Abstract
Introduction
Tobacco use is a risk factor for COVID-19 severity. This study explores an association between tobacco use and COVID-19–linked all-cause healthcare utilization in ambulatory patients.
Methods
Among 49,588 University of Maryland Medical System patients who tested positive for COVID-19 between February 1, 2020, and October 31, 2021, 20,621 ambulatory patients at first presentation were analyzed using a cross-sectional study. A multinomial multivariable logistic regression model was used to test the impact of tobacco use on hospital and emergency department utilization, with COVID-19 severity, comorbid diagnoses, obesity, chronic steroid use, sex, race, age, and Social Vulnerability Index included as covariates.
Results
Of the 20,621 patients, 2,030 (9.84%) were current users; 4,586 (22.24%) were former users; and 14,005 (67.92%) never used tobacco. A total of 16,518 (80.10%) patients remained ambulatory during their COVID-19 illness; 1,786 (8.66%) utilized the emergency department; and 2,317 (11.24%) were admitted to the hospital. Both former (AOR=1.286; 95% CI=1.132, 1.462; p=0.0001) and current (AOR=1.450; 95% CI=1.246, 1.689; p<0.0001) tobacco users were more likely to visit the emergency department than never users. However, only former users were significantly more likely to be hospitalized than never users (AOR=1.320; 95% CI=1.187, 1.468; p<0.0001).
Conclusions
Tobacco users with COVID-19 are more likely to have increased healthcare utilization than never users, with current users more likely to use the emergency department and former users more likely to utilize the hospital. Ambulatory patients who use tobacco should receive closer COVID-19 quarantine management to prevent severe outcomes and healthcare overutilization.
Keywords: COVID-19, tobacco, smoking, healthcare utilization, ambulatory
Highlights
-
•
This cross-sectional study assessed tobacco use and healthcare utilization.
-
•
The study sample focused on patients with COVID-19 ambulatory at first presentation.
-
•
Tobacco users with COVID-19 are likely to have increased healthcare utilization.
-
•
The level of utilization varied on the basis of tobacco use status (current versus former).
INTRODUCTION
Tobacco has detrimental health effects, and its use is the leading cause of preventable disease in the U.S., resulting in $240 billion in healthcare costs in 2018.1,2 Tobacco also adversely impacts coronavirus disease 2019 (COVID-19) infection outcomes, corresponding to greater use of emergency department (ED) and inpatient hospital services.3, 4, 5 Tobacco use and advanced age, hypertension, chronic heart disease, diabetes, sex, chronic kidney disease, and obesity are associated with a higher prevalence of severe COVID-19 as well as being predictors of hospitalization.3,6, 7, 8, 9, 10 In addition, long-term steroid use has been linked to decreased immune response, which could potentially also exacerbate COVID-19 symptoms and lead to hospitalization.11
Recent studies have found that whereas people who formerly smoked and were hospitalized with COVID-19 had a higher percentage of severe clinical outcomes than people who never smoked, people who currently smoke had a lower percentage of hospitalizations and death than people who formerly or never smoked.12,13
A study conducted on the COVID electronic health record (EHR) Cohort at the University of Wisconsin (CEC-UW) found that people who formerly smoked and were hospitalized with COVID-19 are at higher risk for severe outcomes; this association was not found for people who currently smoke and were hospitalized with COVID-19. The CEC-UW is a National Cancer Institute (NCI)–supported retrospective cohort study that consists of EHR data from 21 participating healthcare systems, including the University of Maryland Medical System (UMMS). The cohort analysis did not include patients with COVID-19 who were ambulatory at first presentation, and the role of tobacco use in ambulatory patients with COVID-19 was not fully investigated, leading to an incomplete picture of tobacco use–related outcomes.13
Using UMMS data from the same cohort as CEC-UW, this study explored the effect of tobacco use on all-cause healthcare utilization in patients with COVID-19 with ambulatory presentation at UMMS.
METHODS
Study Population
The COVID-19 + Smoking project is a subproject from the NCI Cancer Moonshot Cancer Center Cessation Initiative that collects data retrospectively on COVID-19–positive patients who use tobacco. As part of the COVID-19 + Smoking project, the NCI COVID-19 EHR Cohort includes data from 21 health systems from across the U.S., including the UMMS Greenebaum Comprehensive Cancer Center.14
The UMMS is a university-based regional healthcare system that provides primary and specialty ambulatory care in 150+ locations as well as inpatient care at 12 hospitals across the state. The UMMS comprises 2,458 licensed beds and saw 100,985 hospital admissions; 1,230,086 outpatient visits; and 329,547 emergency visits in fiscal year 2022.15
The CEC-UW study was initially approved in May 2020 by the University of Wisconsin-Madison Health Sciences Minimal Risk IRB to collect deidentified EHR data and then modified in February 2021 to a limited data set. Approval was also granted by the University of Maryland, Baltimore IRB. Study reporting follows STROBE guidelines (Text S1; 14) as well as criteria outlined through the JBI critical appraisal checklist for analytical cross-sectional studies.16 Presented results include extracted UMMS Epic EHR data from February 1, 2020, to October 31, 2021. To be included in the analysis, a patient must have presented to UMMS with COVID-19 during the study period. Those who were not ambulatory at first presentation were excluded.
Patients with unknown tobacco use status generally did not receive the bulk of their health care through the UMMS system. Owing to the transitional nature of these encounters, the patients may be less likely to have a complete medical history in their UMMS EHR patient file and would not contribute reliable results in the analysis. As such, these patients were also excluded from the study population.
Measures
The source data files contained patient- and encounter-level information on (1) sociodemographic and health characteristics, including ICD-10 codes for medical problems; (2) tobacco use status, including types; and (3) dates for all clinical encounters, including treatment site (e.g., inpatient, outpatient), encounter-based ICD-10 diagnoses, presenting signs and symptoms, and other clinical data.
All-cause healthcare utilization was the primary outcome, reflecting the maximum level of care the patient received during the observation period (time of first presentation to October 31, 2021). Utilization level was assessed for 3 mutually exclusive categories: (1) ambulatory care alone (reference), (2) ED care without hospitalization, and (3) ever hospitalized. All instances of patient care that occurred within the study period were included in the analysis.
Tobacco use status was the primary predictor variable and was categorized as never user, former user, and current user. Status was obtained from the time of the first presentation or, if missing, at the first nonmissing record. All records occurred within 1 year of the first presentation. Any form of tobacco use was determined to be a positive response.
Measures used as covariates include COVID-19 severity, comorbid diagnoses, obesity, chronic steroid use, sex, race, age, and Social Vulnerability Index (SVI). COVID-19 severity was measured on the basis of symptom expression at the time of the first presentation. Three mutually exclusive categories were defined using NIH guidance:17
-
1.
Low severity: asymptomatic (positive test for severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2] but no symptoms consistent with COVID-19) or mild illness (symptoms consistent with COVID-19 but no shortness of breath [SOB] or pneumonia);
-
2.
Medium severity: moderate illness (symptoms consistent with COVID-19, including SOB or pneumonia, and an oxygen saturation level ≥94%); and
-
3.
High severity: severe or critical illness (symptoms consistent with COVID-19, including SOB or pneumonia, and an oxygen saturation level <94%).
Comorbid diagnoses were synthesized using an Elixhauser Comorbidity Index score weighted to predict the risk of 30-day, all-cause readmission. The index is a composite score comprising 38 individual comorbidity measures, with each measure defined through specific ICD-10-CM codes.18 Because obesity was included as a separate covariate, the obesity comorbidity measure was excluded from the index score. The index score was analyzed as a categorical variable in this study, with scores ≤0 placed in the lowest category and all remaining scores divided into quartiles on the basis of the overall sample distribution. ICD-10 diagnoses codes present in the patient’s EHR record between February 1, 2020, and the time of the first presentation were included in the calculation of the index score.
Obesity and chronic steroid use were included separately owing to being strong predictors of COVID-19 outcomes and were expressed as binary variables.8, 9, 10, 11,19 Any instance of an ICD-10 code related to obesity (R93.9, E66, Z68.3, Z68.4, O99.21) or chronic steroid use (Z79.52) between February 1, 2020, and the time of the first presentation was determined to be a positive response.18,20
Race was divided into White non-Hispanic, Black non-Hispanic, other non-Hispanic, and Hispanic.21 Age was expressed in years, and all patients aged >90 years were recorded as aged 90 years.
The 2018 Centers for Disease Control and Prevention/Agency for Toxic Substances and Disease Registry SVI ranks each U.S. Census tract on 15 social factors and provides an overall ranking, with higher scores indicating lower SES.22 ZIP code crosswalk files produced by the U.S. Department of Housing and Urban Development and the U.S. Postal Service were used to match each Census tract’s Centers for Disease Control and Prevention FIPS 11-digit identification to its corresponding ZIP code.23 A median SVI was then calculated for each patient’s ZIP code.
Statistical Analysis
Descriptive statistics for the study sample were calculated, with frequency and percentage reported for categorical variables and mean and SD reported for continuous variables. Significance of associations between healthcare utilization and each of the independent variables were analyzed using bivariate analysis; chi-square tests were used for categorical variables, and F-tests were used for continuous variables. The association between tobacco use and healthcare utilization was tested using a multinomial multivariable logistic regression model, with and without adjustment of the covariates. Statistical significance for all tests was set at α=0.05 (2-tailed test), and all analyses were computed using SAS, Version 9.4 (SAS Institute Inc, Cary, NC).
RESULTS
Data for 49,588 UMMS patients who tested positive for COVID-19 between February 1, 2020, and October 31, 2021, were extracted. After matching the source file data and excluding patients who were not ambulatory at first visit or had missing data, the final study sample size was 20,621 patients (Appendix Figure 1, available online).
Patient demographic and health characteristics are presented in Table 1. Among the 20,621 patients, 2,030 (9.84%) were current tobacco users; 4,586 (22.24%) were former users; and 14,005 (67.92%) never used tobacco. Most current and former tobacco users used combustible tobacco products, with only 160 patients indicating that they used smokeless tobacco products (Appendix Table 1, available online).
Table 1.
Demographic and Clinical Characteristics of the Study Population by Tobacco-Use Status
| Characteristic | Total, n (%) | Current tobacco user, n (%) | Former tobacco user, n (%) | Never used, n (%) |
|---|---|---|---|---|
| Full sample | 20,621 (100) | 2,030 (9.84) | 4,586 (22.24) | 14,005 (67.92) |
| Healthcare utilization | ||||
| Ambulatory only | 16,518 (80.10) | 1,557 (76.70) | 3,434 (74.88) | 11,527 (82.31) |
| Emergency department | 1,786 (8.66) | 245 (12.07) | 406 (8.85) | 1,135 (8.10) |
| Hospital admission | 2,317 (11.24) | 228 (11.23) | 746 (16.27) | 1,343 (9.59) |
| COVID-19 severity | ||||
| Low | 18,870 (91.51) | 1,885 (92.86) | 4,108 (89.58) | 12,877 (91.95) |
| Medium | 1,567 (7.60) | 131 (6.45) | 422 (9.20) | 1,014 (7.24) |
| High | 184 (0.89) | 14 (0.69) | 56 (1.22) | 114 (0.81) |
| Elixhauser Comorbidity Indexa | ||||
| ≤0 | 11,679 (56.64) | 992 (48.87) | 2,040 (44.48) | 8,647 (61.74) |
| 1–3 | 2,382 (11.55) | 234 (11.53) | 609 (13.28) | 1,539 (10.99) |
| 4–6 | 2,434 (11.80) | 252 (12.41) | 538 (11.73) | 1,644 (11.74) |
| 7–11 | 2,018 (9.79) | 248 (12.22) | 565 (12.32) | 1,205 (8.60) |
| 12–66 | 2,108 (10.22) | 304 (14.98) | 834 (18.19) | 970 (6.93) |
| Obesity | ||||
| Yes | 4,601 (22.31) | 422 (20.79) | 1,308 (28.52) | 2,871 (20.50) |
| No | 16,020 (77.69) | 1,608 (79.21) | 3,278 (71.48) | 11,134 (79.50) |
| Chronic steroid use | ||||
| Yes | 137 (0.66) | 13 (0.64) | 57 (1.24) | 67 (0.48) |
| No | 20,484 (99.34) | 2,017 (99.36) | 4,529 (98.76) | 13,938 (99.52) |
| Sex | ||||
| Male | 7,959 (38.60) | 974 (47.98) | 2,039 (44.46) | 4,946 (35.32) |
| Female | 12,662 (61.40) | 1,056 (52.02) | 2,547 (55.54) | 9,059 (64.68) |
| Race | ||||
| Non-Hispanic White | 11,683 (56.66) | 1,018 (50.15) | 3,113 (67.88) | 7,552 (53.92) |
| Non-Hispanic Black | 6,599 (32.00) | 866 (42.66) | 1,168 (25.47) | 4,565 (32.60) |
| Non-Hispanic other | 1,415 (6.86) | 96 (4.73) | 173 (3.77) | 1,146 (8.18) |
| Hispanic | 924 (4.48) | 50 (2.46) | 132 (2.88) | 742 (5.30) |
| Mortality | ||||
| Yes | 187 (0.91) | 17 (0.84) | 88 (1.92) | 82 (0.59) |
| No | 20,434 (99.09) | 2,013 (99.16) | 4,498 (98.08) | 13,923 (99.41) |
| Age | ||||
| Mean ± SD | 47.57 ± 18.56 | 47.24 ± 15.19 | 57.43 ± 15.84 | 44.39 ± 18.70 |
| Median Social Vulnerability Index | ||||
| Mean ± SD | 0.35 ± 0.23 | 0.42 ± 0.25 | 0.34 ± 0.22 | 0.34 ± 0.22 |
The weighting used for the Elixhauser Comorbidity Index predicts the risk of 30-day, all-cause readmission.
A higher percentage of current and former tobacco users had higher comorbidity score, and went to the ED or were hospitalized than patients who never used tobacco. Former users tended to be oldest (57.43 ± 15.84 years), and those who never used tobacco were generally the youngest (44.39 ± 18.70 years). The median SVI reflected higher social vulnerability among current users (0.42), compared with that of former users and never users (0.34). A greater percentage of former users had high COVID-19 severity (1.22%) than current users (0.69%) and never users (0.81%). A higher percentage of former users were obese (28.52%) or were using steroids (1.24%) than current (obesity: 20.79%; chronic steroid use: 0.64%) and never (obesity: 20.50%; chronic steroid use: 0.48%) users. Among the study population, 17 (0.84%) current users, 88 (1.92%) former users, and 82 (0.59%) patients who never used tobacco died during the study period.
Overall, 2,317 (11.24%) patients with COVID-19 diagnosis were hospitalized; 1,786 (8.66%) patients went to the ED but were not hospitalized; and 16,518 (80.10%) patients received ambulatory care alone during the study period. Those who had ever used tobacco had a greater association with going to the ED or being hospitalized than those who did not use tobacco (p<0.0001). In addition, COVID-19 severity, comorbid diagnoses, obesity, chronic steroid use, sex, race, age, and SVI were also associated with the level of healthcare utilization (p<0.0001) (Table 2).
Table 2.
Association Between Healthcare Utilization and Demographic and Clinical Characteristics (Bivariate Analysis)
| Categorical patient variable | Ambulatory only, n=16,518 n (%) |
Emergency department, n=1,786 n (%) |
Hospital admission, n=2,317 n (%) |
χ2 (p-value) |
|---|---|---|---|---|
| Tobacco-use status | 196.5651 (<0.0001) | |||
| Current user | 1,557 (9.43) | 245 (13.72) | 228 (9.84) | |
| Former user | 3,434 (20.79) | 406 (22.73) | 746 (32.203) | |
| Never used | 11,527 (69.78) | 1,135 (63.55) | 1,343 (57.96) | |
| COVID-19 severity | 125.5649 (<0.0001) | |||
| Low | 15,150 (91.72) | 1,653 (92.55) | 2,067 (89.21) | |
| Medium | 1,264 (7.65) | 121 (6.77) | 182 (7.85) | |
| High | 104 (0.63) | 12 (0.67) | 68 (2.93) | |
| Elixhauser Comorbidity Indexa | 908.9657 (<0.0001) | |||
| ≤0 | 9,835 (59.54) | 921 (51.57) | 923 (39.84) | |
| 1–3 | 1,934 (11.71) | 221 (12.37) | 227 (9.80) | |
| 4–6 | 1,951 (11.81) | 247 (13.83) | 236 (10.19) | |
| 7–11 | 1,522 (9.21) | 175 (9.80) | 321 (13.85) | |
| 12–66 | 1,276 (7.72) | 222 (12.43) | 610 (26.33) | |
| Obesity | ||||
| Yes | 3,479 (21.06) | 431 (24.13) | 691 (29.82) | 93.7166 (<0.0001) |
| No | 13,039 (78.94) | 1,355 (75.87) | 1,626 (70.18) | |
| Chronic steroid use | ||||
| Yes | 67 (0.41) | 17 (0.95) | 53 (2.29) | 111.4819 (<0.0001) |
| No | 16,451 (99.59) | 1,769 (99.05) | 2,264 (97.71) | |
| Sex | 21.0141 (<0.0001) | |||
| Male | 6,480 (39.23) | 603 (33.76) | 876 (37.81) | |
| Female | 10,038 (60.77) | 1,183 (66.24) | 1,441 (62.19) | |
| Race | 238.0431 (<0.0001) | |||
| Non-Hispanic White | 9,559 (57.87) | 886 (49.61) | 1,238 (53.43) | |
| Non-Hispanic Black | 4,993 (30.23) | 754 (42.22) | 852 (36.77) | |
| Non-Hispanic Other | 1,281 (7.76) | 58 (3.25) | 76 (3.28) | |
| Hispanic | 685 (4.15) | 88 (4.93) | 151 (6.52) | |
| Continuous patient variable, mean ± SD | ||||
| Age, years | 46.91 ± 18.18 | 44.87 ± 19.12 | 54.42 ± 19.30 | 190.48 (<0.0001) |
| Median Social Vulnerability Index | 0.34 ± 0.22 | 0.41 ± 0.23 | 0.38 ± 0.22 | 111.57 (<0.0001) |
Note: Boldface indicates statistical significance (p<0.05).
The weighting used for the Elixhauser Comorbidity Index predict the risk of 30-day, all-cause readmission.
Table 3 summarizes the results from unadjusted and adjusted models testing the associations between tobacco use status and all-cause healthcare utilization. Full results from the adjusted model are presented in Appendix Table 2 (available online). Both former (AOR=1.286; 95% CI=1.132, 1.462; p=0.0001) and current (AOR= 1.450; 95% CI=1.246, 1.689; p<0.0001) tobacco users were more likely to go to the ED, whereas only former users were significantly more likely to be hospitalized than never users (AOR=1.320; 95% CI=1.187, 1.468; p<0.0001).
Table 3.
Association Between Healthcare Utilization and Demographic and Clinical Characteristics (Results of Multinomial Multivariate Logistic Regression Modeling)
| Outcome/model (ref=ambulatory only) |
Unadjusted |
Adjusteda |
||||
|---|---|---|---|---|---|---|
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| Emergency department | ||||||
| Never used (ref) | 1.000 | — | — | 1.000 | — | — |
| Former tobacco user | 1.201 | 1.065, 1.353 | 0.0027 | 1.286 | 1.132, 1.462 | 0.0001 |
| Current tobacco user | 1.598 | 1.378, 1.853 | <0.0001 | 1.450 | 1.246, 1.689 | <0.0001 |
| Hospital admission | ||||||
| Never used (ref) | 1.000 | — | — | 1.000 | — | — |
| Former tobacco user | 1.865 | 1.692, 2.055 | <0.0001 | 1.320 | 1.187, 1.468 | <0.0001 |
| Current tobacco user | 1.257 | 1.082, 1.460 | 0.0028 | 1.022 | 0.874, 1.196 | 0.7838 |
Note: Boldface indicates statistical significance (p<0.05).
Among other predictors, patients with high COVID-19 severity had over fourfold greater odds of being hospitalized (OR=4.088; 95% CI=2.945, 5.675; p<0.0001) than patients with low COVID-19 severity. Patients with an Elixhauser Comorbidity Index score of 12–66 had over threefold greater odds of being hospitalized than patients with a score ≤0 (OR=3.327; 95% CI=2.916, 3.797; p<0.0001). The odds of increased healthcare utilization for patients who were obese were only statistically significant for hospital admissions (OR=1.177; 95% CI=1.058, 1.308; p=0.0026), whereas those who used steroids chronically had statistically significant greater odds of going to the ED (OR=1.834; 95% CI=1.063, 3.164; p=0.0293) and being hospitalized (OR=2.500; 95% CI=1.707, 3.662; p<0.0001).
Females had greater odds of going to the ED (OR=1.220; 95% CI=1.099, 1.356; p=0.0002) or being hospitalized (OR=1.111; 95% CI=1.011, 1.221; p=0.0283) than males. Non-Hispanic Blacks had greater odds of going to the ED (OR=1.205; 95% CI=1.069, 1.359; p=0.0023) or being hospitalized (OR=1.168; 95% CI=1.044, 1.307; p=0.0066) than non-Hispanic Whites, whereas Hispanic patients were over 2 times more likely to be hospitalized (OR=2.161; 95% CI=1.777, 2.628; p<0.0001). In contrast, non-Hispanic other patients had lower odds of going to the ED (OR=0.481; 95% CI=0.365, 0.633; p<0.0001) or being hospitalized (OR=0.611; 95% CI=0.478, 0.780; p<0.0001) than non-Hispanic Whites.
As patients’ age increases, their likelihood of going to the ED decreases (OR=0.992; 95% CI=0.989, 0.995; p<0.0001), but they were increasingly more likely to be admitted to the hospital (OR=1.015; 95% CI=1.012, 1.017; p<0.0001). Higher median SVI scores were also associated with increased healthcare utilization; every 1 unit increase in median SVI scores resulted in approximately 2 times greater odds of going to the ED (OR=2.495; 95% CI=1.961, 3.174; p<0.0001) or hospital (OR=1.839; 95% CI=1.488, 2.327; p<0.0001).
DISCUSSION
During the COVID-19 pandemic, there was an urgent need to identify predictors of COVID-19 severity that would result in hospitalization or ED use. This study aimed to shed light on tobacco use as a risk factor for increased healthcare utilization among ambulatory patients with COVID-19 by utilizing real-world data extracted from the Epic EHR of a large health system.
Both former and current tobacco users had higher odds of increased all-cause healthcare utilization than never users. However, current users were more likely to utilize the ED than former users. The association between ED utilization and tobacco use persisted after controlling for other factors, aligning with the observations of other researchers.13,24,25 Most ambulatory patients with COVID-19 are managed in primary care, and knowledge of a patient’s tobacco usage could trigger closer supervision during quarantine for illness as well as appropriate therapeutics use and hospitalization.
Among patients who required hospital admission, current users were less likely to be hospitalized than former users, even after adjusting for other factors. Former users were generally older and had higher Elixhauser Comorbidity Index scores than current users, suggesting that the increased hospitalization rate may be due to the presence of chronic illnesses related to advanced age. Another potential explanation for former users having higher odds of hospitalization may be due to patients either quitting or temporarily pausing their tobacco use when they fall ill and become hospitalized, leading to their categorization as former users rather than current users. It is well documented that EHR records on smoking status among hospitalized individuals is poor.26 Further analysis of patients’ tobacco use timelines would be beneficial to increased understanding of this data trend.
The literature presents mixed results regarding the smoker’s paradox theory; however, there has been renewed interest recently owing to COVID-19. A systematic review conducted by Gonzalez-Rubio et al.27 concluded that there may be a protective effect among people who currently smoke against hospitalization. The findings among this study cohort seem to support a slight protective effect among people who currently smoke; however, this does not outweigh the numerous proven adverse health effects of tobacco use.
Research suggests that smoking is one of the most immediately modifiable risk factors to reduce the COVID-19 morbidity burden.28 The results of this study support the well-known short- and long-term health benefits of quitting tobacco, which start as soon as an individual quits and continue for years.25 For future pandemic readiness, there should be considerations for increasing tobacco cessation efforts in vulnerable populations. Policymakers, public health officials, and relevant healthcare stakeholders should ensure that efforts are undertaken to support and maintain tobacco cessation efforts to reduce the impact of the COVID-19 pandemic on already strained healthcare systems.
Data results from this study support consideration of revisions to ambulatory care policies to promote new workflows for all tobacco users positive for COVID-19 to receive closer quarantine management and monitoring for vital signs to facilitate early detection of severe COVID-19 and early therapeutic interventions to minimize preventable increased healthcare utilization. The use of remote patient monitoring or telemedicine can assist with managing patients in quarantine while increasing access and convenience for patients and reducing COVID-19 transmission risk to clinical staff.
Patients with higher COVID-19 severity, greater number of comorbid diagnoses, and other chronic conditions such as obesity or steroid use as well as demographics such as females, non-Hispanic Blacks and Hispanics, and those with higher social vulnerability are more likely to have increased healthcare utilization. Tobacco use also increases the likelihood of these outcomes, compounding the effect among the previously mentioned populations. New patient-centered clinical workflows and evaluations of social determinants of health are necessary to ensure that care delivery is sensitive to the needs of each individual and enhances health equity.
Limitations
This cohort includes data from the UMMS catchment area, and the resulting observations may not be generalizable to other parts of the U.S. In addition, the study population was limited to patients with COVID-19 who were ambulatory at the first visit; those who were admitted to the ED or hospital at the first visit were not included. Owing to the nature of this data collection process and the short time frame of the study period, information on patients’ hospitalizations prior to their first ambulatory COVID-19 visit was not available, and the effect of this predictor was not analyzed.
Second, tobacco-use status was a snapshot variable that was determined at the first nonmissing record and does not differentiate between duration and intensity of tobacco use nor different tobacco types. In addition, the UMMS EHR system is designed so that the most recently inputted social history status overwrites the previous record. Therefore, efforts toward tobacco cessation or other status changes during the study period were not considered.
CONCLUSIONS
Ambulatory patients with COVID-19 who are former or current tobacco users are more likely to have greater healthcare utilization. These patients may need closer surveillance during COVID-19 quarantine by clinical practices for early intervention with therapeutics and transfer to hospital. Tobacco users in ambulatory care should be offered tobacco cessation interventions to prevent severe outcomes and healthcare overutilization in future pandemics.
Future research should further examine the outcome differences between former and current tobacco users. A longitudinal study would be beneficial to capturing the effects of any tobacco status changes during the study period, including duration of use or time elapsed since quitting. Furthermore, incorporating analyses of individual tobacco product types and the amount of use will provide further clarification on their association with COVID-19 illness severity and degree of healthcare utilization.
CRediT authorship contribution statement
Niharika Khanna: Conceptualization, Data curation, Funding acquisition, Methodology, Supervision, Writing – review & editing. Carissa S. Kwan: Data curation, Formal analysis, Methodology, Visualization, Writing – original draft, Writing – review & editing. Elena N. Klyushnenkova: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Visualization, Writing – review & editing. Michael B. Dark: Funding acquisition, Project administration, Writing – review & editing. Janaki Deepak: Conceptualization, Funding acquisition, Methodology, Writing – review & editing.
ACKNOWLEDGMENTS
Acknowledgments: The following individuals contributed substantively to this work: Colleen Kernan, MPH; Elizabeth McMahon, MS; and Adam Gaynor, MPH at Department of Family and Community Medicine, University of Maryland School of Medicine, and Julia Melamed, BSN, RN at University of Maryland Medical Center, University of Maryland School of Medicine. Data are not publicly available owing to Health Insurance Portability and Accountability Act regulations.
Disclaimer: Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NIH.
Funding: This data was collected as a subproject of the NCI Moonshot C3I project suring the COVID-19 pandemic. This data was provided to the NCI COVID-19 EHR Cohort Data Registry.
Previous presentation: This manuscript has been submitted solely to the American Journal of Preventive Medicine and has not been previously published. The manuscript has not been made available as a preprint. Preliminary data results from this study were presented at a poster session during the Society for Research on Nicotine and Tobacco (SRNT) 2022 Annual Meeting. An abstract was also accepted to the AcademyHealth 2023 Annual Research Meeting and the 2023 Annual SRNT-E Conference poster sessions.
Declaration of interest: Financial support was provided by the National Cancer Institute. No other financial disclosures were reported.
Footnotes
Supplementary material associated with this article can be found in the online version at doi:10.1016/j.focus.2025.100352.
Contributor Information
Niharika Khanna, Email: nkhanna@som.umaryland.edu.
Janaki Deepak, Email: jadeepak@som.umaryland.edu.
Appendix. Supplementary materials
REFERENCES
- 1.Smoking cessation: a report of the surgeon general – key findings. Office of the Surgeon General, HHS.https://www.hhs.gov/surgeongeneral/reports-and-publications/tobacco/2020-cessation-sgr-factsheet-key-findings/index.html. Updated January 23, 2020. Accessed November 17, 2022.
- 2.National Center for Chronic Disease Prevention and Health Promotion (US) Office on Smoking and Health. The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta (GA): Centers for Disease Control and Prevention (US); 2014. https://www.ncbi.nlm.nih.gov/books/NBK179276/. Accessed December 14, 2024. [PubMed]
- 3.Mileham K.F., Bruinooge S.S., Aggarwal C., et al. Changes over time in COVID-19 severity and mortality in patients undergoing cancer treatment in the United States: initial report from the ASCO registry. JCO Oncol Pract. 2022;18(4):e426–e441. doi: 10.1200/OP.21.00394. https://DOI.ORG/10.1200/OP.21.00394 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Navas Alcántara M.S., Montero Rivas L., Guisado Espartero M.E., et al. Influence of smoking history on the evolution of hospitalized in COVID-19 positive patients: results from the SEMI-COVID-19 registry. Med Clin (Barc) 2022;159(5):214–223. doi: 10.1016/j.medcli.2021.10.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sharma P., Zeki AA. Does vaping increase susceptibility to COVID-19? Am J Respir Crit Care Med. 2020;202(7):1055–1056. doi: 10.1164/rccm.202005-2103LE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Shastri M.D., Shukla S.D., Chong W.C., et al. Smoking and COVID-19: what we know so far. Respir Med. 2021;176 doi: 10.1016/j.rmed.2020.106237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chen Z., Russo N.W., Miller M.M., Murphy R.X., Burmeister DB. An observational study to develop a scoring system and model to detect risk of hospital admission due to COVID-19. J Am Coll Emerg Physicians Open. 2021;2(2) doi: 10.1002/emp2.12406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Demeulemeester F., de Punder K., van Heijningen M., van Doesburg F. Obesity as a risk factor for severe COVID-19 and complications: a review. Cells. 2021;10(4):933. doi: 10.3390/cells10040933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Michalakis K., Panagiotou G., Ilias I., Pazaitou-Panayiotou K. Obesity and COVID-19: a jigsaw puzzle with still missing pieces. Clin Obes. 2021;11(1) doi: 10.1111/cob.12420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Foo O., Hiu S., Teare D., Syed A.A., Razvi S. A global country-level analysis of the relationship between obesity and COVID-19 cases and mortality. Diabetes Obes Metab. 2021;23(12):2697–2706. doi: 10.1111/dom.14523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hasan S.S., Capstick T., Zaidi S.T.R., Kow C.S., Merchant HA. Use of corticosteroids in asthma and COPD patients with or without COVID-19. Respir Med. 2020;170 doi: 10.1016/j.rmed.2020.106045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Puebla Neira D., Watts A., Seashore J., Polychronopoulou E., Kuo Y.F., Sharma G. Smoking and risk of COVID-19 hospitalization. Respir Med. 2021;182 doi: 10.1016/j.rmed.2021.106414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Piasecki T.M., Smith S.S., Baker T.B., et al. Smoking status, nicotine medication, vaccination, and COVID-19 hospital outcomes: findings from the COVID EHR Cohort at the University of Wisconsin (CEC-UW) study. Nicotine Tob Res. 2023;25(6):1184–1193. doi: 10.1093/ntr/ntac201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Nolan M.B., Piasecki T.M., Smith S.S., et al. Relations of current and past cancer with severe outcomes among 104,590 hospitalized COVID-19 patients: the COVID EHR cohort at the University of Wisconsin. Cancer Epidemiol Biomarkers Prev. 2023;32(1):12–21. doi: 10.1158/1055-9965.EPI-22-0500. https://DOI.ORG/10.1158/1055-9965.EPI-22-0500 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.UMMS fact sheet 2023. University of Maryland Medical System. https://www.umms.org/-/media/files/umms/about-us/member-hospitals/fact-sheets/umms-factsheet-2023.pdf?upd=20230130142713. Updated January 2023. Accessed February 10, 2023.
- 16.Moola S., Munn Z., Tufanaru C., et al. In: JBI Manual for Evidence Synthesis. Aromataris E., Munn Z., editors. Joanna Briggs Institute; Miami, FL: 2020. Chapter 7. Systematic reviews of etiology and risk.https://jbi-global-wiki.refined.site/space/MANUAL/355598596/7.+Systematic+reviews+of+etiology+and+risk Accessed December 14, 2024. [Google Scholar]
- 17.Clinical spectrum of SARS-CoV-2 infection. 2024. COVID-19 treatment guidelines. https://www.ncbi.nlm.nih.gov/books/NBK570371/pdf/Bookshelf_NBK570371.pdf. Accessed December 14, 2024.
- 18.User guide: Elixhauser comorbidity software refined for ICD-10-CM diagnoses, v2022.1.https://www.hcup-us.ahrq.gov/toolssoftware/comorbidityicd10/CMR-User-Guide-v2022-1.pdf. Updated October 2021. Accessed January 19, 2023.
- 19.Underlying conditions and the higher risk for severe COVID-19. Centers for Disease Control and Prevention.https://www.cdc.gov/covid/hcp/clinical-care/underlying-conditions.html. Updated Ferbruary 6, 2025. Accessed May 7, 2025. [PubMed]
- 20.ICD-10-code for long term (current) use of systemic steroids Z79.52. Codify by AAPC. https://www.aapc.com/codes/icd-10-codes/Z79.52. Accessed December 13, 2024.
- 21.Khanna N., Klyushnenkova E.N., Kaysin A. Association of COVID-19 with race and socioeconomic factors in family medicine. J Am Board Fam Med. 2021;34(suppl):S40–S47. doi: 10.3122/jabfm.2021.S1.200338. [DOI] [PubMed] [Google Scholar]
- 22.CDC. SVI; 2018. Documentation. https://www.atsdr.cdc.gov/place-health/php/svi/index.html. Updated July 22, 2024. Accessed December 13, 2024.
- 23.HUD USPS ZIP code crosswalk files. HUD User. https://www.huduser.gov/portal/datasets/usps_crosswalk.html#codebook. Updated August 2022. Accessed February 10, 2023.
- 24.What are the physical health consequences of tobacco use? National Institute on Drug Abuse. https://nida.nih.gov/publications/research-reports/tobacco-nicotine-e-cigarettes/what-are-physical-health-consequences-tobacco-use. Updated January 2020. Accessed December 30, 2022.
- 25.Benefits of quitting. Centers for Disease Control and Prevention. https://www.cdc.gov/tobacco/about/benefits-of-quitting.html?CDC_AAref_Val=https://www.cdc.gov/tobacco/quit_smoking/how_to_quit/benefits/index.html. Updated May 15, 2024. Accessed December 30, 2022.
- 26.Alla F., Berlin I., Nguyen-Thanh V., et al. Tobacco and COVID-19: a crisis within a crisis? Can J Public Health. 2020;111(6):995–999. doi: 10.17269/s41997-020-00427-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.González-Rubio J., Navarro-López C., López-Nájera E., et al. A systematic review and meta-analysis of hospitalised current smokers and COVID-19. Int J Environ Res Public Health. 2020;17(20):7394. doi: 10.3390/ijerph17207394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Reddy R.K., Charles W.N., Sklavounos A., Dutt A., Seed P.T., Khajuria A. The effect of smoking on COVID-19 severity: a systematic review and meta-analysis. J Med Virol. 2021;93(2):1045–1056. doi: 10.1002/jmv.26389. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.