The COVID-19 pandemic abruptly upended cancer care in many countries including the US. The need to reduce community spread and reserve hospital capacity for the most severe COVID-19 cases led to rescheduling or postponement of cancer care appointments (Nelson, 2020; Patt et al., 2020; Ueda et al., 2020; Zheng et al., 2021; Wenger et al., 2022). These control measures significantly decreased cancer-related patient encounters in the early phase of the pandemic, particularly for cancer screenings (Patt et al., 2020). Comparing March to July 2020 with the same period in 2019, there was a substantial decrease in cancer screenings, biopsies, surgeries, office visits, and therapy; the decreases varied by service location and cancer type (Patt et al., 2020). For example, breast cancer screenings decreased by 89.2% and colorectal by 84.5% (Warner et al., 2020). Patients reported delays in receiving cancer care, including follow-up clinic appointments and cancer therapies, such as radiation, infusion therapies, and surgeries (London et al., 2022; Riera et al., 2021).

Cancer care delays due to the COVID-19 pandemic are anticipated to lead to increased cancer morbidity and mortality (Blay et al., 2021; Malagón et al., 2022). One study found an association between surgical and screening delays and increased cancer mortality among patients diagnosed with colorectal, lung, and prostate cancer during the pandemic (Zheng et al., 2021). Delayed mammography and computed tomography for lung cancer were associated with advanced stage of cancer at diagnosis (Zheng et al., 2021). Another study determined delayed surgery for lung cancer was associated with worse survival (Mayne et al., 2021). For breast screenings, some evidence suggests that patients were reluctant to return for mammograms after care disruptions (Miller et al., 2021). Thus, cancer care disruptions during COVID-19 could have detrimental future impacts on cancer outcomes and may require changes to public health and clinical strategies across the cancer prevention and control continuum.

It is unclear if patients felt comfortable returning to care in the context of rapidly changing information and guidelines related to COVID-19 and even now that guidelines are more consistent and vaccines are available. There is concern about whether patients will prioritize immediate unmet social needs that might be a result of or exacerbated by COVID-19, such as food insecurity, employment loss, and housing challenges, over disease prevention. Furthermore, people of color, including African Americans, Latinx, and Native communities, as well as those employed in low-wage occupations, are likely to have greater concerns over COVID-19 safety, in addition to the immediate concerns noted above (Cancino et al., 2020). Rural communities that already experience limited access to cancer care, have less capacity to manage COVID-19 (Segel et al., 2021). Finally, hospitals rapidly increased the use of telehealth to continue cancer care during COVID-19, but older people and those who lived in low-income and rural areas, or were less likely to have commercial insurance were less likely to participate (Darcourt et al., 2021; Jaffe et al., 2020). This combination of factors may exacerbate existing disparities (Cancino et al., 2020).

This survey study was conducted by National Cancer Institute (NCI)—designated Siteman Cancer Center to elucidate: (1) to what extent cancer care appointments (including preventive screenings and treatment) in the bi-state Midwestern catchment area were canceled/postponed, (2) patients’ needs for returning to care, and (3) correlates of care disruption across the catchment area. This study aligns with the NCI’s goal to support population health assessments of their cancer center’s catchment areas. In our catchment area, the cancer burden is significantly greater than the US averages for multiple cancers. Moreover, racial and geographical disparities persist such that African American patients have higher incidence and mortality for lung, colorectal, late-stage breast cancer diagnoses, and prostate cancers compared to White patients. Rural counties also have higher mortality (but not incidence) for melanoma, breast, and prostate cancer compared to urban areas (National Cancer Institute, 2022).

Thus, we explore how socio-contextual factors impact cancer health disparities across the continuum of cancer control and prevention during COVID-19 in this bi-state Midwestern catchment area. This analysis is guided by the theory that social identities like race, ethnicity, social class, and gender shape many contextual factors related to cancer, COVID, and other outcomes and are ultimately the fundamental drivers of disease. We stratify our results by race because of the differential impact of COVID-19 on communities of color and the over-representation of socioeconomic factors such as low-income, low-wage work often experienced by communities of color (Acosta et al., 2021; Athavale et al., 2021; Millett et al., 2020).

Using primary data collected from residents across the 82-county Siteman catchment area overlapping Missouri and Illinois, we learned that across different type of screenings, 21% to 57% of cancer screening or treatment appointments were canceled/postponed from March 2020 through the end of 2020. Across all races, residents with higher educational attainment had 1.25 higher odds of care disruption for general or cancer care compared to residents with lower educational attainment; this association remained significant among Black and White residents. Additionally, White residents of older age, assigned female at birth, or having scheduled a telehealth appointment, also had higher odds of care disruption. Interestingly, while the sample size is small, we did see a trend that suggested that Black people with 3+comorbidities were more likely to cancel/postpone care. Finally, knowing their doctor’s office or clinic is taking the appropriate COVID-related safety precautions was the greatest reported need for returning to care (33.8%).

Delays in cancer screening can lead to stage shifts where patients are diagnosed at later stages and thus have a higher risk for cancer morbidity and mortality. Understanding which screenings were impacted locally and for whom and identifying patient concerns can inform community outreach and engagement efforts. This allows programs to target groups in their catchment, most likely to have delayed screening and draft messaging that can alleviate patient concerns and in turn facilitate a return to care. Moreover, knowing that those with 3+comorbidities, who likely have additional health needs were more likely to cancel/postpone appointments could also inform recruitment methods in getting people to return to care. It is possible that these cancellations and postponements were due to an increased risk of contracting and dying of COVID; however, it is critical to get this high need population to return to care so widening disparities are avoided.

Mammograms and Pap tests are an area of increased interest for our catchment area given the high number of women scheduled for screening. Approximately 38% of the 170 women who were scheduled for mammograms had canceled/postponed appointments. Similarly, 45% of the 188 women scheduled for Pap tests had canceled/postponed appointments. Delays in colorectal cancer screening impacted a smaller number of people, but colorectal cancer screening is an important area given the high proportion of cancellations/postponements, overall low number of scheduled appointments in general, and high colorectal cancer disparities in the region. Of the 51 people scheduled for a colonoscopy, 57% canceled/postponed appointments, and of the 38 scheduled for a blood stool test, 29% canceled/postponed appointments as well. To help healthcare systems reduce the cancer screening deficit, community outreach and engagement strategies need to address these needs. For example, employing mobile strategies such as the use of mobile mammography and home-based cervical and colorectal cancer screening tests could serve those most impacted.

These data are consistent with prior literature that suggests a reduction in general medical and cancer-related appointments (Patt et al., 2020; Wenger et al., 2022; London et al., 2022; Czeisler et al., 2020). This study allows us to understand the magnitude of the impact across eastern/southeastern Missouri and southern Illinois. Future research exploring whether those with higher educational attainment were more likely to cancel/postpone appointments because they were more likely to have better access to scheduling future appointments could further elucidate the extent of educational disparities in healthcare access.

These cross-sectional data cannot infer causality however, many of the correlates of interest (e.g. race, educational attainment) pre-date COVID-19 and the need to consider postponing clinical care. Thus, it is unlikely these results are subject to reverse causation. Also, those excluded due to missing data were more likely to be uninsured. If uninsured persons were also more likely to have canceled/postponed appointments, this could potentially bias results about care disruptions by insurance status towards the null and underestimate the impact. Finally, unmeasured confounding is possible in cross-sectional studies like ours. Despite these limitations, this is a significant study that can improve our understanding of COVID-19 impacts on cancer prevention and control and offer specific insights into the region. In our data, those with higher education were more likely to cancel/postpone care. This indicates that any trends seen in increasing late-stage diagnosis might occur across socioeconomic categories. Additionally, while Black and White people of higher educational attainment both had increased odds of care disruption, having a scheduled telehealth visit was significantly associated with higher odds of care disruption only for White residents. This suggests that while White people were canceling/postponing in-person care, this care may have been substituted with telehealth appointments. Many of these screenings cannot be done virtually, yet this warrants further investigation to understand if care disruption does not always equate to being disconnected from healthcare for some and the subsequent impact on racial disparities in cancer care.

The data that support the findings of this study are openly available on Open Science Framework at https://osf.io/p5x3s/. Please cite as data from this publication if used.

1. 1. Kia L D
   2. Nicole A
   3. Lisa K
   4. Nora L
   5. Sarah H
   6. Bettina D
   7. Vetta LST

   (2023) Open Science Framework

   ID p5x3s. Siteman Cancer Center COVID and cancer data.

   https://osf.io/p5x3s/

1. 1. Acosta AM
   2. Garg S
   3. Pham H
   4. Whitaker M
   5. Anglin O
   6. O’Halloran A
   7. Milucky J
   8. Patel K
   9. Taylor C
   10. Wortham J
   11. Chai SJ
   12. Kirley PD
   13. Alden NB
   14. Kawasaki B
   15. Meek J
   16. Yousey-Hindes K
   17. Anderson EJ
   18. Openo KP
   19. Weigel A
   20. Monroe ML
   21. Ryan P
   22. Reeg L
   23. Kohrman A
   24. Lynfield R
   25. Bye E
   26. Torres S
   27. Salazar-Sanchez Y
   28. Muse A
   29. Barney G
   30. Bennett NM
   31. Bushey S
   32. Billing L
   33. Shiltz E
   34. Sutton M
   35. Abdullah N
   36. Talbot HK
   37. Schaffner W
   38. Ortega J
   39. Price A
   40. Fry AM
   41. Hall A
   42. Kim L
   43. Havers FP

   (2021) Racial and Ethnic Disparities in Rates of COVID-19-Associated Hospitalization, Intensive Care Unit Admission, and In-Hospital Death in the United States From March 2020 to February 2021

   JAMA Network Open 4:e2130479.

   https://doi.org/10.1001/jamanetworkopen.2021.30479

   • PubMed
   • Google Scholar
2. 1. Athavale P
   2. Kumar V
   3. Clark J
   4. Mondal S
   5. Sur S

   (2021) Differential Impact of COVID-19 Risk Factors on Ethnicities in the United States

   Frontiers in Public Health 9:743003.

   https://doi.org/10.3389/fpubh.2021.743003

   • PubMed
   • Google Scholar
3. Software

   1. Blake K

   (2019)

   Health information national trends survey (HINTS) 5

   Cycle 2 Dataset.
4. 1. Blay JY
   2. Boucher S
   3. Le Vu B
   4. Cropet C
   5. Chabaud S
   6. Perol D
   7. Barranger E
   8. Campone M
   9. Conroy T
   10. Coutant C
   11. De Crevoisier R
   12. Debreuve-Theresette A
   13. Delord JP
   14. Fumoleau P
   15. Gentil J
   16. Gomez F
   17. Guerin O
   18. Jaffré A
   19. Lartigau E
   20. Lemoine C
   21. Mahe MA
   22. Mahon FX
   23. Mathieu-Daude H
   24. Merrouche Y
   25. Penault-Llorca F
   26. Pivot X
   27. Soria JC
   28. Thomas G
   29. Vera P
   30. Vermeulin T
   31. Viens P
   32. Ychou M
   33. Beaupere S

   (2021) Delayed care for patients with newly diagnosed cancer due to COVID-19 and estimated impact on cancer mortality in France

   ESMO Open 6:100134.

   https://doi.org/10.1016/j.esmoop.2021.100134

   • PubMed
   • Google Scholar
5. 1. Cancino RS
   2. Su Z
   3. Mesa R
   4. Tomlinson GE
   5. Wang J

   (2020) The Impact of COVID-19 on Cancer Screening: Challenges and Opportunities

   JMIR Cancer 6:e21697.

   https://doi.org/10.2196/21697

   • PubMed
   • Google Scholar
6. Website

   1. Centers for Disease Control and Prevention

   (2018) Behavioral Risk Factor Surveillance System Survey Questionnaire

   Accessed April 8, 2020.

   https://www.cdc.gov/brfss/questionnaires/index.htm
7. 1. Czeisler MÉ
   2. Marynak K
   3. Clarke KEN
   4. Salah Z
   5. Shakya I
   6. Thierry JM
   7. Ali N
   8. McMillan H
   9. Wiley JF
   10. Weaver MD
   11. Czeisler CA
   12. Rajaratnam SMW
   13. Howard ME

   (2020) Delay or Avoidance of Medical Care Because of COVID-19-Related Concerns - United States, June 2020

   MMWR. Morbidity and Mortality Weekly Report 69:1250–1257.

   https://doi.org/10.15585/mmwr.mm6936a4

   • PubMed
   • Google Scholar
8. 1. Darcourt JG
   2. Aparicio K
   3. Dorsey PM
   4. Ensor JE
   5. Zsigmond EM
   6. Wong ST
   7. Ezeana CF
   8. Puppala M
   9. Heyne KE
   10. Geyer CE
   11. Phillips RA
   12. Schwartz RL
   13. Chang JC

   (2021) Analysis of the Implementation of Telehealth Visits for Care of Patients With Cancer in Houston During the COVID-19 Pandemic

   JCO Oncology Practice 17:e36–e43.

   https://doi.org/10.1200/OP.20.00572

   • PubMed
   • Google Scholar
9. Book

   1. Grasso DJ
   2. Briggs-Gowan MJ
   3. Ford JD
   4. Carter AS

   (2020)

   The Epidemic – Pandemic Impacts Inventory (EPII)

   University of Connecticut School of Medicine.

   • Google Scholar
10. 1. Jaffe DH
    2. Lee L
    3. Huynh S
    4. Haskell TP

    (2020) Health Inequalities in the Use of Telehealth in the United States in the Lens of COVID-19

    Population Health Management 23:368–377.

    https://doi.org/10.1089/pop.2020.0186

    • PubMed
    • Google Scholar
11. Website

    1. Killermann S

    (2020) How you can make the gender question on an application form more inclusive

    Accessed May 21, 2020.

    https://www.itspronouncedmetrosexual.com/2012/06/how-can-i-make-the-gender-question-on-an-application-form-more-inclusive
12. 1. London JW
    2. Fazio-Eynullayeva E
    3. Palchuk MB
    4. McNair C

    (2022) Evolving Effect of the COVID-19 Pandemic on Cancer-Related Encounters

    JCO Clinical Cancer Informatics 6:e2100200.

    https://doi.org/10.1200/CCI.21.00200

    • PubMed
    • Google Scholar
13. 1. Malagón T
    2. Yong JHE
    3. Tope P
    4. Miller WH
    5. Franco EL
    6. McGill Task Force on the Impact of COVID-19 on Cancer Control and Care

    (2022) Predicted long-term impact of COVID-19 pandemic-related care delays on cancer mortality in Canada

    International Journal of Cancer 150:1244–1254.

    https://doi.org/10.1002/ijc.33884

    • PubMed
    • Google Scholar
14. 1. Mayne NR
    2. Elser HC
    3. Darling AJ
    4. Raman V
    5. Liou DZ
    6. Colson YL
    7. D’Amico TA
    8. Yang C-FJ

    (2021) Estimating the Impact of Extended Delay to Surgery for Stage I Non-small-cell Lung Cancer on Survival

    Annals of Surgery 273:850–857.

    https://doi.org/10.1097/SLA.0000000000004811

    • PubMed
    • Google Scholar
15. 1. Miller MM
    2. Meneveau MO
    3. Rochman CM
    4. Schroen AT
    5. Lattimore CM
    6. Gaspard PA
    7. Cubbage RS
    8. Showalter SL

    (2021) Impact of the COVID-19 pandemic on breast cancer screening volumes and patient screening behaviors

    Breast Cancer Research and Treatment 189:237–246.

    https://doi.org/10.1007/s10549-021-06252-1

    • PubMed
    • Google Scholar
16. 1. Millett GA
    2. Jones AT
    3. Benkeser D
    4. Baral S
    5. Mercer L
    6. Beyrer C
    7. Honermann B
    8. Lankiewicz E
    9. Mena L
    10. Crowley JS
    11. Sherwood J
    12. Sullivan PS

    (2020) Assessing differential impacts of COVID-19 on black communities

    Annals of Epidemiology 47:37–44.

    https://doi.org/10.1016/j.annepidem.2020.05.003

    • Google Scholar
17. Website

    1. National Cancer Institute

    (2022) State Cancer Profiles

    Accessed October 1, 2020.

    https://statecancerprofiles.cancer.gov/index.html
18. 1. Nelson B

    (2020) Covid-19 is shattering US cancer care

    BMJ 369:m1544.

    https://doi.org/10.1136/bmj.m1544

    • PubMed
    • Google Scholar
19. 1. Patt D
    2. Gordan L
    3. Diaz M
    4. Okon T
    5. Grady L
    6. Harmison M
    7. Markward N
    8. Sullivan M
    9. Peng J
    10. Zhou A

    (2020) Impact of COVID-19 on Cancer Care: How the Pandemic Is Delaying Cancer Diagnosis and Treatment for American Seniors

    JCO Clinical Cancer Informatics 4:1059–1071.

    https://doi.org/10.1200/CCI.20.00134

    • PubMed
    • Google Scholar
20. 1. Peek ME
    2. Nunez-Smith M
    3. Drum M
    4. Lewis TT

    (2011)

    Adapting the everyday discrimination scale to medical settings: reliability and validity testing in a sample of African American patients

    Ethnicity & Disease 21:502–509.

    • PubMed
    • Google Scholar
21. Website

    1. Penedo FJ
    2. Cohen L
    3. Bower J
    4. Antoni MH

    (2020) COVID-19: Impact of the pandemic and HRQOL in cancer patients and survivors

    Accessed June 1, 2020.

    https://clinicaltrials.gov/ct2/show/NCT04447222
22. 1. Qi AC
    2. Peacock K
    3. Luke AA
    4. Barker A
    5. Olsen MA
    6. Joynt Maddox KE

    (2019) Associations between social risk factors and surgical site infections after colectomy and abdominal hysterectomy

    JAMA Network Open 2:e1912339.

    https://doi.org/10.1001/jamanetworkopen.2019.12339

    • PubMed
    • Google Scholar
23. Software

    1. Qualtrics

    (2020)

    Provo

    USA Version.
24. 1. Riera R
    2. Bagattini ÂM
    3. Pacheco RL
    4. Pachito DV
    5. Roitberg F
    6. Ilbawi A

    (2021) Delays and Disruptions in Cancer Health Care Due to COVID-19 Pandemic: Systematic Review

    JCO Global Oncology 7:311–323.

    https://doi.org/10.1200/GO.20.00639

    • PubMed
    • Google Scholar
25. 1. Saez-Clarke E
    2. Otto AK
    3. Prinsloo S
    4. Natori A
    5. Wagner RW
    6. Gomez TI
    7. Ochoa JM
    8. Tworoger SS
    9. Ulrich CM
    10. Hathaway CA
    11. Peoples AR
    12. Antoni MH
    13. Bower JE
    14. Cohen L
    15. Penedo FJ

    (2023) Development and initial psychometric evaluation of a COVID-related psychosocial experiences questionnaire for cancer survivors

    Quality of Life Research 2023:1–20.

    https://doi.org/10.1007/s11136-023-03456-4

    • PubMed
    • Google Scholar
26. 1. Segel JE
    2. Ross HI
    3. Edwards JL
    4. Braun KA
    5. Davis LA

    (2021) The Unique Challenges Facing Rural Providers in the COVID-19 Pandemic

    Population Health Management 24:304–306.

    https://doi.org/10.1089/pop.2020.0151

    • PubMed
    • Google Scholar
27. 1. Ueda M
    2. Martins R
    3. Hendrie PC
    4. McDonnell T
    5. Crews JR
    6. Wong TL
    7. McCreery B
    8. Jagels B
    9. Crane A
    10. Byrd DR
    11. Pergam SA
    12. Davidson NE
    13. Liu C
    14. Stewart FM

    (2020) Managing Cancer Care During the COVID-19 Pandemic: Agility and Collaboration Toward a Common Goal

    Journal of the National Comprehensive Cancer Network pp. 1–4.

    https://doi.org/10.6004/jnccn.2020.7560

    • PubMed
    • Google Scholar
28. Website

    1. United States Census Bureau

    (2022) Population and Housing Unit Estimates

    Accessed October 1, 2022.

    https://www.census.gov/programs-surveys/popest.html
29. Website

    1. United States Department of Agriculture

    (2019) Rural-Urban Continuum Codes

    Accessed May 14, 2020.

    https://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx
30. 1. Wadhera R.K
    2. Choi E
    3. Shen C
    4. Yeh RW
    5. Joynt Maddox KE

    (2019) Trends, causes, and outcomes of hospitalizations for homeless individuals: a retrospective cohort study

    Medical Care 57:21–27.

    https://doi.org/10.1097/MLR.0000000000001015

    • PubMed
    • Google Scholar
31. 1. Wadhera RK
    2. Yeh RW
    3. Joynt Maddox KE

    (2020) Database inaccuracies and disparities in care among homeless adults hospitalized for cardiovascular conditions-reply

    JAMA Internal Medicine 180:614–615.

    https://doi.org/10.1001/jamainternmed.2020.0021

    • PubMed
    • Google Scholar
32. 1. Warner ET
    2. Restrepo E
    3. Benjamin C
    4. Fairley R
    5. Roudebush L
    6. Eshraghi L
    7. Hertz C
    8. Du S
    9. Carfang L

    (2020) Abstract S11-02: Patient-reported impact of the COVID-19 pandemic on breast cancer screening, diagnosis, and treatment: A national survey

    Clinical Cancer Research 26:S11-02.

    https://doi.org/10.1158/1557-3265.COVID-19-S11-02

    • PubMed
    • Google Scholar
33. 1. Wenger NS
    2. Stanton AL
    3. Baxter-King R
    4. Sepucha K
    5. Vavreck L
    6. Naeim A

    (2022) The Impact of COVID-19 on Routine Medical Care and Cancer Screening

    Journal of General Internal Medicine 37:1450–1456.

    https://doi.org/10.1007/s11606-021-07254-x

    • Google Scholar
34. 1. Zheng NS
    2. Warner JL
    3. Osterman TJ
    4. Wells QS
    5. Shu XO
    6. Deppen SA
    7. Karp SJ
    8. Dwyer S
    9. Feng Q
    10. Cox NJ
    11. Peterson JF
    12. Stein CM
    13. Roden DM
    14. Johnson KB
    15. Wei WQ

    (2021) A retrospective approach to evaluating potential adverse outcomes associated with delay of procedures for cardiovascular and cancer-related diagnoses in the context of COVID-19

    Journal of Biomedical Informatics 113:103657.

    https://doi.org/10.1016/j.jbi.2020.103657

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Understanding disruptions in cancer care to reduce increased cancer burden: A cross-sectional study" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and I have overseen the evaluation in my dual role of Reviewing Editor and Senior Editor.

Essential revisions:

As is customary in eLife, the reviewers have discussed their critiques with one another. What follows below is an edited compilation of the essential and ancillary points provided by reviewers in their critiques and in their interaction post-review. Please submit a revised version that addresses these concerns directly. Although we expect that you will address these comments in your response letter, we also need to see the corresponding revision clearly marked in the text of the manuscript. Some of the reviewers' comments may seem to be simple queries or challenges that do not prompt revisions to the text. Please keep in mind, however, that readers may have the same perspective as the reviewers. Therefore, it is essential that you attempt to amend or expand the text to clarify the narrative accordingly.

Reviewer #1 (Recommendations for the authors):

The study lacks focus and significance. The methods used are not adequate for a rigorous study. There were disruptions in cancer care and other treatments during the first year of Covid (June-August 2020). While it is important to investigate whether there were disparities or inequities in access to care, the data used – a cross-sectional analysis of primary data of 650 people – is not the right data to be used to assess the disruption in cancer care.

To investigate disruption in care, I suggest using the difference-in-differences or a similar method. I would also suggest the authors use national survey data or other large datasets to conduct a trend analysis (before and after). This can provide useful information on the type of illnesses (or type of cancer treatments) or patient subpopulations that were impacted more than others by Covid – interruption in care. To conduct any type of disparity or inequity analysis, investigators need to ensure that they match subpopulations that they like to compare to one another based on their health needs. Otherwise, there are confounding variables that most certainly would bias the results.

What conceptual model did the authors use to select the explanatory variables? It seems to me that authors just throw a bunch of variables in a logistic regression model. If this study examined disruption to care, the most important explanatory variables to include in the model were the basic health status or severity and type of cancer.

The sample size was small, and with the cross-sectional nature of the study, no conclusion can be achieved. The characteristics of people with missing data were not presented and discussed. This can easily create a biased sample to begin with. The variables used and the rationale to include them have not been discussed.

Reviewer #2 (Recommendations for the authors):

I appreciate the opportunity to review your manuscript submitted to eLife. The manuscript addresses an important topic, and it will be a valuable contribution to the literature on the impact of COVID-19 on cancer care.

1. Outcome measures were adapted from an unpublished questionnaire by Penedo et al. Please provide more information about how these measures were developed and/or whether they were validated.

2. Figure 1 shows that 53 participants had appointments for cancer-related care other than screening. Because this finding is presented in Figure 1, it should be discussed in the results. A footnote in Supplemental Table 1 incorrectly states that the variable was not included in the manuscript.

3. The analysis should address the assumptions for conducting logistic regression. Specifically, authors should test for (1) linearity between continuous independent variables and the log-odds of care disruption, and (2) absence of multicollinearity. In addition, the authors treated education and income as continuous variables -- please explain this decision.

4. Cancer care disruptions were stratified by race. Did the "Patient needs for rescheduling" likewise vary based on race or other factors?

5. Please consider adding a supplemental table to show the univariate characteristics of Non-Hispanic Black or African American and Non-Hispanic White subgroups, which are stratified in the multivariate analysis in Table 2. The analysis was stratified by race because of the differential socioeconomic factors and COVID-19 impacts among communities of color -- it would be helpful to see the extent to which the combination of factors is reflected in the sample.

Essential revisions:

As is customary in eLife, the reviewers have discussed their critiques with one another. What follows below is an edited compilation of the essential and ancillary points provided by reviewers in their critiques and in their interaction post-review. Please submit a revised version that addresses these concerns directly. Although we expect that you will address these comments in your response letter, we also need to see the corresponding revision clearly marked in the text of the manuscript. Some of the reviewers' comments may seem to be simple queries or challenges that do not prompt revisions to the text. Please keep in mind, however, that readers may have the same perspective as the reviewers. Therefore, it is essential that you attempt to amend or expand the text to clarify the narrative accordingly.

Reviewer #1 (Recommendations for the authors):

The study lacks focus and significance. The methods used are not adequate for a rigorous study. There were disruptions in cancer care and other treatments during the first year of Covid (June-August 2020). While it is important to investigate whether there were disparities or inequities in access to care, the data used – a cross-sectional analysis of primary data of 650 people – is not the right data to be used to assess the disruption in cancer care.

To investigate disruption in care, I suggest using the difference-in-differences or a similar method. I would also suggest the authors use national survey data or other large datasets to conduct a trend analysis (before and after). This can provide useful information on the type of illnesses (or type of cancer treatments) or patient subpopulations that were impacted more than others by Covid – interruption in care. To conduct any type of disparity or inequity analysis, investigators need to ensure that they match subpopulations that they like to compare to one another based on their health needs. Otherwise, there are confounding variables that most certainly would bias the results.

Thank you for this comment. We agree that it is important to understand national trends in COVID-related care disruptions and using causal methods. However, this manuscript was not intended to be generalizable to the nation. Instead, it aimed to examine the local impact of COVID care disruptions. We focused on the Siteman Cancer Center’s (SCC) catchment area because the co-author team includes the SCC’s Associate Director of Community Outreach and Engagement (COE) program, the SCC Associate Director for Diversity, Equity, and Inclusion, multiple members of the SCC COE leadership team. Thus, we are uniquely positioned to mobilize and identify outreach opportunities and/or programs that address any gaps we discover. Moreover, this focus on our catchment area aligns with the National Cancer Institute’s priorities to characterize cancer-relevant knowledge across cancer center catchment areas. Finally, this is our first cross-sectional study of our catchment area, so we could not implement a difference-in-difference analysis. We will regularly survey residents in our catchment in the future. We will be able to track trends over time and consider additional methods like difference-in-difference once we have multiple data points. We have added text in the manuscript to emphasize the purpose of this study as a local, area-level assessment and to indicate that confounding bias is possible in cross-sectional studies, including this one.

What conceptual model did the authors use to select the explanatory variables? It seems to me that authors just throw a bunch of variables in a logistic regression model.

Our analysis was guided by the theory that social identities related to race, ethnicity, class, and gender shape access to healthcare and disease processes and are the fundamental drivers of health. We edited the sentence below in the methods section to help clarify this.

“This analysis is guided by the theory that social identities like race, ethnicity, social class, and gender shape many contextual factors related to cancer, COVID, and other health outcomes and are ultimately, the fundamental drivers of disease”.

If this study examined disruption to care, the most important explanatory variables to include in the model were the basic health status or severity and type of cancer.

We did not ask severity of cancer, but we conducted a posthoc analysis to understand care disruption among those with multiple comorbidities to understand basic health status and determine who may have additional health needs. We found that among Black residents, those that had 3+ comorbidities were more likely to have care disruptions. We have included this in the results and discussed implications in the discussion. We note that the sample size is small.

The sample size was small, and with the cross-sectional nature of the study, no conclusion can be achieved.

This was meant to be an area-level, local assessment to help Siteman Cancer Center understand how to deploy resources to mitigate the potential for health disparities. There were no causal interpretations.

The characteristics of people with missing data were not presented and discussed. This can easily create a biased sample to begin with. The variables used and the rationale to include them have not been discussed.

We checked for differential exclusion for those who were categorized as does not apply for care disruption. This sentence appears in the analytic procedures section: “We dropped those who reported that canceling an appointment did not apply to them (n=84) with more males, uninsured people, and those without telehealth appointments reflected in this exclusion.”

We also added the text below to describe the 17 people missing from the regression model. Given that this is less than 2% of our total sample, it is unlikely that this has introduced bias.

“Additionally, there were 17 residents who had missing data and were not included in the logistic regression models. This missing data was due to responding to sex assigned at birth or sexual orientation questions with “prefer not to answer” or having a missing value on another question included in the model (see footnote on Table 1 for details).”

Reviewer #2 (Recommendations for the authors):

I appreciate the opportunity to review your manuscript submitted to eLife. The manuscript addresses an important topic, and it will be a valuable contribution to the literature on the impact of COVID-19 on cancer care.

1. Outcome measures were adapted from an unpublished questionnaire by Penedo et al. Please provide more information about how these measures were developed and/or whether they were validated.

Since our article was submitted, the questionnaire has been published. The questions were drawn from validated measures assessing the impact of pandemics such as H1N1, and major life disruptions such as natural disasters. This language was updated in the manuscript as were the references.

Relevant Reference: Saez-Clarke, E., Otto, A. K., Prinsloo, S., Natori, A., Wagner, R. W., Gomez, T. I., … and Penedo, F. J. (2023). Development and initial psychometric evaluation of a COVID-related psychosocial experiences questionnaire for cancer survivors. Quality of Life Research, 1-20.

2. Figure 1 shows that 53 participants had appointments for cancer-related care other than screening. Because this finding is presented in Figure 1, it should be discussed in the results. A footnote in Supplemental Table 1 incorrectly states that the variable was not included in the manuscript.

We renamed this Supplementary file 1 as directed by eLife editors. We also added the language below to the Results section.

Approximately 53 people in the sample reported having cancer. Among those, 26% reported having to cancel/postpone their cancer-related care. We also corrected the table to add that this is included in the paper, but not included in the outcome measure.

3. The analysis should address the assumptions for conducting logistic regression. Specifically, authors should test for (1) linearity between continuous independent variables and the log-odds of care disruption, and (2) absence of multicollinearity.

We tested for multicollinearity by calculating variance inflation factors. All values were less than 2, indicating no evidence of multicollinearity. Additionally, a correlation matrix of all intendent variables in the model showed no values over 0.8 (indicative of a high correlation and potential for multicollinearity); all correlations were under 0.5.

In addition, the authors treated education and income as continuous variables -- please explain this decision.

We have 5 education categories and 8 income categories. We decided to model them as continuous given the multiple categories and to increase power.

4. Cancer care disruptions were stratified by race. Did the "Patient needs for rescheduling" likewise vary based on race or other factors?

In a supplemental analysis, we stratified patient care needs by race. We have added the test below to our results. Results were mostly similar.

Results stayed similar when looking at the highest need for rescheduling by race, except that the third highest need for Black residents was time to schedule the appointment (14.2%), and for White residents, it was dealing with other things and not ready to reschedule yet (14.5%). The top two needs to reschedule appointments were the same across race though the proportion of those who selected each need varied somewhat by race. The top two needs were knowing if their doctor’s office or clinic is taking the appropriate COVID-related safety precautions (Non-Hispanic Black: 34.2%; Non-Hispanic White: 33.8%; All Other: 32.7%) and not needing anything for all three racial groups (Non-Hispanic Black: 19.2%; Non-Hispanic White: 18.8%; All Other: 12.2%).

5. Please consider adding a supplemental table to show the univariate characteristics of Non-Hispanic Black or African American and Non-Hispanic White subgroups, which are stratified in the multivariate analysis in Table 2. The analysis was stratified by race because of the differential socioeconomic factors and COVID-19 impacts among communities of color -- it would be helpful to see the extent to which the combination of factors is reflected in the sample.

Thank you for this suggestion. In the newly added Supplementary file 2, we show the variables of interest by race strata.

Author details

1. Kia L Davis

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Formal analysis, Supervision, Investigation, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   DavisKL@wustl.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1338-3018

2. Nicole Ackermann

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7411-3233

3. Lisa M Klesges

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Writing – review and editing

   Competing interests

   has received consulting fees from Dana Farber Cancer Institute. The author is on the Board of Directors for American College of Epidemiology and Neighborhood Preservation, Inc. The author has no other competing interests to declare

4. Nora Leahy

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Investigation, Project administration, Writing – review and editing

   Competing interests

   No competing interests declared

5. Callie Walsh-Bailey

   Brown School, Washington University in St. Louis, St Louis, United States

   Contribution

   Writing – review and editing

   Competing interests

   No competing interests declared

6. Sarah Humble

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0694-091X

7. Bettina Drake

   Department of Surgery, Public Health Sciences, School of Medicine, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

8. Vetta L Sanders Thompson

   Brown School, Washington University in St. Louis, St Louis, United States

   Contribution

   Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review and editing

   Competing interests

   has received consulting fees from Novaris and Chan-Zuckerburg Initiave via National Academies of Science. The author has received payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events from St. Jude Hospital, Scholar Strategies, University of Missouri St. Louis, Ohio State University Health Services and Management Program Management Institute Annual Conference, and Nebraska Conference on Health Equity Key Note. The author is a Board Director, Vice Chair and Programmatic Strategies Chairperson for Missouri Foundation for Health, and the author receives no financial compensation for these roles. The author has no other competing interests to declare

• 74

  Page views

• 19

  Downloads

• 0

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.