STUDY 2
Chang PW, Newman TB. A simpler prediction rule for rebound hyperbilirubinemia. Pediatrics. 2019 Jul;144(1):e20183712.
Background
Hyperbilirubinemia (jaundice) is estimated to affect 50%-60% of all newborns. Rebound hyperbilirubinemia – a rise in bilirubin after cessation of phototherapy – is common and can lead to recently discharged infants being readmitted for additional therapy. Lack of clear guidelines regarding when to discharge infants with hyperbilirubinemia has likely contributed to practice variation and some trepidation regarding whether a bilirubin level is “low enough” to discontinue therapy.
Study overview and results
The authors had previously proposed a three-factor hyperbilirubinemia risk model and sought to simplify their rule further.4 They examined a retrospective cohort of 7,048 infants greater than or equal to 35 weeks’ gestation using a random split sample. The authors derived a two-factor model using the same methods and compared its performance to the three-factor model. The two-factor formula was shown to be a good fit as a logistic regression model (Hosmer-Lemeshow test 9.21; P = .33), and the AUROC (area under the receiver operating characteristic) curves for the derivation and validation cohorts were similar between the two-factor (0.877 and 0.876, respectively) and three-factor risk models (0.887 and 0.881, respectively).
Limitations
These data are limited to infants receiving their first treatment of phototherapy and have not been externally validated. An important variable, serum bilirubin at phototherapy termination, was estimated in most subjects, which may have affected the accuracy of the prediction rule. Whether infants received home phototherapy was based only on equipment orders, and some infants may have received phototherapy unbeknownst to investigators. Last, infants with rebound hyperbilirubinemia at less than 72 hours after phototherapy discontinuation may have been missed.
Important findings and implications
This prediction model provides evidence-based, concrete data that can be used in making joint decisions with families regarding discharge timing of infants with hyperbilirubinemia. It also could be beneficial when deciding appropriate follow-up time after discharge.
Dr. Christopher Russo
STUDY 3
Ramgopal S et al. Risk of serious bacterial infection in infants aged ≤60 days presenting to emergency departments with a history of fever only. J Pediatr. 2019 Jan;204:191-195. doi: 10.1016/j.jpeds.2018.08.043.
Background
Febrile infants aged 60 days and younger are at risk for serious bacterial infections (SBI) including urinary tract infections (UTI), bacteremia, and meningitis. As physical exam is a poor discriminator of SBI in this age group, providers frequently rely on laboratory values and risk factors to guide management. Infants presenting with documented fevers by caregivers but found to have no fever in the emergency department are a challenge, and there are limited data regarding SBI frequency in this population.
Study overview and results
The authors performed a secondary analysis of a prospectively gathered cohort of infants aged 60 days and younger within the Pediatric Emergency Care Applied Research Network (PECARN) who had blood, urine, and CSF data available. Notable exclusions included infants who were premature, had a focal infection, were clinically ill, had recent antibiotic use, did not have blood, urine, and CSF data available, or were lost to telephone follow-up at 7 days to ensure wellness. The study cohort included 6,014 infants, 1,233 (32%) who were febrile by history alone. Rates of overall SBI were lower in the afebrile group (8.8% vs. 12.8%). For infants 0-28 days, rates of UTI were lower for the afebrile group (9.5% vs. 14.5%), but there was no difference in the rates of bacteremia or meningitis. For infants 29-60 days, rates of UTI (6.6% vs. 9.3%) and bacteremia (.5% vs. 1.7%) were lower in the afebrile group.
Limitations
Neither the use of home antipyretics nor the method of temperature taking at home were studied. Also, as this was a secondary analysis, it is possible that not all infants who presented with history of fever only were captured, as work-up was dictated by individual treating providers who may have chosen not to work up certain afebrile infants.
Important findings and implications
Nearly one-third of infants presenting for fever evaluation are afebrile on arrival. Although overall rates of SBI were lower in the group with fever by history only, this difference is largely accounted for by differing rates of UTI. Rates of bacteremia and meningitis remained substantial between groups, particularly for infants aged 0-28 days. Because of the significant morbidity associated with these infections, it is reasonable to suggest that absence of fever on presentation alone should not alter clinical or laboratory work-up, particularly in infants 0-28 days.
STUDY 4
Humphrey-Murto S et al. The influence of prior performance information on ratings of current performance and implications for learner handover: A scoping review. Acad Med. 2019 Jul;94(7):1050-7.
Background
Learner Handover (LH) or “forward feeding” occurs when information about trainees is shared between faculty supervisors. Although this can be helpful to tailor educational experiences and build upon previous assessments, it risks stigmatizing trainees and adding bias to future feedback and assessments as the trainee never really has a “clean slate.” In this study, the authors sought to uncover the key concepts of how prior performance information (PPI) influences assessments and any implications for medical education.
Study overview and results
The authors performed a cross-disciplinary scoping review looking at over 17,000 articles published between 1980 and 2017 across the domains of psychology, sports, business, and education. Seven themes were identified with the following notable findings. Raters exposed to positive PPI scored a learner’s performance higher, and vice versa. There was a dose-response relationship with more positive and more negative PPI resulting in higher and lower assessments, respectively. General standards, such as a direction to complete all work in a timely manner, caused an assimilation effect, while specific standards, such as a direction to complete a certain task by a certain day, did not. More motivated and more experienced raters are less affected by PPI, and those who believe that people can change (incremental theorists) are less affected by PPI while those who believe personal attributes are fixed (entity theorists) are more affected.
Limitations
The heterogeneity of the studies and the fact that they were largely conducted in experimental settings may limit generalizability to medical education. Slightly less than half of the studies included a control arm. Last, most of the studies looked at the ratings of only one target performance, not multiple performances over time.
Important findings and implications
Ratings of current performance displace toward PPI direction, with negative PPI more influential than positive PPI. In a formative setting, PPI may help the assessor focus on areas of possible weakness. In contrast, for a summative assessment, PPI may be prejudicial and have an impact on the rating given to the student. Clinicians should be mindful of the information they share with future raters about learners and the potential bias on future assessments that can manifest as a result.
