Table 1.
Demographics characteristics and ICD diagnosis of patients for which the gold standard corpora were developed and used for evaluating the performance of the NLP and DL methods at the three sites.
| WCM | NM | UF | |
|---|---|---|---|
| Total patients (n) | 134 | 313 | 341 |
| Total notes (N) | 301 | 400 | 400 |
| Avg age (SD) on date of note documented | 48.43(18.26) | 44(14.52) | 52.18(15.17) |
| Age categories | |||
| < 18 | 2(1.5%) | 1(0.3%) | 4(1.2%) |
| 18–39 | 47(35.1%) | 127(40.6%) | 66(19.4%) |
| 40–59 | 41(30.6%) | 129(41.2%) | 157(46%) |
| ≥60 | 44(32.8%) | 56(17.9%) | 114(33.4%) |
| Sex | |||
| Female | 81(60.4%) | 198(63.3%) | 231(67.7%) |
| Male | 53(39.6%) | 115(36.7%) | 110(32.3%) |
| Race | |||
| White | 86(64.2%) | 229(73.2%) | 203(59.5%) |
| Black/AA | 10(7.5%) | 44(14.1%) | 124(36.4%) |
| Native/Alaskan | 0(0.0%) | 2(0.6%) | NA |
| Asian | 4(3.0%) | 5(1.6%) | 2(0.6%) |
| Other/Declined /Unknown | 34(25.4%) | 32(10.2%) | 12(3.5%) |
| Ethnicity | |||
| Hispanic | 12(9.0%) | 44(14.1%) | 10(2.9%) |
| Non-Hispanic | 120(89.6%) | 255(81.5%) | 328(96.2%) |
| Other/Declined/Unknow | 1(0.7%) | 32(10.2%) | 3(0.9%) |
| Marital Status | |||
| Single | 83(61.9%) | 167(53.4%) | NA |
| Married | 30(22.4%) | 99(31.6%) | NA |
| Divorced | 13(9.7%) | NA | NA |
| Widowed | 7(5.2%) | NA | NA |
| Other | 1(0.7%) | 47(15%) | NA |
| ICD diagnosis of SI on analyzed encounter | |||
| Yes | 1(0.7%) | 9(2.9%) | 5(1.5%) |
| No | 133(99.3%) | 304(97.1%) | 336(98.5%) |
| ICD diagnosis of SB on analyzed encounter | |||
| Yes | 0(0.0%) | 6(1.9%) | 2(0.6%) |
| No | 134(100.0%) | 307(98.1%) | 339(99.4%) |
| ICD diagnosis of SI before analyzed encounter | |||
| Yes | 1(0.7%) | 31(9.9%) | 27(7.9%) |
| No | 133(99.3%) | 282(90.1%) | 314(92.1%) |
| ICD diagnosis of SB before analyzed encounter | |||
| Yes | 1(0.7%) | 23(7.3%) | 17(5.0%) |
| No | 133(99.3%) | 290(92.7%) | 324(95.0%) |
| Provider specialty | |||
| Psychiatric | 31(23.1%) | 669(21.1%) | 2(0.6%) |
| Non-Psychiatric | 103(76.9%) | 247 (78.9%) | 339(99.4%) |