Internal validation sub-studies compare an observed imperfect measurement of a variable with its gold standard measurement in a sample of participants. The gold standard is often expensive or difficult to collect, so the validation sample is usually small. We simulated a small cohort study with a nested, balanced validation sub-study and used the simulated cohorts to evaluate three analytic strategies with the objective of determining which of them, alone or in combination, improved the performance of quantitative bias adjustments informed by the validation sub-study. The strategies were: (1) incorporating a weak prior to the beta distributions assigned to the bias parameters, as opposed to no prior (2) restricting beta distribution assignment to the three predictive values informed by sampling, as opposed to all four predictive values, and (3) restricting imputation of gold-standard measurement to study cohort members who were not sampled for the validation sub-study, as opposed to imputing for all members of the study cohort. All twelve combinations of bias adjustment strategies returned bias-adjusted estimates with less bias and higher coverage than the simulated misclassified cohorts, although all had higher root mean squared errors and wider intervals because of the uncertainty added by incorporating the sparse validation data into the analysis. We found that assigning a prior may sometimes have utility, whereas restricting assignment of beta distributions for use only with sampling and restricting imputation to only cohort members not selected for the validation study adds computational complexity with little utility in return.
Background: The Tanner scale is the reference for assessing pubertal development. In epidemiologic studies, self- or parent-assessments are frequently used, though their accuracy is not regularly discussed. We compared this measurement with clinical evaluations (as gold standard) among 12-year-old children in the general French population and examined factors explaining discrepancies. Methods: Within the PELAGIE mother–child cohort conducted in Brittany (France), clinical examinations of 498 children were performed by trained medical staff to assess pubertal development (breasts, gonads, and pubic hair) using Tanner stages. Children and parents completed assessments within the previous days, and had the option to declare two stages in case of hesitation. Agreement between the self- and parent reports and the clinical assessments was evaluated using weighted kappa, intraclass correlation coefficient (ICC), and Correspondence Analysis. We used robust Poisson regressions to examine potential correlates (determined a priori ) of discordant assessments. Results: Among girls, 68.1% correctly assessed breast stage compared with clinical assessment (ICC=0.72), though overestimation was common (31.1%). Accuracy for pubic hair was 68.6% in girls (ICC=0.79) and 60.3% in boys (ICC=0.73). Only 38.3% of boys accurately reported their gonadal stage (ICC=0.38), often underestimating (68.1%). Self-reporting of hesitant responses improved concordance with clinical assessments. Pubertal and morphological indicators, such as voice change, contributed to reducing discrepancies in self-assessment. Parents’ reports showed similar trends to children’s, but with frequent non-responses. Conclusion: The accuracy of self-assessment of pubertal stage in our French population of 12-year-olds was generally good, supporting its use as a reliable tool, particularly among girls. For gonadal development, a potential systematic and informative errors were present, with underestimation. The use of reporting hesitant responses in children appears to be a promising approach while the added value of parental assessment remains limited.
Introduction: Administrative health data are essential for quantifying incidence and long-term outcomes associated with pediatric critical illness but rely on reliable capture and coding of hospitalizations with pediatric intensive care unit (PICU) admission. The accuracy of administrative data capture of pediatric critical care admission in Canada is unknown. Methods: We assessed the performance of the Discharge Abstract Database (DAD), which captures all hospitalizations in British Columbia, Canada for identifying pediatric critical care admission by comparing records to the Virtual Pediatric Systems (VPS, LLC, Los Angeles, CA), a prospectively collected database of PICU admissions at British Columbia Children’s Hospital, British Columbia’s only dedicated pediatric quaternary care hospital. We identified all hospitalizations for children younger than 18 years at British Columbia Children’s Hospital between 2014 and 2023 and used presence of PICU days in the DAD to identify possible PICU admissions. We then matched the DAD records to the VPS records via exact admission date to determine the performance of this case definition (presence of PICU days) both overall, and across subgroups. We evaluated the accuracy of PICU admission characteristics and outcomes reporting in the DAD using VPS as the reference standard. Results: A total of 63,256 admissions were included, of which 8891 had an indicator for PICU admission in the DAD. Of these, 8729 had a positive validation record for PICU admission in the VPS database at British Columbia Children’s Hospital. There were 123 admissions to the PICU in the VPS database that were not captured in the DAD. The case definition had excellent sensitivity (96.3%; 95% CI = 95.9, 96.7) and specificity (99.7%; 95% CI = 99.7, 99.7). Results were consistent across subgroups. DAD definitions of admission type, PICU length of stay, inter-hospital transfer status, receipt of invasive mechanical ventilation, and mortality all showed sensitivity and specificity ≥ 85%. Conclusion: Using a large validation sample of PICU admissions from a quaternary care centre, we found that routinely collected administrative data on hospitalizations in Canada had high accuracy for classifying and characterizing PICU admissions.
