Correction to: Pediatric Research https://doi.org/10.1038/s41390-021-01478-7, published online 6 April 2021
In the article “Validation of published rebound hyperbilirubinemia risk prediction scores during birth hospitalization after initial phototherapy: a retrospective chart review” by Vincent So, Helen Coo, and Faiza Khurshid (Pediatric Research, 2021), an error was brought to our attention in the data entry of phototherapy threshold values in a small fraction (~2%) of the dataset. Since the threshold value is utilized to calculate scores presented in our paper, it became apparent that a number of the values reported (and used to derive the graphs) are incorrect, albeit in most instances the differences are minor. This is also the case for our main findings (calibration curves and discrimination as measured by AUROC). No changes in p values or statistical analysis were noted in any analyses performed. We apologize for these regrettable errors. The errors are listed in detail below. The original article has been corrected.
1. The data entry of threshold values led to minor errors propagated in calculation of the AUROC, 2-Variable, 3-Variable Scores, 2-Variable Probability, 3-Variable Probability
Reported value | Corrected values | |
---|---|---|
2-Variable Score (SD) | RHB: 12.98 (12.15)*** Non-RHB: 5.40 (12.17)*** | RHB: 12.67 (12.11)*** Non-RHB: 5.47 (12.20)*** |
3-Variable Score (SD) | RHB: 38.35 (12.70)*** Non-RHB: 32.74 (10.94)*** | RHB: 40.07 (10.73)*** Non-RHB: 32.33 (10.89)*** |
2-Variable Probability (IQR) | RHB: 0.162 (0.077–0.314)*** Non-RHB: 0.081 (0.039–0.175)*** | RHB: 0.146 (0.077–0.314)*** Non-RHB: 0.086 (0.039–0.174)*** |
3-Variable Probability (IQR) | RHB: 0.175 (0.086–0.356)*** Non-RHB: 0.109 (0.063–0.192)*** | RHB: 0.189 (0.109–0.366)*** Non-RHB: 0.110 (0.059–0.183)*** |
2-Variable AUROC (95% CI) | 0.672 (0.600–0.745) | 0.662 (0.590–0.735) |
3-Variable AUROC (95% CI) | 0.634 (0.552–0.716) | 0.691 (0.619–0.763) |
Pearson’s coefficient between 2- and 3-Variable Scores | 0.650*** | 0.846*** |
Spearman’s coefficient between 2- and 3-variable probability | 0.672*** | 0.857*** |
2. In Table 2, the data entry resulted in minor changes in the following:
Reported value | Corrected values | |
Difference between threshold and TSB levels at PT end (1st treatment cycle) (SD) | All patients: −2.66 (1.72) Non-RHB: −2.90 (1.74)*** | All patients: −2.63 (1.72) Non-RHB: −2.86 (1.74)*** |
TSB levels PT end (2nd treatment cycle) (SD) | 13.85 (2.15) | 13.86 (2.15) |
Relative change in TSB levels between PT start and PT end, mg/dL (2nd treatment cycle) (SD) | −3.16 (2.85) | −3.16 (2.86) |
Difference between threshold and TSB levels at PT end, mg/dL (SD) | −2.31 (3.14) | −3.58 (1.97) |
In Table 3, data entry errors led to minor changes in bivariate and multivariate odds ratios. Also, the original value for “Difference Between Threshold and TSB Levels at PT End (in 3-Variable Score)” represented relative TSB (TSB-threshold). We felt it would be better represented as threshold-TSB (1/original value) for better comparison with the 2-Variable Score variable.
Odds ratios | Reported value | Corrected values | |
GA (35–37 weeks) (95% CI) | MV | 4.53 (1.71–12.03)** | 4.75 (1.79–12.61)** |
GA (38–39 weeks) (95% CI) | MV | 4.58 (1.91–11.00)** | 4.84 (2.01–11.66)*** |
Birthweight (<3000 g) (95% CI) | MV | 1.04 (0.45–2.39) | 1.02 (0.44–2.35) |
Birthweight (>3500 g) (95% CI) | MV | 1.05 (0.50–2.24) | 1.07 (0.50–2.26) |
DAT positive (95% CI) | MV | 0.88 (0.42–1.84) | 0.90 (0.43–1.88) |
Male sex (95% CI) | MV | 1.26 (0.67–2.36) | 1.25 (0.67–2.33) |
Resuscitation at birth (95% CI) | MV | 1.69 (0.69–3.72) | 1.61 (0.69–3.73) |
Age of PT start in days (in 3-Variable Score) (95% CI) | MV | 0.54 (0.29–0.99)* | 0.50 (0.27–0.93)* |
Difference between threshold and TSB levels at PT end (in 3-Variable Score) | MV | 1.44 (1.12–1.85)** | 0.67 (0.52–0.87)** |
Difference starting threshold and ending TSB (as in 2-Variable Score) (95% CI) | MV | 0.99 (0.81–1.22) | 1.04 (0.85–1.28) |
Difference between threshold and TSB levels at PT end (in 3-Variable Score) | BV | 1.48 (1.22–1.80)*** | 0.68 (0.56–0.83)*** |
Difference starting threshold and ending TSB (as in 2-Variable Score) (95% CI) | BV | 0.81 (0.72–0.91)*** | 0.82 (0.73–0.92)** |
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
So, V., Coo, H. & Khurshid, F. Correction to: Validation of published rebound hyperbilirubinemia risk prediction scores during birth hospitalization after initial phototherapy: a retrospective chart review. Pediatr Res 91, 1009–1010 (2022). https://doi.org/10.1038/s41390-021-01858-z
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41390-021-01858-z