Table 3.
Clinical Characteristics by Ethnicity and Logistic Regression of Odds of Death
| Characteristic | White (n = 235) | Black (n = 133) | Asian (n = 94) | Other (n = 17) | Missing (n = 135) | P Value |
|---|---|---|---|---|---|---|
| Male sex | 146 (62.13) | 76 (57.14) | 60 (63.83) | 12 (70.59) | 88 (65.19) | .63 |
| Mean age, y (SD) | 70.52 (15.69) | 63.58 (19.08) | 66.4 (15.88) | 54.82 (15.25) | 60.7 (18.11) | <.01 |
| Mean Elixhauser score (SD) | 6.29 (7.25) | 5.23 (6.85) | 6.63 (7.15) | 2.53 (4.37) | 2.99 (5.12) | <.01 |
| Mean NEWS-2 score (SD) | 5.09 (3.2) | 4.93 (3.04) | 4.74 (3.31) | 5.41 (2.67) | 5.39 (2.83) | .28 |
| Mean days to admission (SD) | 6.99 (5.75) | 7.33 (5.98) | 6.15 (4.82) | 7.29 (6.47) | 8.16 (5.08) | .49 |
| Outcomes | ||||||
| FiO2 ≥60% | 100 (42.55) | 61 (45.86) | 43 (45.74) | 5 (29.41) | 70 (51.85) | .31 |
| Admitted to ICU | 23 (9.79) | 20 (15.04) | 15 (15.96) | 2 (11.76) | 27 (20.00) | .10 |
| Received IVS | 22 (9.36) | 18 (13.53) | 14 (14.89) | 2 (11.76) | 24 (17.78) | .21 |
| Died in hospital | 67 (28.51) | 41 (30.83) | 30 (31.91) | 3 (17.65) | 37 (27.41) | .76 |
| Discharged alive | 149 (63.40) | 78 (58.65) | 54 (57.45) | 14 (82.35) | 86 (63.70) | .31 |
| Pending outcome | 19 (8.09) | 14 (10.53) | 10 (10.64) | 0 (0.00) | 12 (8.89) | .62 |
| Comorbidities | ||||||
| Any comorbidity | 179 (76.17) | 112 (84.21) | 78 (82.98) | 11 (64.71) | 100 (74.07) | .10 |
| Ischemic heart disease | 33 (14.04) | 16 (12.03) | 18 (19.15) | 0 (0.00) | 13 (9.63) | .12 |
| Chronic heart failure | 19 (8.09) | 11 (8.27) | 1 (1.06) | 0 (0.00) | 6 (4.44) | .07 |
| Hypertension | 97 (41.28) | 73 (54.89) | 40 (42.55) | 7 (41.18) | 69 (51.11) | .08 |
| Hyperlipidemia | 57 (24.26) | 32 (24.06) | 27 (28.72) | 2 (11.76) | 33 (24.44) | .66 |
| Diabetes | 61 (25.96) | 63 (47.37) | 41 (43.62) | 4 (23.53) | 47 (34.81) | <.01 |
| CKD | 39 (16.60) | 26 (19.55) | 26 (27.66) | 0 (0.00) | 11 (8.15) | <.01 |
| Peripheral vascular disease | 11 (4.68) | 3 (2.26) | 0 (0.00) | 0 (0.00) | 4 (2.96) | .19 |
| Stroke | 28 (11.91) | 13 (9.77) | 11 (11.70) | 1 (5.88) | 4 (2.96) | .05 |
| Atrial fibrillation | 43 (18.30) | 10 (7.52) | 11 (11.70) | 2 (11.76) | 9 (6.67) | <.01 |
| DVT/PE history | 4 (1.70) | 2 (1.50) | 0 (0.00) | 0 (0.00) | 1 (0.74) | .68 |
| Hemiplegia | 6 (2.55) | 0 (0.00) | 0 (0.00) | 0 (0.00) | 0 (0.00) | .04 |
| Dementia | 42 (17.87) | 9 (6.77) | 10 (10.64) | 0 (0.00) | 9 (6.67) | <.01 |
| Asthma | 18 (7.66) | 14 (10.53) | 8 (8.51) | 3 (17.65) | 13 (9.63) | .65 |
| COPD | 22 (9.36) | 3 (2.26) | 3 (3.19) | 0 (0.00) | 2 (1.48) | <.01 |
| Connective tissue disease | 4 (1.70) | 3 (2.26) | 1 (1.06) | 0 (0.00) | 1 (0.74) | .82 |
| Peptic ulcer | 8 (3.40) | 2 (1.50) | 1 (1.06) | 0 (0.00) | 3 (2.22) | .61 |
| Liver (noncirrhotic) | 14 (5.96) | 12 (9.02) | 10 (10.64) | 3 (17.65) | 5 (3.70) | .09 |
| Liver (cirrhotic) | 2 (0.85) | 4 (3.01) | 6 (6.38) | 0 (0.00) | 1 (0.74) | .02 |
| Solid tumor | 33 (14.04) | 9 (6.77) | 8 (8.51) | 1 (5.88) | 8 (5.93) | .06 |
| Hematologic tumor | 4 (1.70) | 1 (0.75) | 1 (1.06) | 0 (0.00) | 2 (1.48) | .92 |
| HIV/AIDS | 1 (0.43) | 7 (5.26) | 1 (1.06) | 0 (0.00) | 0 (0.00) | <.01 |
| Logistic regression of odds of death by ethnicity | ||||||
| Unadjusted, OR (95% CI) | Intercept | 1.17 (.73–1.88) | 1.24 (.73–2.10) | 0.48 (.13–1.71) | 0.96 (.59–1.55) | |
| Adjusteda model, OR (95% CI) | Intercept | 1.85* (1.06–3.24) | 1.78 (.97–3.29) | 0.95 (.24–3.80) | 1.63 (.93–2.84) |
Data are presented as no. (%) unless otherwise indicated.
Abbreviations: CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DVT/PE, deep vein thrombosis/pulmonary embolism; FiO2, fraction of inspired oxygen; NEWS-2, New Early Warning Score; ICU, intensive care unit; HIV, human immunodeficiency virus; IVS, invasive ventilation support; SD, standard deviation.
aAdjusted regression model for age, male sex, Elixhauser score, and NEWS-2 differences by ethnic groups. Additional logistic regression models adjusted for any comorbidity, diabetes, and CKD were also carried out, but were inferior predictors compared to the selected model. COPD, cirrhotic liver disease, and HIV/AIDS were not used in logistic regression models, as they had low n values for ethnic groups.
*P < .05.