From: Treatment effect prediction with adversarial deep learning using electronic health records
Characteristics | No. of participants (n = 736) | Readmission in one year (n = 461) | Non-readmission in one year (n = 275) | P-value |
---|---|---|---|---|
Age (years), mean (SD) | 64.29 ± 13.55 | 63.66 ± 13.55 | 65.34 ± 13.53 | 0.104 |
Female sex (T/F) | 508/227 | 331/130 | 177/97 | 0.050 |
Hypertension (T/F) | 526/210 | 323/138 | 203/72 | 0.314 |
Diabetes mellitus (T/F) | 466/270 | 283/178 | 183/92 | 0.185 |
Renal insufficiency (T/F) | 592/144 | 359/102 | 233/42 | 0.030 |
SBP (mmHg), mean (SD) | 133.41 ± 20.41 | 130.33 ± 20.02 | 138.57 ± 20.04 | <  0.001 |
DBP (mmHg), mean (SD) | 77.13 ± 13.66 | 76.15 ± 13.86 | 78.76 ± 13.19 | 0.012 |
Heart rate (b.p.m) mean (SD) | 79.98 ± 16.37 | 81.17 ± 17.01 | 78.00 ± 15.06 | 0.011 |
Creatinine (umol/L), mean (SD) | 100.35 ± 64.5 | 106.77 ± 72.85 | 89.61 ± 45.50 | <  0.001 |
LVEF (%), mean (SD) | 43.74 ± 11.86 | 41.92 ± 12.12 | 46.80 ± 10.76 | <  0.001 |
CK (umol/L), mean (SD) | 87.79 ± 82.04 | 89.71 ± 80.56 | 84.60 ± 84.50 | 0.414 |
cTnT (ng/ml), mean (SD) | 0.058 ± 0.38 | 0.077 ± 0.47 | 0.025 ± 0.057 | 0.068 |
Treatment | ||||
Diuretics (T/F) | 536/200 | 344/117 | 202/73 | <  0.001 |
ACEI (T/F) | 442/294 | 279/182 | 163/112 | 0.797 |
ARB (T/F) | 480/256 | 296/165 | 184/91 | 0.507 |
Beta-blocker (T/F) | 588/148 | 367/94 | 221/54 | 0.879 |
CCB (T/F) | 454/282 | 307/154 | 147/128 | <  0.001 |
Statin (T/F) | 536/200 | 322/139 | 214/61 | 0.023 |
Digoxin (T/F) | 457/279 | 257/204 | 200/75 | <  0.001 |
Nitrates (T/F) | 454/282 | 274/187 | 180/95 | 0.122 |
Aspirin (T/F) | 513/223 | 314/147 | 199/76 | 0.258 |
Clopidogrel (T/F) | 379/357 | 244/217 | 140/135 | 0.650 |
Warfarin (T/F) | 638/98 | 399/62 | 239/36 | 0.979 |
Spironolactone (T/F) | 402/334 | 288/173 | 161/114 | 0.328 |
Antibiotics (T/F) | 713/23 | 446/15 | 267/8 | 0.967 |
Antiacid (T/F) | 589/147 | 367/94 | 222/53 | 0.786 |