Evaluating the impact of hurricanes and the COVID‐19 pandemic on colorectal cancer incidence in Puerto Rico: An interrupted time‐series analysis

Abstract

Background

Major events, such as Hurricanes Irma and Maria and the coronavirus disease 2019 (COVID‐19) pandemic disrupted Puerto Rico's health system. Lack of access to colorectal cancer (CRC) screening services may have impeded timely diagnosis. The authors examined the impact of these events on CRC incidence in Puerto Rico.

Methods

The Puerto Rico Central Cancer Registry database allowed the authors to obtain CRC cases from 2012 to 2021. An interrupted time‐series analysis was performed to examine changes in CRC incidence immediately after and during the periods after the hurricanes and the pandemic. Analysis periods included: pre‐hurricanes, post‐hurricanes, and post‐COVID‐19 lockdown restrictions.

Results

We observed a level change of −8.3 CRC cases was observed in the month the hurricanes struck Puerto Rico, corresponding to an immediate decrease of 17.5%. After a slight upward trend, a second decline of 39.4 CRC cases was estimated after the COVID‐19 lockdown restrictions, representing an immediate change of −24.2%. By the end of the study, the estimated numbers of patients with early stage CRC patients and those aged 50–75 years did not reach the expected numbers. In addition, CRC cases in patients with late‐stage disease and in those aged younger than 50 years and aged 76 years and older exceeded the expected numbers.

Conclusions

Hurricanes Irma and Maria and the COVID‐19 pandemic caused a decrease in CRC incidence in Puerto Rico. This analysis suggests that limited access to CRC screening services during these events likely hindered CRC diagnoses. To fully understand the long‐term effects, monitoring of CRC trends will be necessary in the coming years.

Short abstract

To the authors' knowledge, this is the first interrupted time‐series analysis conducted in Puerto Rico analyzing the impact of Hurricanes Irma and Maria and the COVID‐19 pandemic on colorectal cancer incidence. The analysis revealed immediate decreases in colorectal cancer incidence after each major event, with late‐stage cases eventually exceeding expected levels.

INTRODUCTION

Colorectal cancer (CRC) is the third most diagnosed cancer and the second leading cause of cancer death worldwide. ¹ In 2023, CRC represented 7.8% of all new cancer cases in the United States. ² In Puerto Rico, a US territory where 98.9% of the population identifies as Hispanic/Latino, ³ CRC is the second most common cancer. The Cancer in Puerto Rico 2016–2020 report revealed that CRC accounted for 11.5% and 10.5% of all cancers in men and women, respectively. ⁴ Furthermore, Puerto Ricans have a higher incidence and mortality of CRC compared with US Hispanics. ⁵ Previous studies have highlighted CRC disparities in Puerto Rico, where the poverty rate is high and many people depend on Medicaid coverage. ³ These socioeconomic factors have been shown to contribute to poorer CRC outcomes. ⁶ , ⁷

Major events, such as hurricanes and pandemics, can hinder access to cancer screening services and delay cancer diagnosis. ⁸ In September 2017, Hurricane Maria hit Puerto Rico just 2 weeks after Hurricane Irma, resulting in extreme destruction across the archipelago. In the aftermath of the hurricanes, few hospitals were operating, which aggravated the already limited access to care. ⁸ , ⁹ In 2020, the coronavirus disease 2019 (COVID‐19) pandemic emerged, and Puerto Rico's government implemented stringent COVID‐19 lockdown measures. ¹⁰ Because people complied with the lockdown and isolation guidelines, they were effective in curbing contagion; however, this inadvertently led to a drop in the use of health services. ¹¹

Patients with chronic diseases, including cancer, are particularly vulnerable in the aftermath of climate‐related disasters. ¹² , ¹³ , ¹⁴ Disruptions to health care systems during and after these events can create significant challenges in managing population health. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ For patients with cancer, timely diagnosis and treatment are critical to effective outcomes. ²⁰ , ²¹ However, during disasters, medical services may be delayed or inaccessible because of damaged infrastructure, overburdened health care facilities, or shortages of medical personnel. ⁸ This can lead to late diagnoses, interruptions in treatment, and an overall worsening of health outcomes. ¹² , ¹⁵ , ¹⁶ The growing threat of climate change complicates this issue, increasing the frequency and intensity of extreme weather events, such as hurricanes. Therefore, building resilient health care systems that can withstand the pressures of climate‐related disasters will be critical to safeguarding the continuity of services for patients with cancer. ¹⁴

To date, no published studies have examined the extent to which hurricanes and the COVID‐19 pandemic affected CRC incidence levels in Puerto Rico. Addressing this gap in knowledge, we describe the impact of Hurricanes Irma and Maria and the COVID‐19 lockdown restrictions on the incidence trends and the changes in levels of CRC cases in Puerto Rico from 2012 to 2021.

MATERIALS AND METHODS

Data sources

We used monthly incidence cases from the Puerto Rico Central Cancer Registry (PRCCR) for this secondary data analysis. The PRCCR collects information on all cancer cases diagnosed and treated in Puerto Rico since 1950 and has reached 95% completeness of all cases since 2010, meeting the quality standards of the Centers for Disease Control and Prevention's National Program of Cancer Registries. ⁴ , ²² Using the PRCCR database allowed us to conduct research on the Puerto Rican population. Moreover, the majority of our research team is Puerto Rican, which ensured that the study is culturally relevant and that the team has a deep understanding of the population being studied.

Study group

From January 1, 2012, to December 31, 2021, there were 18,537 first‐time diagnosed CRC cases among Puerto Rican residents. These CRC cases were identified using the International Classification of Diseases for Oncology codes C18.0–C18.9, C19.9, C20.9, and C26.0, excluding histology types 9050–9055, 9140, and 9590–9993. We excluded 448 cases with an unknown month of diagnosis, leaving us with an overall cohort of 18,089 patients. In addition, for each stratified analysis, we excluded one patient with unknown age and 1985 patients with unknown stage.

Study variables

We classified stage at diagnosis into five categories using the National Cancer Institute's Surveillance, Epidemiology, and End Results general summary staging system: (1) in situ, in which the abnormal or precancerous cells have not spread to nearby tissues; (2) localized, in which the cancer is only located in the place of origin; (3) regional, in which the cancer has begun to spread to nearby tissues or organs; (4) distant, in which the cancer has spread to a different part of the body; and (5) unknown, in which there is not enough information to determine the stage. ²³ In our analysis, we stratified the stage as early (in situ and localized) and late (regional and distant). ²⁴ In addition, we divided the patients into three age groups (younger than 50 years, 50–75 years, and 76 years and older) based on the recommended screening age during the study period. ²⁵

Statistical analysis

We examined the changes in CRC incidence using an interrupted time series (ITS) analysis, a statistical method used to assess the effect of a public health intervention or event on a time‐series data set. ITS analyses identify trends and level changes at a specific point in time and are more effective when the effects of the interruption can be observed relatively soon after it occurs. ²⁶ , ²⁷ To enrich our ITS analysis, we used the lincom command to calculate the expected number of cases for each month if the event had not occurred (counterfactual analysis). ²⁸ The interruptions of interest were Hurricanes Irma and Maria, which were considered as a single event because they made landfall just 2 weeks apart in September 2017, and the start of the COVID‐19 lockdown restrictions, which began in April 2020. Specifically, we compared the trends in CRC cases across three distinct periods: (1) pre‐hurricanes (January 2012 to August 2017), (2) post‐hurricanes (September 2017 to March 2020), and (3) post–COVID‐19 lockdown restrictions (April 2020 to December 2021). Our data set included the number of new CRC cases diagnosed every month for a period of 10 years.

We executed the ITS analysis using the itsa command, a user‐written feature in Stata version 18.0 (StataCorp LLC). ²⁷ To run this command, we first selected our historical data and the interruption periods. For the standard ITS analysis, the following regression model is used:

$$Y_t={\beta }_0+{\beta }_1{T}_t+{\beta }_2{X}_t+{\beta }_3{X}_t{T}_{t,}$$

where Y _t is the aggregated value of the main outcome (e.g., monthly cases), T represents the time passed since the start of the study in the frequency at which the observations are taken (e.g., months), X _t is a dichotomous variable representing the interruption (0 = preinterruption; 1 = postinterruption), and X _t T _t is an interaction term. The itsa command provided estimates for the coefficient values as follows: β ₀ indicates the estimated value of our outcome variable at the beginning of the first period, β ₁ indicates the slope or trend of the outcome variable between the start of the study and the first interruption, β ₂ is the level change from the end of the trend line before the interruption and the start of the trend line after the interruption, and β ₃ is the difference in slope between the trend line before the interruption and the trend line after the interruption. ²⁷ Thus significant p values in β ₂ and β ₃ indicate an immediate and over‐time effect; respectively. ²⁸

By using the itsa command, we obtained various coefficients, such as the trends for each period, the level change between trends interrupted by the event, and the starting point of the first period. Next, we used the lincom command, based on the coefficients from the itsa command, to obtain point estimates of the expected incidence cases without the event, along with their p values and 95% confidence intervals (CIs). ²⁸ Then, we used the level change calculated by the ITS analysis and the expected number of cases calculated using lincom to evaluate the immediate change in CRC cases caused by the hurricanes (first interruption) or COVID‐19 lockdown restrictions (second interruption). This change was represented by a percentage difference compared with expected cases without the interruptions. ²⁹ The percentage difference formula was defined as:

$$\left(\frac{\text{Immediate}\text{change}\text{in}\text{cases}\text{after}\text{the}\text{Interruption}}{\text{Expected}\text{cases}\text{in}\text{the}\text{month}\text{of}\text{the}\text{interruption}}\right)\times 100.$$

In addition, when using ITS analysis, it is important to account for autocorrelation in the statistical estimation because of possible seasonal patterns in the data. This means that the number of cases per month may be uneven at different time periods or that other events near the interruption could influence the results, such as weather conditions or new health policies. ²⁶ Therefore, we conducted a sensitivity analysis with two regression‐based models; the first was Newey, with different lags to account for autocorrelation, and the second was Prais. The Newey option uses the ordinary least‐squares method with a Newey–West standard error. The Prais option transforms the estimate to fit a first‐order autoregression, or AR(1), model. ²⁷ Both methods produced similar results. Therefore, we opted to use Prais–Winsten AR(1) regression to model our data.

RESULTS

Table 1 lists the results for each study period by cancer stage and patient age at diagnosis. The initial estimated numbers of CRC cases for the entire group and for different subgroups also are included. During the pre‐hurricanes period, a trend line with a slope close to zero was observed overall (Figure 1). When stratifying by stage at diagnosis, the trend shows a slight increase (0.15 cases per month) for early stage CRC (p > .05; Figure 2A) but significantly decreases (−0.20 cases per month) for late‐stage CRC (p < .05; Figure 2B). When evaluating the same period stratified by age at diagnosis, the slope of the trend line is close to zero for all three age groups (Figure 3).

TABLE 1.

Impact of hurricanes and the coronavirus disease 2019 lockdown on colorectal cancer cases by stage and age group in Puerto Rico, 2012–2021.

Characteristic	Time point interruption
	Start of study	Hurricanes Irma and Maria			Start of post–COVID‐19 restrictions			End of study
	January 2012	September 2017			April 2020			December 2021
	Estimated initial no. of cases (95% CI)	Immediate change in the no. of cases (95% CI)	Expected no. of cases (95% CI) a	Percentage difference (95% CI), % b	Immediate change in the no. of cases (95% CI)	Expected no. of cases (95% CI) a	Percentage difference (95% CI), % b	Point estimate in the no. of cases	Expected no. of cases (95% CI) a
Overall	159.1 (146.7–171.6)	−28.3 (−49.9, −6.8)	161.4 (148.5–174.4)	−17.5 (−19.1, −16.2)	−39.4 (−66.8, −12.0)	162.5 (140.3–184.6)	−24.2 (−28.1, −21.3)	159.7	163.1 (134.8–191.5)
Stage
Early	63.1 (55.9–70.3)	−23.4 (−35.6, −11.2)	73.2 (65.7–80.7)	−32.0 (−35.6, −29.0)	−25.5 (−40.7, −10.3)	77.7 (64.9–90.5)	−32.8 (−39.3, −28.2)	68.2	80.6 (64.3–96.9)
Late	83.5 (77.8–89.1)	−3.3 (−13.4, 6.7)	69.8 (63.9–75.8)	−4.7 (−5.2, −4.4)	−6.1 (−19.2, 7.0)	63.7 (53.6–73.8)	−9.6 (−11.4, −8.3)	71.5	59.8 (46.8–72.6)
Age
<50 years	14.4 (12.4–16.4)	−0.9 (−4.5, 2.7)	15.7 (13.6–17.8)	−5.7 (−6.6, −5.1)	−1.5 (−6.2, 3.2)	16.3 (12.7–19.9)	−9.2 (−11.8, −7.5)	18.4	16.7 (12.1–21.3)
50–75 years	99.5 (90.0–109.0)	−19.5 (−35.8, −3.2)	102.7 (92.9–112.6)	−19.0 (−21.0, −17.3)	−26.4 (−47.0, −5.8)	104.2 (87.3–121.0)	−25.3 (−30.2, −21.8)	96.6	105.1 (83.6–126.7)
≥76 years	45.7 (42–49.3)	−6.1 (−12.7, 0.4)	42.3 (38.5–46.2)	−14.4 (−15.8, −13.2)	−10.3 (−18.9, −1.8)	40.9 (34.3–47.4)	−25.2 (−30.0, −21.7)	44.4	39.9 (31.5–48.2)

Abbreviations: CI, confidence interval; COVID‐19, coronavirus disease 2019.

^a Expected cases using a counterfactual approach; we assume the same pattern as the first period.

^b The immediate decline or increase caused by the event, represented by the percentage differences compared with the expected cases during the same month in the postevent period ([immediate change/expected cases] * 100); negative percentages represent fewer cases than expected after applying the previous formula.

FIGURE 1.

The monthly number of colorectal cancer cases from January 2012 to December 2021 in Puerto Rico. Solid lines in each period describe the trends in the expected number of cases per month, a red dashed line indicates the expected number of cases based on the first trend line (counterfactual analysis), dots are the observed number of colorectal cancer cases per month, and vertical dashed lines represent the start of each study period: September 2017 (Hurricanes Irma and Maria; first interruption) and April 2020 (coronavirus disease 2019 lockdown; second interruption). The asterisk indicates p < .05.

FIGURE 2.

The monthly number of colorectal cancer cases by cancer stage from January 2012 to December 2021 in Puerto Rico. (A, B) Solid lines in each period describe the trends of the expected number of cases per month, a red dashed line indicates the expected number of cases based on the first trend line (counterfactual analysis), dots are the observed number of colorectal cancer cases per month, and vertical dashed lines represent the start of each study period: September 2017 (Hurricanes Irma and Maria; first interruption) and April 2020 (coronavirus disease 2019 lockdown; second interruption). The asterisk in A indicates p < .05.

FIGURE 3.

The number of monthly colorectal cancer cases by age group from January 2012 to December 2021 in Puerto Rico. (A–C) Solid lines in each period describe the trends of the expected number of cases per month, a red dashed line indicates the expected number of cases based on the first trend line (counterfactual analysis), dots are the observed number of colorectal cancer cases per month, and vertical dashed lines represent the start of each study period: September 2017 (Hurricanes Irma and Maria; first interruption) and April 2020 (coronavirus disease 2019 lockdown; second interruption). Asterisks in A and C indicate p < .05.

In September 2017, Hurricanes Irma and Maria made landfall in Puerto Rico, marking the first interruption evaluated in our ITS analysis. In the absence of this interruption, 161.4 cases (95% CI, 148.5–174.4 cases) would have been expected. However, the observed number of CRC cases in September was 82 (see Table S1). The level change was estimated at −28.3 cases (95% CI, −49.9, −6.8 cases), representing an immediate change of −17.5%. Similarly, significant declines were observed by stage (−32.0% for early stage CRC and −4.7% for late‐stage CRC) and by age at diagnosis (−5.7% for patients younger than 50 years, −19.0% for those aged 50–75 years, and −14.4% for those aged 76 years and older; p < .05; Table 1). During the post‐hurricanes period, an upward trend of 0.50 overall monthly CRC cases was observed, although it was not statistically significant (p > .05; Figure 1). A similar pattern was observed when stratifying by stage (Figure 2) and by age at diagnosis (Figure 3), but these trends also were not statistically significant (p > .05).

April 2020 marked the second interruption evaluated in our analysis, with the beginning of the COVID‐19 lockdown restrictions. That month, the observed number of CRC cases was 50 (see Table S1). The expected number of cases without the interruptions would have been 162.5 (95% CI, 140.3–184.6 cases). The level change in the number of cases was −39.4 (95% CI, −66.8, −12.0), representing an immediate change of −24.2%. When stratifying by stage, the immediate change was −32.8% for early state CRC and −9.6% for late‐stage CRC. When stratifying by age at diagnosis, the immediate change was −9.2% for patients younger than 50 years, −25.3% for those aged 50–75 years, and −25.2% for those aged 76 years and older (p < .05; Table 1).

During the post–COVID‐19 lockdown restrictions period, an increase of 2.52 CRC cases per month was observed (p < .05; Figure 1). By using this trend, we described the point estimates of the number of cases for December 2021. In the overall analysis, the point estimate for this month was 159.7 cases, which nearly matched the expected number of cases without the interruptions (163.1 expected cases; Table 1). When stratifying by stage at diagnosis, early and late CRC stages showed a monthly increase of 1.94 (p < .05) and 0.26 (p > .05) cases, respectively (Figure 2). The early stage CRC point estimate for December 2021 was lower than the expected number of cases (68.2 point estimate vs. 80.6 expected cases in counterfactual analysis). Conversely, the late‐stage CRC point estimate for December 2021 surpassed the expected number of cases (71.5 point estimate vs. 59.8 expected cases in counterfactual analysis; Table 1). When stratifying by age at diagnosis, significant increases of 1.57 and 0.72 cases per month were observed for patients aged 50–75 years and those aged 76 years and older, respectively (p < .05; Figure 3A,C). However, a nonsignificant increase of 0.20 cases per month was observed for patients younger than 50 years (p > .05; Figure 3B). The point estimate in the number of cases for December 2021 among patients aged 50–75 years did not reach the expected number of cases (96.6 point estimate vs. 105.1 expected cases in counterfactual analysis; Table 1). Meanwhile, the point estimate in the number of cases for December 2021 slightly exceeded the expected cases among patients younger than 50 years (18.4 point estimate vs. 16.7 expected cases in counterfactual analysis) and those aged 76 years and older (44.4 point estimate vs. 39.9 expected cases in counterfactual analysis; Table 1). However, based on the 95% CIs of the expected cases in the counterfactual analysis, there was no evidence of significant differences in these comparisons.

DISCUSSION

In the current study, we evaluated the effect of two consecutive hurricanes and the COVID‐19 lockdown restrictions on CRC incidence in Puerto Rico from 2012 to 2021. We analyzed overall CRC cases as well as CRC cases stratified by age and disease stage. In all groups, disruptions were observed after each major event. Immediately after both interruptions, the number of CRC cases was lower than expected. However, the number of monthly cases increased during the remainder of each study period. Early stage CRC cases experienced the largest decrease after each interruption than any other stratified group. Conversely, late‐stage CRC cases had a smaller change in the number of cases after each interruption. By the end of the study period, the estimated number of late‐stage CRC cases and cases in patients outside the recommended screening age range (younger than 50 years and 76 years and older) exceeded the expected counts for each group.

We believe that these changes in trends after the hurricanes and the COVID‐19 pandemic were caused by disruptions in the availability and continuity of oncology services. Screening is the most effective approach for improved CRC outcomes because it enables early detection and intervention. ³⁰ Recently, the US Preventive Services Task Force updated its CRC screening recommendations in 2021 to lower the initial screening age to 45 years. ³¹ As of 2022, the Behavioral Risk Factor Surveillance System indicates that 55.5% of individuals aged 45–75 years in Puerto Rico met the full screening recommendations. This percentage is lower compared with the United States population, in which 66.9% of individuals received the recommended CRC screening test. ³²

A study on Puerto Rico's Medicaid population indicated that CRC screening rates in 2017 (after the hurricanes) and 2020 (during the COVID‐19 pandemic) were 25% and 39% lower, respectively, compared with the rates in 2016. ³³ Similarly, a study in Ontario, Canada, reported 41% fewer cancer screening tests in 2020 than in 2019. ³⁴ From 2018 to 2020, CRC screening rates in Puerto Rico increased from 55.6% to 71.8%. ³⁵ This upward trend aligns with the rise in CRC cases observed in the post‐hurricane period. However, the most recent data indicate a new decline in screening rates, which may be influenced by the updated guidelines that lowered the recommended screening age. ³¹ , ³²

Our study demonstrates how disruptive disasters can be to early diagnosis: early stage cases had the highest percentage difference compared with the expected number of CRC cases. In part, this may be because patients with early stage CRC are more likely to be asymptomatic. Meanwhile, patients with late‐stage CRC usually have severe symptoms that may lead them to seek medical care. ³⁶ , ³⁷ This delay in diagnosis can lead to an increase in late‐stage diagnoses in subsequent months or years, resulting in poorer prognoses and survival rates. ³⁸ In Canada, one study estimated that a 3‐month interruption in CRC screening could result in more CRC diagnoses than expected in the future, with 60% being diagnosed at an advanced stage. ³⁹ This was consistent with our data because the estimated number of late‐stage CRC cases exceeded the expected count after both events. Therefore, CRC cases that went undiagnosed during the hurricanes and the pandemic may have been detected later and at a more advanced stage than they would have been under normal circumstances. Because there is evidence that an increase in late‐stage diagnosis can increase the risk of death for patients with CRC, monitoring CRC outcomes is imperative. ⁴⁰ A study in England estimated an increase of 15.3%16.6% in CRC mortality up to 5 years after diagnosis. ¹⁵ Therefore, these additional deaths are expected to occur as a result of the COVID‐19 pandemic aftermath.

We also observed that the estimated number of CRC cases for patients in the recommended screening age (50–75 years) did not reached the expected number of cases at the end of the study. One possible explanation is that older patients face greater barriers to accessing health care and have higher risks associated with COVID‐19, which may have led to delays in seeking medical care. ¹² , ⁴¹ In Puerto Rico, the largest group of patients with CRC are older than 50 years. ⁴ Therefore, interventions to prevent health care service interruptions and promote CRC screening in these populations are critical.

Puerto Rico is no stranger to hurricanes, despite their often unprecedented consequences. Hurricane warnings are meant to allow time for preparation, evacuation, and emergency response. ⁸ However, the 2017 hurricanes severely damaged essential infrastructure, limiting access to water, electricity, and communication. ⁴² During the COVID‐19 pandemic, health care challenges were primarily structural and operational, including equipment shortages and lack of personnel. ¹⁹ In Puerto Rico, lockdowns may have delayed necessary cancer screenings, even when leaving home was permitted. ⁴³ To assess the geographic aspect of both events, we compared the metropolitan region with other regions. CRC diagnoses declined similarly across both areas during the COVID‐19 lockdown. However, during the hurricanes, the decline was smaller in the metropolitan region, likely because of its higher concentration of hospitals. This suggests that, although hurricanes disrupted health care access unevenly, the pandemic's executive orders led to a more systemic, island‐wide disruption (see Table S2 and Figure S1).

The aftermath of both events, together with the data from our study, underscores the urgent need to update emergency plans, specifically considering oncology services throughout the entire cancer care continuum. Efforts are needed to develop specific plans that are effective in mitigating the effect of these events on cancer, taking into consideration the vulnerabilities of our population. This will help mitigate the effect of such events on the populations well‐being and expedite disaster recovery. ¹⁴

Our study has some limitations that should be noted. First, because data after December 2021 were not yet available, we could not analyze how the CRC incidence trends behaved during the whole pandemic period. However, we were able to assess CRC trends when the first COVID‐19 restrictions were put in effect. As additional data become available, an evaluation of the whole COVID‐19 pandemic period and its long‐term effects may be feasible. Second, more research should be done to identify other factors, such as emigration, that may have affected CRC trends after these major events. Nevertheless, the PRCCR participates in data exchange with some registries in the United States. Therefore, the data that we used included incidence cases of Puerto Ricans temporarily staying in the United States. Third, approximately 11% of cancer stage data were missing, leading to their exclusion in the stratified analysis. Despite these limitations, to our knowledge, this is the first ITS study performed in Puerto Rico to analyze the effect of hurricanes and pandemics on CRC incidence. It represents an opportunity to use this methodology more often to observe interruptions or interventions during specific time points and their effect on cancer. Other studies, such as a study in Zambia, have used ITS analyses to understand the effect of the COVID‐19 pandemic and have used their findings to improve diagnostic tools. ²⁹

CONCLUSION

In this analysis, we have demonstrated that major events, such as climate‐related disasters and pandemics, have a definitive effect on CRC incidence. Because it has been established that climate‐related events worsen outcomes for patients with cancer, assessing the implications of this decline in CRC incidence is crucial. ⁴⁴ Lower incidence immediately after the disasters could lead to more CRC cases diagnosed at later stages in the future, which potentially can lower survival rates. Understanding the association of these major events with cancer incidence is necessary to develop public health responses, such as proactive planning and preparation to mitigate disruptions in cancer care. Therefore, strategies must be developed that focus on providing more support for cancer screening, diagnosis, and treatment during and after climate‐related disasters and other major events.

AUTHOR CONTRIBUTIONS

Tonatiuh Suárez‐Ramos: Conceptualization; data curation; formal analysis; methodology; visualization; validation. writing–original draft; and writing–review and editing. Samantha Verganza: Writing–original draft; conceptualization; formal analysis; and validation. Yisel Pagán‐Santana: Methodology; visualization; writing–original draft; and writing–review and editing. Maira A. Castañeda‐Avila: Writing–review and editing. Carlos R. Torres‐Cintrón: Conceptualization and writing–review and editing. Eduardo J. Santiago‐Rodríguez: Methodology and writing–review and editing. Karen J. Ortiz‐Ortiz: Conceptualization; data curation; supervision; and writing–review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declared no conflicts of interest.

Supporting information

Supplementary Material

Suárez‐Ramos T, Verganza S, Pagán‐Santana Y, et al. Evaluating the impact of hurricanes and the COVID‐19 pandemic on colorectal cancer incidence in Puerto Rico: An interrupted time‐series analysis. Cancer. 2025;e35793. doi: 10.1002/cncr.35793

DATA AVAILABILITY STATEMENT

The data that support this study's findings come from the Puerto Rico Central Cancer Registry (PRCCR) database. Because of a confidentiality agreement between the PRCCR and the authors, the clinical data from this study are not publicly available. However, investigators may obtain these data through the PRCCR by following its confidentiality procedures and requesting the data online at https://rcpr.org/Datos‐de‐C%C3%A1ncer/Acceso‐a‐Datos. The case count data used for the interrupted time‐series analysis have been de‐identified and are included in Table S1.