Table 1 Baseline characteristics of the patients in training and testing dataset
From: Identifying who are unlikely to benefit from total knee arthroplasty using machine learning models
Cohorts | Knee dataset | Training set | Testing set |
|---|---|---|---|
Number of images | 7224 | 5778 | 1446 |
Number of participants | 5720 | 4575 | 1145 |
Demographics | |||
Male | 1513 (21) | 1232 (21) | 281 (19) |
Age, year | 67.0\(\pm\)7.6 | 67.2\(\pm\)7.5 | 67.0\(\pm\)7.7 |
BMI | |||
Normal (BMI\(<\)24) | 1462 (21) | 1151 (20) | 311 (22) |
Overweight (24\(\le\)BMI\(<\)28) | 2693 (37) | 2138 (37) | 555 (38) |
Obese (28\(\le\)BMI\(<\)32) | 1954 (27) | 1609 (28) | 345 (24) |
Severely obese (BMI\(\ge\)32) | 1115 (15) | 880 (15) | 235 (16) |
Ethnicity | |||
Chinese | 6288 (87) | 5024 (87) | 1264 (87) |
Malay | 500 (6.9) | 410 (7.1) | 90 (6.2) |
Indian | 370 (5.1) | 295 (5.1) | 75 (5.2) |
Others | 66 (0.9) | 49 (0.8) | 17 (1.2) |
Clinical characteristics | |||
Hypertension | 4501 (62) | 3,619 (63) | 882 (61) |
Diabetes | 1332 (18) | 1094 (19) | 238 (16) |
Hyperlipidemia | 3515 (49) | 2828 (49) | 687 (48) |
Ischemic heart disease | 596 (8.3) | 483 (8.4) | 113 (7.8) |
Arthritis other than knee OA | 240 (3.3) | 188 (3.3) | 52 (3.6) |
Depression | 66 (0.9) | 43 (0.7) | 23 (1.6) |
Outcomes at 6 months | |||
Not achieve KSS MCID | 2035 (28) | 1631 (28) | 404 (28) |
Not achieve OKS MCID | 857 (12) | 556 (9.6) | 134 (9.3) |
Not achieve SF-MCS MCID | 5309 (73) | 4262 (74) | 1047 (72) |
Not achieve SF-PCS MCID | 2734 (38) | 2196 (38) | 538 (37) |
Outcomes at 2 years | |||
Not achieve KSS MCID | 1808 (25) | 1432 (25) | 376 (26) |
Not achieve SF-MCS MCID | 5172 (72) | 4137 (72) | 1035 (72) |
Not achieve SF-PCS MCID | 2320 (32) | 1841 (32) | 479 (33) |
Not achieve OKS MCID | 559 (7.7) | 346 (6.0) | 88 (6.1) |
