Abstract
Background:
Individuals with a history of cancer may be more susceptible to severe COVID-19 due to immunosuppression, comorbidities, or ongoing treatment. We linked inpatient claims data on COVID-19 hospitalizations to cancer diagnoses from the New York State Cancer Registry (NYSCR) to examine associations between prior cancer diagnoses and hospitalizations for COVID-19, and factors associated with death at discharge after COVID-19 hospitalization.
Methods:
New York State (NYS) residents diagnosed with invasive cancer before July 1, 2021, who were alive on January 1, 2020, were identified from NYSCR data. We obtained claims data for discharge year 2020 and the first half of 2021 from NYS's Statewide Planning and Research Cooperative System (SPARCS), and we linked inpatient records with COVID-19 as the primary diagnosis to cancer data from the NYSCR using deterministic matching methods. We calculated descriptive statistics and conducted multivariable-adjusted logistic regression analyses to examine associations of cancer case characteristics with COVID-19 hospitalization and with vital status at discharge among patients with a history of cancer. All analyses were conducted in SAS 9.4.
Results:
Our analysis included 1,257,377 individuals with a history of cancer, 10,210 of whom had a subsequent primary COVID-19 hospitalization. Individuals with a history of cancer were 16% more likely to be hospitalized with COVID-19, compared to the general population of NYS, after adjusting for age and sex (95% CI, 14%-19%). Factors independently associated with COVID-19 hospitalization among cancer patients included older age, male sex, non-Hispanic Black race or Hispanic ethnicity, diagnosis with late-stage cancer or with multiple tumors, more recent cancer diagnosis, and New York City (NYC) residency at the time of cancer diagnosis. Factors independently associated with death at discharge among individuals with COVID-19 hospitalization and a prior cancer diagnosis included older age, male sex, non-Hispanic Black or non-Hispanic Asian/Pacific Islander race or Hispanic ethnicity, residence in NYC at the time of COVID-19 hospitalization, and an active cancer diagnosis claim code at the time of COVID-19 hospitalization.
Conclusion:
This claims-based study identified higher risks of COVID-19 hospitalization and death at discharge among individuals with a history of cancer, and particularly those in certain demographic and diagnostic groups.
Keywords: Cancer, claims, COVID-19, hospitalization, New York State
Introduction
Cancer is a major public health issue worldwide. In the United States, there were an estimated 18.1 million individuals with any history of cancer (excluding basal cell or squamous cell skin cancers and in situ cancers other than urinary bladder) as of January 1, 2022, representing approximately 5.4% of the total US population.1-2 Individuals with a history of cancer may be more susceptible to severe COVID-19 due to immunosuppression, comorbidities, or ongoing treatment. Cancer is one of several underlying medical conditions that is considered to be conclusively associated with higher risk for severe COVID-19, based on a review of the literature by the Centers for Disease Control and Prevention (CDC).3-4
However, previous studies investigating the relationship between past cancer diagnoses and severe COVID-19 outcomes, including hospitalization and death, have had mixed results. Several studies observed that patients with a cancer history had a significantly higher risk for severe COVID-19 outcomes,5-12 while others did not13-14 or only found that patients with a recent cancer diagnosis or those who had received recent cancer treatment were at a higher risk of severe outcomes of COVID-19.15-16
The New York State (NYS) Statewide Planning and Research Cooperative System (SPARCS) is a comprehensive all-payer database that captures patient-level information for all encounters from hospital inpatient and outpatient visits, making it a valuable resource for data on COVID-19 hospitalizations in NYS.17 We conducted a comprehensive analysis of early COVID-19 hospitalizations among NYS residents with a history of cancer to examine in detail associations between cancer history and severe COVID-19 outcomes. We linked claims data on COVID-19 hospi-talizations from SPARCS to cancer diagnoses from the New York State Cancer Registry (NYSCR) to assess which patient demographics, tumor characteristics, and cancer types were associated with an elevated risk for COVID-19 hospitalization. We then examined associations between prior cancer diagnosis and death at discharge among all COVID-19 hospitalizations obtained from SPARCS, as well as the demographics, tumor characteristics, and cancer types that were associated with an increased risk of death at discharge.
Methods
Data Sources and Case Selection
We retrieved data for 1,262,264 patients with a history of cancer (including 142,114 with more than 1 tumor) from the NYSCR Surveillance, Epidemiology, and End Results Data Management System (SEER*DMS) database for NYS residents who were diagnosed with invasive cancer before July 1, 2021, and who were alive on January 1, 2020. We then obtained claims data from SPARCS, which includes patient-level data on diagnoses from hospital inpatient and outpatient (ambulatory, emergency department, and outpatient services) visits. We retained data for inpatient claims only for NYS residents with discharge dates in 2020 and the first half of 2021 to allow for identification of hospitalizations related to COVID-19. We retrieved a total of 2,977,403 inpatient claims records meeting these inclusion criteria.
Since there is no direct identifier in the SPARCS data, we defined an individual using the date of birth (DOB), sex, and a unique personal identifier (UPI) variable, which is a combination of the first 2 and last 2 letters of the last name, the first 2 letters of the first name, and the last 4 digits of the Social Security number (SSN). When more than 1 claims record had the same UPI, DOB, and sex, we considered them to be the same patient. If the SSN component of UPI was missing, the claims were considered to be from the same patient if they had the same first 6 characters of the UPI and the same DOB, sex, and either patient zip code or both treating facility and medical record number (MRN). From the 2,977,403 inpatient claims records obtained from SPARCS, we identified a total of 2,041,781 unique patients.
Linkage Between Cancer Data and SPARCS Claims Records
We linked SPARCS inpatient records and cancer data using deterministic matching methods by comparing UPI, DOB, sex, reporting facility identifier (PFI), MRN, and patient zip code at diagnosis. The linkage process included 9 sequential steps, followed by manual review to resolve duplicate matches. The 9 steps included linkage of records with: (1) same UPI, DOB, and sex; (2) same UPI, DOB, PFI, and MRN; (3) same UPI, sex, PFI, and MRN; (4) same UPI, DOB, and zip code; (5) same UPI and DOB; (6) same UPI, sex, and either same birth year or same birth month and day; (7) same UPI without SSN, plus same DOB, sex, PFI and MRN; (8) same UPI without SSN, plus same PFI and MRN; and (9) same UPI without SSN, plus same DOB, sex, and zip code.
Of the 1,262,264 cancers retrieved from SEER*DMS, we excluded 4,715 with unknown or implausible age (defined as >110 years), where age was calculated as the difference between the patient's date of birth and either the earliest admission date from SPARCS or, for unlinked cases, the date of the midpoint of the study period (September 1, 2020). We additionally excluded 172 cases with nonmale/ nonfemale sex, due to the small number of these cases. After these exclusions, a total of 1,257,377 cancers were included in the study. Of these, 251,304 (20.0%) matched with 1 or more SPARCS inpatient claims records, among which we identified 30 duplicate matches where 2 cancer cases matched to the same inpatient claims record(s). After manual review, we removed 1 case from each duplicate. We also removed 3,878 ineligible matches because their date of cancer diagnosis was later than the date of admission from inpatient claims records, indicating that they had not been diagnosed with cancer prior to the time of their hospitalization. The remaining 247,398 (19.7%) were considered to be good matches, while 1,009,979 cancer cases (80.3%) did not have a documented hospitalization during the time frame of interest.
Identification of COVID-19 Hospitalizations
We identified COVID-19 hospitalizations using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) primary diagnosis code U07.1 for discharge year 2020 and the first half of 2021 and code B97.29 for discharge dates from January 1 to March 31, 2020.18 When a COVID-19 diagnosis code was the principal diagnosis on 1 or more linked inpatient claims records for a patient, it was counted as 1 COVID-19 hospitalization, and only the record with the earliest admission date was included in the analysis. Among the 2,041,781 patients with a hospitalization during the time frame of interest, we identified COVID-19 as a principal diagnosis code for 77,338 cases. Four of these cases had unknown sex and were excluded, leaving a total of 77,334 COVID-19 hospitalizations for analysis, including 10,210 COVID-19 hospitalizations among linked cancer cases.
Statistical Analysis
We calculated descriptive statistics for cancer patients with (vs without) COVID-19 hospitalization for the following demographic and tumor characteristics: age, sex, race/ethnicity, region of residency (New York City [NYC] or the rest of NYS, categorized based on address at cancer diagnosis from SEER*DMS), stage, history of multiple tumors, and time since cancer diagnosis. We used χ2 tests to examine statistically significant differences in the distribution of each covariate by COVID-19 hospitalization status and student's t test to examine differences in mean age by COVID-19 hospitalization.
We conducted multivariable-adjusted logistic regression analyses to calculate odds ratios (OR) and 95% CIs for associations between the above-mentioned variables and COVID-19 hospitalization among patients with a history of cancer. All variables examined were significantly associated with COVID-19 hospitalization and were adjusted for in the final model. For patients with a history of 2 or more invasive tumors, we included only the most recent invasive tumor in the analysis, based on the assumption that the more recently diagnosed tumor would have had a greater impact on the patient's recent health.
Next, we examined COVID-19 hospitalization by cancer type to determine if patients with a history of certain cancers were more likely to be hospitalized with COVID-19. We categorized cancer diagnoses into 24 cancer types: oral cavity and pharynx; esophagus; stomach; colorectal; liver and intrahepatic bile duct; pancreas; larynx; lung and bronchus; melanoma; breast; cervix uteri; corpus uterus and not otherwise specified (NOS); ovary; prostate; testis; urinary bladder (including in situ); kidney and renal pelvis; brain and other nervous system; thyroid; Hodgkin lymphoma; non-Hodgkin lymphomas; myeloma; leukemias; and other malignancies of hematopoietic or lymphopoietic origin. We included in situ urinary bladder cancers in the analysis based on the SEER rules for determining multiple primary cancers and for calculating incidence rates, which specify that in situ bladder cancers are counted along with invasive cancers when reporting bladder cancer incidence (and for no other type of cancer).
We calculated the ratio of observed to expected (O/E) COVID-19 hospitalizations among cancer cases overall and by cancer type. We estimated the expected counts of COVID-19 hospitalizations using age- and sex-specific rates of COVID-19 hospitalization among NYS residents, which were calculated by dividing age- and sex-specific counts of hospitalizations in NYS from SPARCS by the corresponding age- and sex-specific population counts for NYS from the 2019 American Community Survey 1-year population estimates. We then calculated the expected number of COVID-19 hospitalizations by cancer type by applying these age- and sex-specific proportions of COVID-19 hospi-talization for all cancers combined to the observed number of cancers by age and sex for each individual cancer type. We used 18 age groups (0-4, 5-9, 10-14, 15-19, 20-24, 25-29, 30-34, 35-39, 40-44, 45-49, 50-54, 55-59, 60-64, 65-69, 70-74, 75-79, 80-84, and >85 years) and 2 categories of sex (male and female) for these calculations. We calculated 95% CIs for the ratios of observed to expected using the Byar's approximation of the exact Poisson distribution.19
Next, we restricted our analysis to all COVID-19 hospitalizations (n = 77,334) retrieved from SPARCS, and we examined vital status at discharge among patients with versus without a history of cancer. In addition, among the patients with a history of cancer, we examined differences in vital status at discharge by the presence or absence of a cancer diagnosis claim code in the SPARCS claim record, where patients with a cancer diagnosis claim code were considered active cancer cases and those without a cancer diagnosis claim code were considered inactive cancer cases. We calculated the crude and age- and sex-adjusted proportions of death for each of these groups, as well as the 95% CIs, using the SAS STDRATE procedure. The age- and sex-adjusted proportions of death by cancer status were calculated using the age- and sex-specific proportions obtained from all 77,334 COVID-19 hospitalization patients as a reference, and the age- and sex-adjusted proportions of death by active status of cancer diagnosis were calculated using the age- and sex-specific proportions obtained from the 10,210 COVID-19 hospitalized patients with a history of cancer as a reference. We used 11 age groups (0-39, 40-44, 45-49, 50-54, 55-59, 60-64, 65-69, 70-74, 75-79, 80-84, and >85 years) to adjust for age in these age- and sex-adjusted calculations.
For the 10,210 patients with COVID-19 hospitalization and a history of cancer, we calculated descriptive statistics by vital status at discharge for the following demographic and tumor characteristics: age, sex, race/ethnicity, region of residency at COVID-19 hospitalization, stage, history of multiple tumors, time since cancer diagnosis, and active or inactive cancer diagnosis based on the presence or absence of a cancer claim code at the time of COVID-19 hospitaliza-tion. We used χ2 tests to examine statistically significant differences in the distribution of each covariate by vital status at discharge from hospitalization for COVID-19 and student's t test to examine differences in mean age by vital status at discharge. We conducted multivariable-adjusted logistic regression analyses to calculate ORs and 95% CIs for associations between each variable of interest and vital status at discharge. All variables of interest were included as covariates in the final multivariable-adjusted model.
Finally, we used methods similar to those described above to calculate the ratios and 95% CIs of observed to expected counts of deaths at discharge by cancer type among individuals with COVID-19 hospitalization and a history of cancer (n = 10,210). We calculated the expected numbers of deaths at discharge by using age- and sex-specific proportions of death for all 10,210 patients with a history of cancer, based on the data retrieved from SPARCS. We then calculated the expected number of deaths for each cancer type by applying these age- and sex-specific proportions of death at discharge to the observed number of COVID-19 hospitalizations for each cancer type. We calculated 95% CIs for the ratios of observed to expected deaths at discharge based on the Byar's approximation, as described above. All analyses were performed using SAS 9.4.
Results
The overall ratio of observed versus expected COVID-19 hospitalizations among NYS residents with a history of cancer was 1.16 and the 95% CI was 1.14 to 1.19 (results not shown), indicating that individuals with a prior diagnosis of cancer were 16% more likely to be hospitalized with COVID-19 compared to the general population of NYS, after adjusting for age and sex.
Table 1 presents the distribution of demographic and tumor characteristics of interest for individuals with a history of cancer by COVID-19 hospitalization status, as well as multivariable-adjusted ORs and 95% CIs for each variable of interest in relation to COVID-19 hospitalization status. Among individuals with a history of cancer, those with versus without COVID-19 hospitalization tended to be older (mean age of 74.0 vs 68.2 years) and were more likely to be male (55.1% vs 45.9%), non-Hispanic Black (18.9% vs 12.2%), Hispanic (14.3% vs 10.1%), reside in NYC at the time of their cancer diagnosis (45.6% vs 37.7%), and to have been diagnosed with distant-stage or systemic cancer (19.4% vs 11.3%), multiple tumors (16.6% vs 11.2%), and within the past 0 to 2 years (18.0% vs 14.0%) or 3 to 5 years (22.4% vs 19.3%) (all P < .0001).
Table 1.
Distribution of Demographic and Tumor Characteristics of Interest and Multivariable-Adjusted Odds Ratios (ORs) and 95% CIs for Each Variable of Interest and COVID-19 Hospitalization Status Among 1,257,377 New York State Residents with a History of Cancer
| Variable | Cancer cases with COVID-19 hospitalization (n = 10,210), n (%) | Cancer cases without COVID-19 hospitalization (n = 1,247,167), n (%) | P value* | OR (95% CI)** | OR (95% CI)*** |
|---|---|---|---|---|---|
| Age in years, mean (SD) | 74.0 (12.9) | 68.2 (15.2) | <.0001 | ||
| Age group (y) | <.0001 | ||||
| 0-19 | 30 (0.3) | 8,340 (0.7) | 0.71 (0.49-1.04) | 0.71 (0.48-1.03) | |
| 20-29 | 38 (0.4) | 14,798 (1.2) | 0.63 (0.45-0.88) | 0.62 (0.44-0.88) | |
| 30-39 | 71 (0.7) | 35,372 (2.8) | 0.52 (0.40-0.68) | 0.52 (0.40-0.68) | |
| 40-49 | 267 (2.6) | 70,967 (5.7) | Ref | Ref | |
| 50-59 | 841 (8.2) | 174,431 (14.0) | 1.28 (1.12-1.47) | 1.27 (1.11-1.46) | |
| 60-69 | 2,041 (20.0) | 314,699 (25.2) | 1.73 (1.52-1.96) | 1.72 (1.51-1.95) | |
| 70-79 | 3,249 (31.8) | 348,433 (27.9) | 2.59 (2.28-2.93) | 2.58 (2.28-2.93) | |
| >80 | 3,673 (36.0) | 280,127 (22.5) | 4.00 (3.53-4.54) | 4.01 (3.54-4.55) | |
| Sex | <.0001 | ||||
| Male | 5,627 (55.1) | 572,339 (45.9) | 1.35 (1.29-1.40) | 1.34 (1.29-1.40) | |
| Female | 4,583 (44.9) | 674,828 (54.1) | Ref | Ref | |
| Race/ethnicity | <.0001 | ||||
| Non-Hispanic White | 6,276 (61.5) | 882,587 (70.8) | Ref | Ref | |
| Non-Hispanic Black | 1,930 (18.9) | 152,041 (12.2) | 1.75 (1.66-1.85) | 1.91 (1.81-2.01) | |
| Non-Hispanic Asian/Pacific Islander | 454 (4.4) | 66,127 (5.3) | 0.99 (0.89-1.09) | 1.09 (0.99-1.20) | |
| Hispanic | 1,457 (14.3) | 126,327 (10.1) | 1.66 (1.56-1.77) | 1.83 (1.73-1.94) | |
| Non-Hispanic other/unknown | 93 (0.9) | 20,085 (1.6) | 0.75 (0.61-0.93) | 0.80 (0.65-0.98) | |
| Region of residency at cancer diagnosis | <.0001 | ||||
| New York City | 4,655 (45.6) | 469,750 (37.7) | 1.21 (1.16-1.27) | NA | |
| Rest of New York State | 5,555 (54.4) | 777,417 (62.3) | Ref | Ref | |
| Stage of Cancer | <.0001 | ||||
| Local | 4,926 (48.2) | 626,01 1 (50.2) | Ref | Ref | |
| Regional | 1,628 (15.9) | 205,539 (16.5) | 1.09 (1.03-1.15) | 1.09 (1.03-1.15) | |
| Distant | 1,983 (19.4) | 141,262 (1 1.3) | 1.69 (1.61-1.79) | 1.67 (1.61-1.79) | |
| Unknown stage | 1,673 (16.4) | 274,355 (22.0) | 0.79 (0.75-0.84) | 0.80 (0.75-0.84) | |
| Multiple Tumors | <.0001 | ||||
| No | 8,512 (83.4) | 1,106,958 (88.8) | Ref | Ref | |
| Yes | 1,698 (16.6) | 140,209 (11.2) | 1.23 (1.16-1.30) | 1.23 (1.16-1.29) | |
| Time since cancer diagnosis | <.0001 | ||||
| 0-2 years (2019-2021) | 1,840 (18.0) | 174,494 (14.0) | 1.53 (1.44-1.62) | 1.51 (1.42-1.61) | |
| 3-5 years (2016-2019) | 2,285 (22.4) | 240,208 (19.3) | 1.42 (1.34-1.50) | 1.41 (1.33-1.49) | |
| 6-10 years (2012-2016) | 2,404 (23.5) | 284,264 (22.8) | 1.25 (1.19-1.32) | 1.25 (1.18-1 .32) | |
| >10 years (1976-201 1) | 3,681 (36.1) | 548,201 (44.0) | Ref | Ref | |
P-values from χ2 tests for categorical variables and student's t test for mean age.
Analyses mutually adjusted for all variables in the table.
Analyses mutually adjusted for all variables in the table with the exception of region of residence at cancer diagnosis.
In multivariable-adjusted logistic regression analyses, we observed lower odds of COVID-19 hospitalization among younger age groups and increased odds among older age groups, with the highest odds among individuals ages 60 years and older. Compared to individuals aged 40-49 years, the ORs (95% CIs) were 1.73 (1.52-1.96) for individuals ages 60-69 years, 2.59 (2.28-2.93) for individuals ages 70-79 years, and 4.00 (3.53-4.54) for individuals 80 years of age and older. In addition, we observed increased odds for males versus females (OR, 1.35; 95% CI, 1.29-1.40) and for non-Hispanic Black (OR, 1.75; 95% CI, 1.66-1.85) and Hispanic (OR, 1.66; 95% CI, 1.56-1.77) versus non-Hispanic White individuals. Risk was also increased for individuals who resided in NYC at the time of cancer diagnosis versus the rest of NYS (OR, 1.21; 95% CI, 1.16-1.27); individuals with regional stage (OR, 1.09; 95% CI, 1.03-1.15) or distant stage (OR, 1.69; 95% CI, 1.61-1.79) versus local stage disease; and individuals with multiple cancers versus a single cancer (OR, 1.23; 95% CI, 1.16-1.30). Individuals with more recent cancer diagnosis had higher risk, with increasing odds for decreasing time since cancer diagnosis. Compared to individuals whose most recent cancer was diagnosed more than 10 years in the past, the ORs were 1.53 for cancers diagnosed in the past 0-2 years (95% CI, 1.44-1.62); 1.42 for 3-5 years (95% CI, 1.34-1.50); and 1.25 for 6-10 years (95% CI, 1.19-1.32).
Since region of residency (NYC or rest of NYS) was determined based on address at cancer diagnosis and not address at the time of the COVID-19 pandemic, due to its unavailability for individuals without a linked record in the SPARCS data, we also considered models that did not adjust for region of residence. We obtained very similar results (displayed in column 6 of Table 1) to those obtained from the analysis adjusted for region of residency for all covariates except race/ethnicity. After removing region of residency from the model, the associations with COVID-19 hospitalization were strengthened for non-Hispanic Black (OR, 1.91; 95% CI, 1.81-2.01) and Hispanic individuals (OR, 1.83; 95% CI, 1.73-1.94) compared to non-Hispanic White individuals.
Table 2 and Figure 1 display the observed versus expected counts of COVID-19 hospitalizations among individuals with a history of cancer by cancer type, where the expected numbers were determined based on age- and sex-adjusted proportions for all cancers combined. We observed a statistically significant higher than expected number of COVID-19 hospitalizations for several cancers including kidney and renal pelvis cancer (O/E, 1.29; 95% CI, 1.171.41), leukemias (O/E, 1.91; 95% CI, 1.74-2.09), liver and intrahepatic bile duct cancer (O/E, 1.44; 95% CI, 1.19-1.72), lung and bronchus cancer (O/E, 1.54; 95% CI, 1.44-1.65), myeloma (O/E, 2.21; 95% CI, 1.95-2.49), non-Hodgkin lymphomas (O/E, 1.38; 95% CI, 1.27-1.50), and other malignancies of hematopoietic or lymphopoietic origin (O/E, 1.46; 95% CI, 1.29-1.66). We observed a statistically significant lower than expected number of COVID-19 hospitalizations for breast cancer (O/E, 0.79; 95% CI, 0.75-0.83), cervix uteri cancer (O/E, 0.73; 95% CI, 0.57-0.91), Hodgkin lymphoma (O/E, 0.74; 95% CI, 0.56-0.97), melanoma (O/E, 0.60; 95% CI, 0.53-0.67), oral cavity and pharynx cancer (O/E, 0.84; 95% CI, 0.72-0.98), prostate cancer (O/E, 0.86; 95% CI, 0.83-0.90), testis cancer (O/E, 0.43; 95% CI, 0.30-0.59), and thyroid cancer (O/E, 0.70; 95% CI, 0.62-0.79).
Table 2.
Ratio and 95% CI of Observed to Expected Number of COVID-19 Hospitalizations Among Individuals with a History of Cancer by Cancer Type
| Cancer type | Number of observed COVID-19 hospitalizations | Number of expected COVID-19 hospitalizations* | Ratio of observed to expected (95% CI) |
|---|---|---|---|
| Brain and other nervous system | 58 | 70 | 0.83 (0.63-1.07) |
| Breast | 1,427 | 1,816 | 0.79 (0.75-0.83) |
| Cervix uteri | 74 | 102 | 0.73 (0.57-0.91) |
| Colorectal | 889 | 836 | 1.06 (0.99-1.14) |
| Corpus uterus and NOS | 408 | 402 | 1.02 (0.92-1.12) |
| Esophagus | 46 | 44 | 1.05 (0.77-1.41) |
| Hodgkin lymphoma | 54 | 73 | 0.74 (0.56-0.97) |
| Kidney and renal pelvis | 437 | 339 | 1.29 (1.17-1.41) |
| Larynx | 71 | 71 | 1.01 (0.79-1.27) |
| Leukemias | 491 | 257 | 1.91 (1.74-2.09) |
| Liver and intrahepatic bile duct | 115 | 80 | 1.44 (1.19-1.72) |
| Lung and bronchus | 852 | 552 | 1.54 (1.44-1.65) |
| Melanoma | 268 | 448 | 0.60 (0.53-0.67) |
| Multiple myeloma | 268 | 121 | 2.21 (1.95-2.49) |
| Non-Hodgkin lymphomas | 586 | 424 | 1.38 (1.27-1.50) |
| Oral cavity and pharynx | 172 | 204 | 0.84 (0.72-0.98) |
| Other malignancies of hematopoietic or lymphopoietic origin | 243 | 166 | 1.46 (1.29-1.66) |
| Ovary | 94 | 102 | 0.93 (0.75-1.13) |
| Pancreas | 100 | 98 | 1.02 (0.83-1.25) |
| Prostate | 2,070 | 2,396 | 0.86 (0.83-0.90) |
| Stomach | 127 | 122 | 1.04 (0.87-1.24) |
| Testis | 36 | 84 | 0.43 (0.30-0.59) |
| Thyroid | 257 | 367 | 0.70 (0.62-0.79) |
| Urinary bladder, including in situ | 550 | 549 | 1.00 (0.92-1.09) |
NOS, not otherwise specified. *Number of expected COVID-19 hospitalizations was calculated using the age- and sex-specific numbers of hospitalizations for all cancers combined.
Figure 1.
Ratio and 95% CI of Observed to Expected Number of COVID-19 Hospitalizations Among Individuals with a History of Cancer by Cancer Type
Table 3 shows the age- and sex-adjusted proportions of death at discharge among all COVID-19 hospitalizations retrieved from SPARCS, by cancer status. The adjusted percentage of individuals with death at discharge was higher among individuals with a prior cancer diagnosis (17.6%) compared to those with no prior cancer diagnosis (15.5%) (P < .0001). Among the individuals with a prior cancer diagnosis, we observed a higher adjusted percentage with death at discharge among individuals with a cancer diagnosis claim code at the time of COVID-19 hospitaliza-tion (27.1%) compared to those with no cancer diagnosis claim code (20.8%) (P < .0001).
Table 3.
Crude and Age- and Sex-Adjusted Proportions of Death at Discharge Among 77,334 COVID-19 Hospitalizations for NYS Residents Retrieved from SPARCS, by Cancer Status, and Among 10,210 COVID-19 Hospitalizations Among NYS Residents with a History of Cancer, by Active Status of Cancer
| Cancer status | Deceased at discharge (n) | Total N | Crude proportion* | Adjusted proportion** | SE | 95% CI | |
|---|---|---|---|---|---|---|---|
| No prior cancer diagnosis | 9,835 | 67,124 | 14.65 | 15.54 | 0.16 | 15.24 | 15.85 |
| Prior cancer diagnosis | 2,328 | 10,210 | 22.80 | 17.62 | 0.46 | 16.73 | 18.52 |
| Inactive cases (no cancer diagnosis claim code) | 1,535 | 7,077 | 21.69 | 20.75 | 0.53 | 19.70 | 21.79 |
| Active cases (with cancer diagnosis claim code) | 793 | 3,133 | 25.31 | 27.10 | 0.99 | 25.16 | 29.05 |
NYS, New York State; SPARCS, Statewide Planning and Research Cooperative System.
χ2 P < .0001 for differences between crude proportions by cancer status and by active status of cancer.
Adjusted proportions were calculated using SAS STDRATE with the direct standardization method. Age- and sex-specific proportions of deaths for 77,334 COVID-19 hospitalizations were used as the reference in the calculation of adjusted proportions by cancer status, and age- and sex-specific proportions of deaths for 10,210 COVID-19 hospitalizations with prior cancer were used as the reference in the calculation of adjusted proportions by active status of cancer. χ2 P < .0001 comparing individuals with versus without prior diagnosis of cancer and comparing active and inactive cancer diagnoses.
Table 4 displays the distribution of different demographic and tumor characteristics of interest and multivariable-adjusted ORs and 95% CIs for each variable in relation to death at discharge among 10,210 individuals with COVID-19 hospitalization and a prior cancer diagnosis. Comparing individuals who were deceased at discharge with those who were not, there were higher percentages of individuals aged 80 years and over, individuals who were male, non-Hispanic Black, non-Hispanic Asian/Pacific Islander (API), Hispanic, living in NYC, diagnosed with distant-stage or unknown-stage cancer or with multiple tumors, or who had active cancer diagnosis claim codes (all P < .01).
Table 4.
Distribution of Demographic Variables and Tumor Characteristics of Interest and Multivariable-Adjusted ORs and 95% CIs for Each Variable and Vital Status at Discharge Among 10,210 Individuals with COVID-19 Hospitalization and Prior Cancer Diagnosis
| Variable | Cancer cases deceased at discharge from COVID-19 hospitalization (n = 2,328), n (%) | Cancer cases alive at discharge from COVID-19 hospitalization (n = 7,882), n (%) | P value* | OR (95% CI)** |
|---|---|---|---|---|
| Age in years, mean (SD) | 77.9 (10.9) | 72.8 (13.3) | <.0001 | |
| Age group (y) | <.0001 | |||
| 0-19 | <11 | 30 (0.4) | NA | |
| 20-29 | <11 | 37 (0.5) | 0.24 (0.03-1.81) | |
| 30-39 | <11 | 69 (0.9) | 0.26 (0.06-1.13) | |
| 40-49 | 26 (1.1) | 241 (3.1) | Ref | |
| 50-59 | 99 (4.3) | 742 (9.4) | 1.29 (0.82-2.04) | |
| 60-69 | 365 (15.7) | 1,676 (21.3) | 2.16 (1.41-3.29) | |
| 70-79 | 726 (31.2) | 2,523 (32.0) | 3.02 (1.99-4.58) | |
| >80 | 1,109 (47.6) | 2,564 (32.5) | 4.85 (3.19-7.38) | |
| <.0001 | ||||
| Male | 1,415 (60.8) | 4,212 (53.4) | 1.33 (1.20-1.46) | |
| Female | 913 (39.2) | 3,670 (46.6) | Ref | |
| Race/ethnicity | 0.001 | |||
| Non-Hispanic White | 1,352 (58.1) | 4,924 (62.5) | Ref | |
| Non-Hispanic Black | 482 (20.7) | 1,448 (18.4) | 1.32 (1.16-1.51) | |
| Non-Hispanic API | 121 (5.2) | 333 (4.2) | 1.39 (1.1 1-1.75) | |
| Hispanic | 356 (15.3) | 1,101 (14.0) | 1.28 (1.11-1.49) | |
| Non-Hispanic other/unknown | 17 (0.7) | 76 (1.0) | 0.85 (0.50-1.47) | |
| Region of residency*** | <.0001 | |||
| New York City | 1,159 (49.8) | 3,359 (42.6) | 1.26 (1.13-1.40) | |
| Rest of New York State | 1,169 (50.2) | 4,523 (57.4) | Ref | |
| Stage of Cancer | 0.005 | |||
| Local | 1,072 (46.1) | 3,854 (48.9) | Ref | |
| Regional | 354 (15.2) | 1,274 (16.2) | 1.07 (0.92-1.23) | |
| Distant | 474 (20.4) | 1,509 (19.1) | 1.12 (0.98-1.30) | |
| Unknown stage | 428 (18.4) | 1,245 (15.8) | 1.12 (0.98-1.29) | |
| Multiple Tumor | 0.012 | |||
| No | 1,901 (81.7) | 6,611 (83.9) | Ref | |
| Yes | 427 (18.3) | 1,271 (16.1) | 1.03 (0.91-1.18) | |
| Time since cancer diagnosis | 0.003 | |||
| 0-2 years (2019-2021) | 411 (17.7) | 1,429 (18.1) | 0.96 (0.83-1.13) | |
| 3-5 years (2016-2019) | 521 (22.4) | 1,764 (22.4) | 1.02 (0.89-1.17) | |
| 6-10 years (2012-2016) | 492 (21.1) | 1,912 (24.3) | 0.87 (0.76-0.99) | |
| >10 years (1976-201 1) | 904 (38.8) | 2,777 (35.2) | Ref | |
| Active cancer diagnosis | <.0001 | |||
| Cancer diagnosis claim code | 793 (34.1) | 2,340 (29.7) | 1.39 (1.23-1.58) | |
| No cancer diagnosis claim code | 1,535 (65.9) | 5,542 (70.3) | Ref | |
P-values from χ2 tests for categorical variables and Student's t test for mean age.
Analyses mutually adjusted for all variables in the table.
Region of residency determined based on patient address at COVID-19 hospitalization from data obtained from SPARCS.
In multivariable-adjusted logistic regression analyses for the associations with death at discharge, we observed increased odds among older age groups, with increasing odds corresponding to increasing age. Compared to individuals 40-49 years of age, we observed ORs (95% CI) of 2.16 (1.41-3.29) for individuals aged 60-69 years, 3.02 (1.99-4.58) for those aged 70-79 years, and 4.85 (3.19-7.38) for those aged 80 years and older. The ORs for younger individuals were not statistically significant, and for individuals aged 0-19, the OR was not estimable due to the number of deaths in this age group. In addition, we observed increased odds of death at discharge for males versus females (OR, 1.33; 95% CI, 1.20-1.46); non-Hispanic Black (OR, 1.32; 95% CI, 1.16-1.51), non-Hispanic API (OR, 1.39; 95% CI, 1.11-1.75), and Hispanic individuals (OR, 1.28; 95% CI, 1.11-1.49) compared to non-Hispanic White individuals. In addition, odds of death at discharge were increased for individuals living in NYC at the time of COVID-19 hospitalization (OR, 1.26; 95% CI, 1.13-1.40) compared to the rest of NYS; and individuals with an active cancer diagnosis claim code (OR, 1.39; 95% CI, 1.23-1.58) compared to those with no cancer diagnosis claim code. We did not observe statistically significant associations with death at discharge for stage of cancer or presence of multiple tumors (all P > .05). For time since cancer diagnosis, there was a suggestion of decreased odds of death at discharge for individuals diagnosed 6-10 years in the past, compared to those diagnosed more than 10 years previously (OR, 0.87; 95% CI, 0.76-0.99), but, overall, the association with time since cancer diagnosis was not statistically significant (P = .11).
Table 5 and Figure 2 show the observed versus expected numbers of deaths at discharge for COVID-19 hospitaliza-tion among individuals with a history of cancer by cancer type, where the expected numbers of deaths were determined based on the age- and sex-specific counts for all 10,210 COVID-19 hospitalizations with a history of cancer. We did not observe a higher than expected number of deaths at discharge for any cancer type. However, we observed a lower than expected number of deaths at discharge for breast cancer (O/E, 0.73; 95% CI, 0.64-0.82), corpus uterus and NOS (O/E, 0.76; 95% CI, 0.60-0.96), melanoma (O/E, 0.64; 95% CI, 0.47-0.85), and thyroid cancer (O/E, 0.67; 95% CI, 0.47-0.93), after adjustment for age and sex.
Table 5.
Ratio and 95% CI of Observed to Expected Number of Deaths at Discharge Among 10,210 Individuals with COVID-19 Hospitalization and a History of Cancer, by Cancer Type
| Cancer type | Number of observed deaths | Number of expected deaths* | Ratio of observed to expected (95% CI) |
|---|---|---|---|
| Brain and other nervous system | <11 | Suppressed | 0.85 (0.39-1.62) |
| Breast | 267 | 366 | 0.73 (0.64-0.82) |
| Cervix uteri | 14 | 17 | 0.83 (0.45-1.40) |
| Colorectal | 207 | 233 | 0.89 (0.77-1.02) |
| Corpus uterus and NOS | 75 | 98 | 0.76 (0.60-0.96) |
| Esophagus | 12 | 10 | 1.14 (0.59-2.00) |
| Hodgkin lymphoma | 12 | 11 | 1.13 (0.58-1.97) |
| Kidney and renal pelvis | 98 | 105 | 0.94 (0.76-1.14) |
| Larynx | 15 | 18 | 0.85 (0.48-1.41) |
| Leukemias | 113 | 110 | 1.03 (0.85-1.24) |
| Liver and intrahepatic bile duct | 33 | 25 | 1.32 (0.91-1.86) |
| Lung and bronchus | 235 | 212 | 1.1 1 (0.97-1.26) |
| Melanoma | 46 | 72 | 0.64 (0.47-0.85) |
| Myeloma | 65 | 62 | 1.04 (0.80-1.33) |
| Non-Hodgkin lymphomas | 140 | 138 | 1.01 (0.85-1.19) |
| Oral cavity and pharynx | 36 | 41 | 0.88 (0.61-1.21) |
| Other malignancies of hematopoietic or lymphopoietic origin | 58 | 61 | 0.94 (0.72-1.22) |
| Ovary | 12 | 20 | 0.61 (0.32-1.07) |
| Pancreas | 24 | 22 | 1.07 (0.68-1.59) |
| Prostate | 532 | 552 | 0.96 (0.88-1.05) |
| Stomach | 27 | 31 | 0.88 (0.58-1.28) |
| Testis | <11 | Suppressed | 0.34 (0.04-1.24) |
| Thyroid | 37 | 55 | 0.67 (0.47-0.93) |
| Urinary bladder including in situ | 153 | 157 | 0.97 (0.83-1.14) |
NOS, not otherwise specified. * Number of expected COVID-19 hospitalizations was calculated using the age- and sex-specific proportions of deaths for all 10,210 cancers.
Figure 2.
Ratio and 95% CI of Observed to Expected Number of Deaths at Discharge Among 10,210 Individuals with COVID-19 Hospitalization and a History of Cancer by Cancer Type
Discussion
In this population-based analysis, we linked data on all NYS residents who had a history of invasive cancer and were alive immediately prior to the start of the COVID-19 pandemic with claims data on hospitalizations for COVID-19 in 2020 and the first half of 2021. We observed that individuals with a history of cancer were 16% more likely to be hospitalized for COVID-19, compared to the general population of NYS. Factors independently associated with COVID-19 hospitalization among cancer patients included older age, male sex, non-Hispanic Black race or Hispanic ethnicity, diagnosis with late-stage cancer or with multiple tumors, more recent cancer diagnosis, and NYC residency at the time of cancer diagnosis. In addition, we observed that individuals with a history of cancer were more likely to die while hospitalized for COVID-19, compared to those with no prior cancer diagnosis, and among individuals with a history of cancer, those with a cancer diagnosis claim code (indicating an active cancer diagnosis) were more likely to die than those without a cancer diagnosis claim code. Factors independently associated with death at discharge among individuals with a prior cancer diagnosis included older age, male sex, non-Hispanic Black or non-Hispanic API race or Hispanic ethnicity, living in NYC at the time of COVID-19 hospitalization, and having an active cancer diagnosis claim code.
The results of this study indicate that individuals with a history of cancer are at an increased risk for severe COVID-19 outcomes, including hospitalization and death, which is in agreement with findings from previous studies.5-12 Consistent with prior studies, we also observed that older age,5,11,12 Black race,7,20 diagnosis with late-stage cancer,16 and more recent cancer diagnosis7,16 were independently associated with severe COVID-19 among individuals with a history of cancer. However, to our knowledge, our findings that diagnosis with multiple tumors and residence in NYC were associated with COVID-19 hospitalization in individuals with a history of cancer have not been identified in previous studies. Although the association with residence in NYC was based on address at the time of cancer diagnosis, and not at the time of COVID-19 hospitalization, this variable likely approximated residence at the time of hospi-talization for a majority of patients. NYC was an epicenter of the COVID-19 pandemic and experienced a large number of COVID-19 cases, hospitalizations, and deaths early on in the pandemic, which likely contributed to this finding of an increased risk of hospitalization compared to residents of the rest of NYS.21,22 We were unable to assess differences in the total number of COVID-19 cases among cancer patients by variables such as region of NYS, race/ethnicity, and sex. However, it is likely that at least part of the difference in the risk of COVID-19 hospitalization by demographic characteristics was related to differences in the overall number of COVID-19 cases across population groups.21,22 A higher risk of severe COVID-19 in certain demographic groups, regardless of cancer status, may have also contributed to the results.22,23
In adjusted analyses of the observed versus expected number of hospitalizations by cancer type, we observed a higher than expected number of COVID-19 hospitalizations for kidney and renal pelvis cancer, leukemias, liver and intrahepatic bile duct cancer, lung and bronchus cancer, multiple myeloma, non-Hodgkin lymphomas, and other malignancies of hematopoietic or lymphopoietic origin. Previous studies have reported similar results for liver,16 lung,16 and hematological malignancies,16,24 but to our knowledge no prior studies have examined as many as 24 cancer types, and we observed associations with certain cancer types that were not previously reported. We did not have data on the presence of comorbidities, including HIV, that are more common in patients with certain cancers and that may increase the risk of hospitalization after diagnosis with COVID-19.4,16,25 It is possible that some of the increased risk of COVID-19 hospitalization for certain cancer types was related to the presence of HIV, which is associated with increased risk of several cancers including non-Hodgkin lymphoma, lung cancer, and liver cancer,26 or other comorbidities.
We observed that older age, male sex, non-Hispanic Black or non-Hispanic API race or Hispanic ethnicity, living in NYC at the time of COVID-19 hospitalization, and having an active cancer diagnosis claim code were associated with higher risk of death at discharge among 10,210 COVID-19 hospitalizations of patients with a history of cancer. Our findings were consistent with other studies that observed increased mortality with older age,10,11,12,16 male sex,10,11 and Black race.16,20 Our analyses did not indicate that diagnosis with late-stage cancer or multiple tumors or more recent cancer diagnosis was associated with higher risk of death at discharge. One previous study reported that the presence of multisite tumors was associated with increased risk of all-cause mortality, although the paper noted that this included any patient where more than 1 cancer site was reported and that cases with subsequent malignancy versus metastasis could not be distinguished.10 To our knowledge, no previous studies have examined associations between residence in NYC and mortality at COVID-19 discharge among patients with a history of cancer. This association may again be related to the severity of the early COVID-19 pandemic in NYC, compared to the rest of NYS. The association between an active cancer diagnosis claim code and higher risk of death at discharge is likely due to immunosuppression related to the cancer itself or treatment, or other effects of recent cancer treatment. Consistent with our findings, a prior study observed that patients with COVID-19 and recent cancer treatment had a higher risk of death (OR, 1.74; 95% CI, 1.54-1.96), while those with no recent cancer treatment did not have increased mortality (OR, 0.93; 95% CI, 0.84-1.02), compared to individuals without cancer.15 Some of the associations between demographic characteristics and death at discharge may be related to poorer outcomes in cancer patients with these characteristics, regardless of COVID-19 status. However, the occurrence of death during hospitalization for COVID-19 and the strength of the associations suggest a clear role of COVID-19 in these outcomes.
In adjusted analyses of the observed versus expected number of deaths by cancer type, we did not observe a statistically significant increased risk of death at discharge for any cancer type. In contrast, some previous studies reported that hematologic malignancies8,9,10,24 and lung cancer10 were associated with an increased risk of mortality after COVID-19. Possible reasons for this inconsistency may be differences in the cancer case selection or comparison group, as we only analyzed the death at discharge among patients with COVID-19-related hospitalizations, rather than all COVID-19 patients, and compared the mortality for each cancer type to that for all cancers combined. The use of a different comparison group, such as all COVID-19-related hospitalizations, would be expected to yield lower counts of the expected number of deaths and higher ratios of the observed to expected number of deaths. However, patients with a history of cancer may differ from other patients hospitalized for COVID-19 in ways that could not be controlled for in this analysis but that would be expected to impact mortality, such as having a higher number of comorbidities.9,15,16
Our results indicate an increased risk of COVID-19 hospitalization and death among cancer patients, and in particular those with certain demographic and tumor characteristics. The explanation for these findings is likely multifactorial and related to both immune function and risk patterns of severe COVID-19. A poorer immune response or worse course of illness in cancer patients that is related to the cancer itself, cancer treatment, or a higher prevalence of cancer-related comorbidities would be expected to lead to poorer outcomes after diagnosis with COVID-19. This is supported by the associations we observed with late-stage disease, multiple tumors, more recent diagnosis, and active cancer. Severity of COVID-19 due to diagnosis early in the pandemic, prior to the availability of effective treatments and vaccines, or reduced access to care likely also contributed to some of the results we observed, including the stronger associations for NYC, an early epicenter of the pandemic where the impact was greatest on vulnerable populations.22 Other associations may be related to both immune-related factors and COVID-19 severity, including the stronger associations observed among older individuals, men, and certain racial and ethnic groups. These demographic characteristics have been associated with more severe COVID-19 regardless of cancer status,27-29 and the associations we observed likely have both cancer-related and independent contributing factors including immune function, comorbidity burden, and patterns of care.
Strengths of this study include the availability of statewide population-based data on cancer diagnoses and COVID-19 hospitalizations, including patient demographics and case characteristics. This allowed us to look at a number of predictors of COVID-19 hospitalization and death at discharge in cancer patients. However, a small proportion of COVID-19 hospitalizations were likely missed, including those that occurred at Veterans Affairs and other military hospitals that are not captured in SPARCS. In addition, the match of the NYSCR and SPARCS data may have missed some cases that were not identified by our deterministic matching methods, which would be expected to result in an underestimation of the risk of COVID-19 hospitalization for cancer patients. For patients with multiple tumors, we used the type, stage, and date of diagnosis for the most recent tumor; however, in some cases this may not be the most relevant cancer diagnosis for the patient's health. Additionally, for analyses of risk of COVID-19 hospitaliza-tion among all patients with a history of cancer, region of NYS was categorized based on address at diagnosis for the most recent tumor and may not reflect a patient's current address, particularly for cases diagnosed further in the past. Finally, by using the first COVID-19 hospitalization record for patients who had multiple hospitalizations, we may have undercounted deaths among all COVID-19 patients or among COVID-19 patients with a prior cancer. Although we did not examine subsequent hospitalizations, a previous study of US electronic health record and administrative data reported that during the period from March to August 2020, 9% of patients hospitalized with COVID-19 were readmitted to the same hospital within 2 months of discharge but less than 0.1% of patients died during readmission, suggesting that only a small number of deaths were missed by omitting subsequent hospitalizations.30
In summary, our results indicate that cancer patients were more likely to be hospitalized for COVID-19 than individuals without a history of cancer, and, among cancer patients, several case characteristics and cancer types were associated with an increased risk of COVID-19 hospi-talization. In addition, patients with a history of cancer had a statistically significant increased risk of death after COVID-19 hospitalization, compared to patients without a history of cancer, and this risk was strongest among certain demographic groups and patients with an active cancer claim code at the time of their COVID-19 hospitalization. Consistent with the results of most previous studies, our results indicate a higher risk of severe COVID-19 outcomes among cancer patients, and in particular those in certain demographic groups or with certain diagnostic characteristics. Although this study focused on hospitalizations and deaths during the early part of the COVID-19 pandemic, prior to the widespread availability of vaccines and treatments for COVID-19, the results highlight the importance of continued vigilance to ensure the best possible outcomes for all patients with a history of cancer.
Footnotes
This work was supported in part by the Centers for Disease Control and Prevention's National Program of Cancer Registries through cooperative agreement 6NU58DP006309 awarded to the New York State Department of Health and by Contract 75N91018D00005 (Task Order 75N91018F00001) from the National Cancer Institute, National Institutes of Health.
References
- 1.Cancer Treatment & Survivorship Facts & Figures 2022-2024. American Cancer Society; 2022. [Google Scholar]
- 2.Graphs. National Cancer Institute Division of Cancer Control and Population Sciences website. Accessed September 25, 2022. https://cancercontrol.cancer.gov/ocs/statistics#graphs [Google Scholar]
- 3.Centers for Disease Control and Prevention. COVID-19: underlying medical conditions associated with higher risk for severe COVID-19: information for healthcare professionals. Updated June 15, 2022. https://www.cdc.gov/coronavirus/2019-ncov/hcp/clinical-care/underlyingconditions.html [PubMed] [Google Scholar]
- 4.National Center for Immunization and Respiratory Diseases (NCIRD), Division of Viral Diseases. Science brief: evidence used to update the list of underlying medical conditions associated with higher risk for severe COVID-19. In: CDC COVID-19 Science Briefs [Internet]. Centers for Disease Control and Prevention; 2020. [PubMed] [Google Scholar]
- 5.Robilotti EV, Babady NE, Mead PA, et al. Determinants of COVID-19 disease severity in patients with cancer. Nat Med. 2020;26(8): 1218–1223. doi: 10.1038/s41591-020-0979-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yang L, Chai P, Yu J, Fan X. Effects of cancer on patients with COVID-19: a systematic review and meta-analysis of 63,019 participants. Cancer Biol Med. 2021;18(1):298–307. doi: 10.20892/j.issn.2095-3941.2020.0559 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wang Q, Berger NA, Xu R. Analyses of risk, racial disparity, and outcomes among US patients with cancer and COVID-19 infection. JAMA Oncol. 2021;7(2):220–227. doi: 10.1001/jamaoncol.2020.6178 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Meng Y, Lu W, Guo E, et al. Cancer history is an independent risk factor for mortality in hospitalized COVID-19 patients: a propensity score-matched analysis. J Hematol Oncol. 2020;13(1):75. doi: 10.1186/s13045-020-00907-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Venkatesulu BP, Chandrasekar VT, Girdhar P, et al. A systematic review and meta-analysis of cancer patients affected by a novel coronavirus. medRxiv. Preprinted posted online May 29, 2020. doi: 10.1101/2020.05.27.20115303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sharafeldin N, Bates B, Song Q, et al. Outcomes of COVID-19 in patients with cancer: report from the National COVID Cohort Collaborative (N3C). J Clin Oncol. 2021;39(20):2232–2246. doi: 10.1200/JCO.21.01074 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Han S, Zhuang Q, Chiang J, et al. Impact of cancer diagnoses on the outcomes of patients with COVID-19: a systematic review and meta-analysis. BMJ Open. 2022; 12(2):e044661. doi: 10.1136/bmjopen-2020-044661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fillmore NR, La J, Szalat RE, et al. Prevalence and outcome of COVID-19 infection in cancer patients: a National Veterans Affairs study. J Natl Cancer Inst. 2021;113(6):691–698. doi: 10.1093/jnci/djaa159 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Joharatnam-Hogan N, Hochhauser D, Shiu KK, et al. Outcomes of the 2019 novel coronavirus in patients with or without a history of cancer: a multi-centre North London experience. Ther Adv Med Oncol. 2020;12:1758835920956803. doi: 10.1177/1758835920956803 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Klein IA, Rosenberg SM, Reynolds KL, et al. Impact of cancer history on outcomes among hospitalized patients with COVID-19. Oncologist. 2021;26(8):685–693. doi: 10.1002/onco.13794 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chavez-MacGregor M, Lei X, Zhao H, Scheet P, Giordano SH. Evaluation of COVID-19 mortality and adverse outcomes in US patients with or without cancer. JAMA Oncol. 2022;8(1):69–78. doi: 10.1001/jamaoncol.2021.5148 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kim Y, Zhu L, Zhu H, et al. Characterizing cancer and COVID-19 outcomes using electronic health records. PLoS One. 2022; 17(5):e0267584. doi: 10.1371/journal.pone.0267584 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Statewide Planning and Research Cooperative System (SPARCS). New York State Department of Health website. Revised August 2022. https://www.health.ny.gov/statistics/sparcs/ [Google Scholar]
- 18.Centers for Medicare and Medicaid Services. Preliminary Medicare COVID-19 Data Snapshot Methodology. https://www.cms.gov/files/document/medicare-covid-19-data-snapshot-methodology.pdf [PubMed] [Google Scholar]
- 19.Breslow NE, Day NE. Statistical Methods in Cancer Research: Vol. II, The Design and Analysis of Cohort Studies. IARC Scientific Publication No. 82. International Agency for Research on Cancer; 1987. [PubMed] [Google Scholar]
- 20.Fu J, Reid SA, French B, et al. Racial disparities in COVID-19 outcomes among black and white patients with cancer. JAMA Netw Open. 2022;5(3):e224304. doi: 10.1001/jamanetworkopen.2022.4304 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bialek S, Bowen V, Chow N, et al. CDC COVID-19 Response Team. Geographic differences in COVID-19 cases, deaths, and incidence— United States, February 12-April 7, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(15):465–471. doi: 10.15585/mmwr.mm6915e4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Thompson CN, Baumgartner J, Pichardo C, et al. COVID-19 Outbreak— New York City, February 29-June 1, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(46):1725–1729. doi: 10.15585/mmwr.mm6946a2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gu T, Mack JA, Salvatore M, et al. Characteristics associated with racial/ethnic disparities in COVID-19 outcomes in an academic health care system. JAMA Netw Open. 2020;3(10):e2025197. doi: 10.1001/jamanetworkopen.2020.25197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wang Q, Berger NA, Xu R. When hematologic malignancies meet COVID-19 in the United States: infections, death and disparities. Blood Rev. 2021;47:100775. doi: 10.1016/j.blre.2020.100775 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Tesoriero JM, Swain CE, Pierce JL, et al. COVID-19 outcomes among persons living with or without diagnosed HIV infection in New York State. JAMA Netw Open. 2021; 4(2):e2037069. doi: 10.1001/jamanetworkopen.2020.37069 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Shiels MS, Islam JY, Rosenberg PS, et al. Projected cancer incidence rates and burden of incident cancer cases in HIV-infected adults in the United States through 2030. Ann Intern Med. 2018;168(12):866–873. doi: 10.7326/M17-2499 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.St Sauver JL, Lopes GS, Rocca WA, et al. Factors associated with severe COVID-19 infection among persons of different ages living in a defined Midwestern US population. Mayo Clin Proc. 2021; 96(10):2528–2539. doi: 10.1016/j.mayocp.2021.06.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Williamson EJ, Walker AJ, Bhaskaran K, et al. Factors associated with COVID-19-related death using OpenSAFELY. Nature. 2020;584(7821):430–436. doi: 10.1038/s41586-020-2521-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Pijls BG, Jolani S, Atherley A, et al. Demographic risk factors for COVID-19 infection, severity, ICU admission and death: a metaanalysis of 59 studies. BMJ Open. 2021; 11 (1):e044640. doi: 10.1136/bmjopen-2020-044640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lavery AM, Preston LE, Ko JY, et al. Characteristics of hospitalized COVID-19 patients discharged and experiencing same-hospital readmis-sion—United States, March-August 2020. MMWR Morb Mortal Wkly Rep. 2020;69:1695–1699. doi: 10.15585/mmwr.mm6945e2 [DOI] [PMC free article] [PubMed] [Google Scholar]