis: Hyeda, da Costa

Provision of study materials or patients: Hyeda, da Costa, Kowalsky

Administrative, technical, or logisticsupport: Hyeda, da Costa, Kowalsky

Supervision: da Costa, Kowalsky

Conflict of Interest Disclosures: The authors reported no conflicts of interest.

Funding/Support: The authors received no financial support for this research.