Abstract
This study examined the sociodemographic and clinical characteristics of patients, as well as their clinical outcomes, during a surge in COVID-19 disease in Houston, Texas.
Texas is experiencing resurgence of coronavirus disease 2019 (COVID-19). We report sociodemographic, clinical, and outcome differences across the first and second surges of COVID-19 hospitalizations at Houston Methodist, an 8-hospital health care system in Houston, Texas.1
Methods
From electronic health records, we identified patients with positive reverse transcriptase–polymerase chain reaction (RT-PCR) nasopharyngeal swab test results for severe acute respiratory syndrome coronavirus 2. We extracted age, sex, race/ethnicity, comorbidity, medication, intensive care unit (ICU) admission, and mortality information. The assessment of race/ethnicity was driven by prior analyses of our data that demonstrated higher SARS-CoV-2 infection rates among racial and ethnic minorities.2 We tracked daily total, ICU, and non-ICU (medical/surgical units) hospital census across the reporting period. We categorized patients into surge 1 for admissions between March 13 and May 15, 2020, and surge 2 between May 16 and July 7, 2020. Surge 2 started 2 weeks after a phased statewide reopening.3
We provided summary statistics as means or medians and proportions for various sociodemographic, clinical, and outcome characteristics of hospitalized COVID-19 patients. Proportional differences with 95% CIs are provided for bivariable comparisons across surges 1 and 2. Extraction and reporting of these data were not deemed human subjects research by the Houston Methodist Institutional Review Board. Analyses were performed with Stata version 16. P values were 2 sided, with statistical significance set at P < .05.
Results
As of July 7, 2020, 2904 unique COVID-19 patients had been hospitalized, representing 774 and 2130 patients during surge 1 and 2, respectively. The Figure presents total, ICU, and non-ICU daily hospital census along with a 7-day mean across the study period. Dates corresponding to various phases of statewide reopening are also highlighted. Patients in surge 2 (vs surge 1) were younger (mean age, 57.3 vs 59.9 years; difference, −2.62 years; 95% CI, −4.04 to −1.20 years), the proportion identifying as Hispanic was higher (43.3% vs 25.7%; difference, 17.64%; 95% CI, 13.89%-28.79%), and the median zip code–based income was lower ($60 765 vs $65 805; difference, −$5040; 95% CI, −$7641 to −$2439). Surge 2 patients had a significantly lower burden of overall and specific comorbidities such as diabetes, hypertension, and obesity (Table).
Table. Sociodemographic, Comorbidity, Clinical, and Outcome Differences Between Surge 1 and Surge 2 of COVID-19 Hospitalizations at Houston Methodist, Texas.
Surge 1: March 13 to May 15 (64 d)a | Surge 2: May 16 to July 7 (53 d)a | Difference (95% CI)b | P value | |
---|---|---|---|---|
No. | 774 | 2130 | ||
Demographic and social characteristics | ||||
Age, mean (SD), y | 59.9 (16.9) | 57.3 (17.4) | −2.62 (−4.04 to −1.20) | <.001 |
Age ≤50 y, No. (%) | 208 (26.9) | 736 (34.6) | 7.68 (3.96 to 11.40) | <.001 |
Non-Hispanic race, No. (%)c | ||||
White | 257 (45.2) | 543 (46.1) | 0.97 (−4.02 to 5.95) | .70 |
Black | 256 (45.0) | 534 (45.4) | 0.38 (−4.60 to 5.36) | .88 |
Asian | 46 (8.1) | 83 (7.1) | −1.03 (−3.71 to 1.64) | .44 |
Other | 10 (1.8) | 17 (1.4) | −0.31 (−1.59 to 0.96) | .62 |
Hispanic or Latino, No. (%)c | 196 (25.7) | 910 (43.3) | 17.64 (13.89 to 28.79) | <.001 |
Insurance, No. (%) | ||||
Commercial | 305 (39.4) | 769 (36.1) | −3.30 (−7.30 to 0.70) | .10 |
Medicare | 333 (43.0) | 774 (36.3) | −6.69 (−10.73 to −2.64) | .001 |
Self-pay | 88 (11.4) | 423 (19.9) | 8.49 (5.68 to 11.30) | <.001 |
Medicaid | 32 (4.1) | 141 (6.6) | 2.49 (0.73 to 4.24) | .01 |
Other | 16 (2.1) | 23 (1.1) | −0.99 (−0.21 to −0.11) | .04 |
Zip code income, median (IQR), $ | 65 805 (48 790 to 86 034) | 60 765 (46 300 to 76 163) | −5040 (−7641 to −2439) | <.001 |
Comorbidity profiled | ||||
Charlson Comorbidity Index, median (IQR) score | 3 (1 to 6) | 2 (1 to 4) | −1 (−1.30 to −0.71) | <.001 |
Diabetes (with or without complications), No. (%) | 312 (40.3) | 475 (32.0) | −8.34 (−12.54 to −4.15) | <.001 |
Hypertension, No. (%) | 427 (55.3) | 583 (38.8) | −16.52 (−20.81 to −12.24) | <.001 |
Obesity (BMI ≥30), No. (%) | 261 (33.9) | 383 (25.7) | −8.19 (−12.20 to −4.18) | <.001 |
Therapies, No. (%) | ||||
Remdesivir | 87 (11.2) | 472 (22.2) | 10.92 (8.08 to 13.76) | <.001 |
Convalescent plasma therapy | 89 (11.5) | 235 (11.0) | −0.47 (−3.08 to 2.15) | .72 |
Enoxaparin | 494 (63.8) | 1546 (72.6) | 8.76 (4.88 to 12.64) | <.001 |
Hydroxychloroquine | 436 (56.3) | 11 (0.5) | −55.81 (−59.32 to −52.31) | <.001 |
Other anticoagulants | 333 (43.0) | 635 (29.8) | −13.21 (−17.20 to −9.22) | <.001 |
Pulmonary diagnoses, severity of care indicators, and outcomes, No. (%)d | ||||
Pneumonia | 658 (85.3) | 1435 (87.8) | 2.53 (−0.43 to 5.50) | .09 |
Acute respiratory distress syndrome | 146 (18.9) | 76 (5.1) | −13.81 (−16.79 to −10.84) | <.001 |
Lower respiratory tract infection | 23 (3.0) | 13 (0.9) | −2.11 (−3.41 to −0.82) | <.001 |
Acute bronchitis | 20 (2.6) | 24 (1.6) | −0.98 (−2.27 to 0.32) | .11 |
ICU admission | 295 (38.1) | 427 (20.1) | −18.07 (−21.89 to −14.25) | <.001 |
Invasive mechanical ventilator use | 186 (24.0) | 230 (10.8) | −13.23 (−16.52 to −9.95) | <.001 |
Ventilation days, median (IQR)e | 13 (5 to 26) | 5.7 (3 to 10.3) | −7.31 (−9.58 to −5.05) | <.001 |
Extracorporeal membrane oxygenation | 15 (1.9) | 6 (0.3) | −1.65 (−2.65 to −0.66) | <.001 |
Currently hospitalized | 0 | 652 (30.6) | 30.61 (28.65 to 32.57) | <.001 |
Died | 94 (12.1) | 75 (3.5) | −8.62 (−11.05 to −6.19) | <.001 |
Died (excluding currently hospitalized)d | 75 (5.1) | −7.07 (−9.63 to −4.51) | <.001 | |
Died (ICU admissions excluding currently hospitalized)d,f | 81 (27.5) | 49 (22.9) | −4.56 (−12.15 to 3.03) | .24 |
Length of hospital stay, median (IQR), dd | 7.1 (3.9 to 14.4) | 4.8 (2.8 to 7.8) | −2.31 (−2.78 to −1.84) | <.001 |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); COVID-19, coronavirus disease 2019; ICU, intensive care unit; IQR, interquartile range.
Missing by variable No. (%): race, 129 (4.4); Hispanic, 41 (1.4); and median zip income, 23 (0.8).
Differences calculated as surge 2 – surge 1. Negative values represent decrease in surge 2. Null value = 0.
Self-reported in accordance with predefined categories. Hispanic ethnicity excluded and analyzed separately. Non-Hispanic race total n: surge 1 = 569; and surge 2 = 1177.
Among discharged patients (n = 2252).
For nonzero ventilator days.
Among patients who received care in the ICU.
A greater proportion of surge 2 patients received remdesivir and enoxaparin. A smaller proportion of surge 2 patients were admitted to the ICU (20.1% vs 38.1%; difference, −18.07%; 95% CI, −21.89% to −14.25%). Length of hospital stay was less (4.8 vs 7.1 days; difference, −2.31 days; 95% CI, −2.78 to −1.84 days). Among dead or discharged patients (n = 2252 [77.5%] overall; n = 774 in surge 1 and n = 1478 in surge 2), surge 2 in-hospital mortality was significantly lower compared with that for surge 1 (5.1% vs 12.1%; difference, −7.07%; 95% CI, −9.63% to −4.51%). In-hospital mortality among discharged and deceased ICU-treated patients during surge 2 was not significantly lower than that during surge 1 (49/214 [22.9%] vs 81/295 [27.5%]; difference, −4.56%; 95% CI, −12.15% to 3.03%). The mean daily proportion of individuals with positive RT-PCR results during surge 1 was 13%, whereas it was 25% during surge 2.
Discussion
An increase in COVID-19 hospitalizations was observed across a major health care system in the greater Houston area, which was temporally related to phased reopening. Throughout the reporting period, hospital admission guidelines were consistently based on risk stratification by evaluation of severity of symptoms, comorbidities, diagnostic findings, and pulse oximetry. During surge 2, the absolute number of RT-PCR tests performed increased, as did the proportion of positive results. Therefore, higher hospital census likely reflects higher rates of community COVID-19 prevalence. Surge 2 data indicated a demographic shift of the pandemic toward a younger, predominantly Hispanic, and lower socioeconomic patient population with an overall lower comorbidity burden, ICU admission rate, and in-hospital mortality. The demographic and socioeconomic shift may reflect return to work and relaxation of COVID-19 transmission mitigation practices. Additionally, in-hospital mortality among ICU-treated surge 2 patients was 4.6% lower than that in surge 1. The overall better outcomes during surge 2 may be explained by a combination of lower comorbidity burden, lesser disease severity, and better medical management.
Limitations of the study include data from a single hospital system that may not be generalizable. The shift toward non-ICU resources implies that different staffing patterns and infection control practices may be needed. Lower acuity and ICU use and shorter lengths of stay may allow for increased capacity and less overall stress on health care resources.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
References
- 1.Tittle S, Braxton C, Schwartz RL, et al. . A guide for surgical and procedural recovery after the first surge of Covid-19. NEJM Catalyst. Published July 2, 2020. Accessed August 4, 2020. https://catalyst.nejm.org/doi/full/10.1056/cat.20.0287
- 2.Vahidy FS, Nicolas JC, Meeks JR, et al. . Racial and ethnic disparities in SARS-CoV-2 pandemic: analysis of a COVID-19 observational registry for a diverse US metropolitan population. medRxiv. Preprint posted online May 12, 2020. doi: 10.1101/2020.04.24.20073148 [DOI] [PMC free article] [PubMed]
- 3.Office of the Texas Governor. Governor’s strike force to open Texas. Published 2020. Accessed July 14, 2020. https://open.texas.gov/