I would like to report an erratum to “A Bayesian latent class approach for EHR‐based phenotyping” by Hubbard et al (38(1):74‐87). Due to a typographical error, the stratified sample sizes and first two rows reported in Table 4 are incorrect. A corrected Table 4 is provided below.
Table 4.
Characteristics of study population of pediatric patients at risk for T2DM stratified according to absence of codes for T1DM and presence of codes for T2DM, metformin prescription, or elevated hemoglobin A1c or glucose
| Total | Codes or Biomarkers Suggesting T2DM | ||
|---|---|---|---|
| Yes | No | ||
| N = 68 265 | N = 804 | N = 67 461 | |
| N (%) | N (%) | N (%) | |
| Male | 36 836 (53.96) | 221 (27.49) | 36 615 (54.28) |
| White | 35 740 (52.35) | 371 (46.14) | 35 369 (52.43) |
| Endocrinologist | 5338 (7.82) | 510 (63.43) | 4828 (7.16) |
| Metformin | 764 (1.12) | 675 (83.96) | 89 (0.13) |
| Insulin | 727 (1.06) | 154 (19.15) | 573 (0.85) |
| T1D Codes | 632 (0.93) | 0 (0) | 632 (0.94) |
| T2D Codes | 275 (0.4) | 221 (27.49) | 54 (0.08) |
| Any glucose measurement | 11 325 (16.59) | 355 (44.15) | 10 970 (16.26) |
| Any HbA1c measurement | 6031 (8.83) | 397 (49.38) | 5634 (8.35) |
| Mean (SD) | Mean (SD) | Mean (SD) | |
| Age | 11.90(2.50) | 13.79 (2.58) | 11.87 (2.49) |
| BMI | 2.02 (0.30) | 2.27 (0.36) | 2.01 (0.30) |
| Glucose | 94.31 (32.51) | 141.39 (104.47) | 92.79 (27.44) |
| Hemoglobin A1c | 5.79 (1.25) | 6.93 (1.94) | 5.71 (1.15) |
Additionally, the number of patients with biomarkers or codes indicative of type 2 diabetes stated on page 81 as 5043 (7.4%) should be 804 (1.2%). No other results in the manuscript were affected by this error.
Hubbard RA. Correction to “A Bayesian latent class approach for EHR‐based phenotyping”. Statistics in Medicine. 2020;39:205–205. 10.1002/sim.8436