Improving Cohort Definitions in Research Using Hospital Administrative Databases-Do We Need Guidelines?

Sugammadex use in pediatric patients with stage IV-V chronic kidney disease in a quaternary referral hospital: a case series

BACKGROUND

Sugammadex is a pharmacologic agent that provides rapid reversal of neuromuscular blockade via encapsulation of the neuromuscular blocking agent (NMBA). The sugammadex-NMBA complex is primarily cleared through glomerular filtration from the kidney, raising the possibility that alterations in renal function could affect its elimination. In pediatric patients, the benefits of sugammadex have led to widespread utilization; however, there is limited information on its application in pediatric renal impairment. This study examined sugammadex use and postoperative outcomes in pediatric patients with severe chronic renal impairment at our quaternary pediatric referral hospital.

METHODS

After IRB approval, we performed a retrospective analysis in pediatric patients with stage IV and V chronic kidney disease who received sugammadex from January 2017 to March 2022. Postoperative outcomes studied included new or increased respiratory requirement, unplanned intensive care unit (ICU) admission, postoperative pneumonia, anaphylaxis, and death within 48 h postoperatively, unplanned deferral of intraoperative extubation, and repeat administrations of NMBA reversal after leaving the operating room.

RESULTS

The final cohort included 17 patients ranging from 8 months to 16 years old. One patient required new postoperative noninvasive ventilation on postoperative day 2, which was credited to hypervolemia. Another patient had bronchospasm intraoperatively resolving with medication, which could not definitively be associated sugammadex administration. There were no instances of deferred extubation, unplanned ICU or need for supplemental oxygen after tracheal extubation identified.

CONCLUSION

No adverse effects directly attributable to sugammadex in pediatric patients with severe renal impairment were detected. There may be a role for utilization of sugammadex for neuromuscular reversal in this population.

Pediatric Respiratory Syncytial Virus Hospitalizations and Respiratory Support After the COVID-19 Pandemic

Importance

After the COVID-19 pandemic, there was a surge of pediatric respiratory syncytial virus (RSV) infections, but national data on hospitalization and intensive care unit use and advanced respiratory support modalities have not been reported.

Objective

To analyze demographics, respiratory support modes, and clinical outcomes of children with RSV infections at tertiary pediatric hospitals from 2017 to 2023.

Design, Setting, and Participants

This cross-sectional study evaluated children from 48 freestanding US children's hospitals registered in the Pediatric Health Information System (PHIS) database. Patients 5 years or younger with RSV from July 1, 2017, to June 30, 2023, were included. Each season was defined from July 1 to June 30. Prepandemic RSV seasons included 2017 to 2018, 2018 to 2019, and 2019 to 2020. The postpandemic season was delineated as 2022 to 2023.

Exposure

Hospital presentation with RSV infection.

Main Outcomes and Measures

Data on emergency department presentations, hospital or intensive care unit admission and length of stay, demographics, respiratory support use, mortality, and cardiopulmonary resuscitation were analyzed. Postpandemic season data were compared with prepandemic seasonal averages.

Results

A total of 288 816 children aged 5 years or younger (median [IQR] age, 8.9 [3.3-21.5] months; 159 348 [55.2%] male) presented to 48 US children's hospitals with RSV from July 1, 2017, to June 30, 2023. Respiratory syncytial virus hospital presentations increased from 39 698 before the COVID-19 pandemic to 94 347 after the pandemic (P < .001), with 86.7% more hospitalizations than before the pandemic (50 619 vs 27 114; P < .001). In 2022 to 2023, children were older (median [IQR] age, 11.3 [4.1-26.6] months vs 6.8 [2.6-16.8] months; P < .001) and had fewer comorbidities (17.6% vs 21.8% of hospitalized patients; P < .001) than during prepandemic seasons. Advanced respiratory support use increased 70.1% in 2022 to 2023 (9094 vs 5340; P < .001), and children requiring high-flow nasal cannula (HFNC) or noninvasive ventilation (NIV) were older than during prepandemic seasons (median [IQR] age for HFNC, 6.9 [2.7-16.0] months vs 4.6 [2.0-11.7] months; for NIV, 6.0 [2.1-16.5] months vs 4.3 [1.9-11.9] months). Comorbid conditions were less frequent after the pandemic across all respiratory support modalities (HFNC, 14.9% vs 19.1%, NIV, 22.0% vs 28.5%, invasive mechanical ventilation, 30.5% vs 38.0%; P < .001).

Conclusions and Relevance

This cross-sectional study identified a postpandemic pediatric RSV surge that resulted in markedly increased hospital volumes and advanced respiratory support needs in older children with fewer comorbidities than prepandemic seasons. These clinical trends may inform novel vaccine allocation to reduce the overall burden during future RSV seasons.

Comparison of Administrative Database-Derived and Hospital-Derived Data for Monitoring Blood Culture Use in the Pediatric Intensive Care Unit

BACKGROUND

Optimizing blood culture practices requires monitoring of culture use. Collecting culture data from electronic medical records can be resource intensive. Our objective was to determine whether administrative data could serve as a data source to measure blood culture use in pediatric intensive care units (PICUs).

METHODS

Using data from a national diagnostic stewardship collaborative to reduce blood culture use in PICUs, we compared the monthly number of blood cultures and patient-days collected from sites (site-derived) and the Pediatric Health Information System (PHIS, administrative-derived), an administrative data warehouse, for 11 participating sites. The collaborative's reduction in blood culture use was compared using administrative-derived and site-derived data.

RESULTS

Across all sites and months, the median of the monthly relative blood culture rate (ratio of administrative- to site-derived data) was 0.96 (Q1: 0.77, Q3: 1.24). The administrative-derived data produced an estimate of blood culture reduction over time that was attenuated toward the null compared with site-derived data.

CONCLUSIONS

Administrative data on blood culture use from the PHIS database correlates unpredictably with hospital-derived PICU data. The limitations of administrative billing data should be carefully considered before use for ICU-specific data.

Leveraging machine learning to identify acute myeloid leukemia patients and their chemotherapy regimens in an administrative database

BACKGROUND

Administrative datasets are useful for identifying rare disease cohorts such as pediatric acute myeloid leukemia (AML). Previously, cohorts were assembled using labor-intensive, manual reviews of patients' longitudinal chemotherapy data.

METHODS

We utilized a two-step machine learning (ML) method to (i) identify pediatric patients with newly diagnosed AML, and (ii) among the identified AML patients, their chemotherapy courses, in an administrative/billing database. Using 2558 patients previously manually reviewed, multiple ML algorithms were derived from 75% of the study sample, and the selected model was tested in the remaining hold-out sample. The selected model was also applied to assemble a new pediatric AML cohort and further assessed in an external validation, using a standalone cohort established by manual chart abstraction.

RESULTS

For patient identification, the selected Support Vector Machine model yielded a sensitivity of 0.97 and a positive predictive value (PPV) of 0.97 in the hold-out test sample. For course-specific chemotherapy regimen and start date identification, the selected Random Forest model yielded overall PPV greater than or equal to 0.88 and sensitivity greater than or equal to 0.86 across all courses in the test sample. When applied to new cohort assembly, ML identified 3016 AML patients with 10,588 treatment courses. In the external validation subset, PPV was greater than or equal to 0.75 and sensitivity was greater than or equal to 0.82 for patient identification, and PPV was greater than or equal to 0.93 and sensitivity was greater than or equal to 0.94 for regimen identifications.

CONCLUSION

A carefully designed ML model can accurately identify pediatric AML patients and their chemotherapy courses from administrative databases. This approach may be generalizable to other diseases and databases.

Tracheostomy, Feeding-Tube, and In-Hospital Postoperative Mortality in Children: A Retrospective Cohort Study

BACKGROUND

Neuromuscular/neurologic disease confers increased risk of perioperative mortality in children. Some patients require tracheostomy and/or feeding tubes to ameliorate upper airway obstruction or respiratory failure and reduce aspiration risk. Empiric differences between patients with and without these devices and their association with postoperative mortality have not been previously assessed.

METHODS

This retrospective cohort study using the Pediatric Health Information System measured 3- and 30-day in-hospital postsurgical mortality among children 1 month to 18 years of age with neuromuscular/neurologic disease at 44 US children's hospitals, from April 2016 to October 2018. We summarized differences between patients presenting for surgery with and without these devices using standardized differences. Then, we calculated 3- and 30-day mortality among patients with tracheostomy, feeding tube, both, and neither device, overall and stratified by important exposures, using Fisher exact test to test whether differences were significant.

RESULTS

There were 43,193 eligible patients. Unadjusted 3-day mortality was 1.3% (549/43,193); 30-day mortality was 2.7% (1168/43,193). Most (79.1%) used neither a feeding tube or tracheostomy, 1.2% had tracheostomy only, 15.5% had feeding tube only, and 4.2% used both devices. Compared to children with neither device, children using either or both devices were more likely to have multiple CCCs, dysphagia, chronic pulmonary disease, cerebral palsy, obstructive sleep apnea, or malnutrition, and a prolonged intensive care unit (ICU) stay within the previous year. They were less likely to present for high-risk surgeries (33% vs 57%). Having a feeding tube was associated with decreased 3-day mortality overall compared to having neither device (0.9% vs 1.3%, P = .003), and among children having low-risk surgery, and surgery during urgent or emergent hospitalizations. Having both devices was associated with decreased 3-day mortality among children having low-risk surgery (0.8% vs 1.9%; P = .013), and during urgent or emergent hospitalizations (1.6% vs 2.9%; P = .023). For 30-day mortality, having a feeding tube or both devices was associated with lower mortality when the data were stratified by the number of CCCs.

CONCLUSIONS

Patients requiring tracheostomy, feeding tube, or both are generally sicker than patients without these devices. Despite this, having a feeding tube was associated with lower 3-day mortality overall and lower 30-day mortality when the data were stratified by the number of CCCs. Having both devices was associated with lower 3-day mortality in patients presenting for low-risk surgery, and surgery during urgent or emergent hospitalizations.

Sociodemographic Differences of Hospitalization and Associations of Resource Utilization for Failure to Thrive

OBJECTIVES

This study examines the sociodemographic differences between elective and nonelective admissions for failure to thrive (FTT). We investigate associations between admission type and hospital resource utilization, including length of stay and feeding tube placement.

METHODS

We included children <2 years old with FTT in the nationwide Kids' Inpatient Database. We described differences between elective and nonelective admissions using Fisher exact and t tests. To assess associations of admission type and hospital resource utilization, we used negative binomial and logistic regression for length of stay and feeding tube placement, respectively.

RESULTS

In this study of 45,920 admissions (37,224 nonelective vs 8696 elective), we found differences by race and ethnicity, income, and insurance type, among other factors. Compared to elective admissions, nonelective admissions had higher proportions of infants who were Black, Hispanic, and of lower-income. Nonelective admissions were associated with longer lengths of stay (incidence rate ratio 1.46; 95% CI: 1.37-1.55), independent of child age, sex, neighborhood income, insurance, admission day, chronic conditions, and location. Nonelective admissions were associated with lower odds of feeding tube placement compared to elective admissions (adjusted odds ratio 0.62; 0.56-0.68). In the stratified analyses, children of racial and ethnic minority groups admitted nonelectively versus electively had relatively higher odds of feeding tube placement, while White children had relatively lower odds of feeding tube placement.

CONCLUSION

There are various sociodemographic differences between elective and nonelective FTT admissions. Future research is warranted to elucidate drivers of these differences, particularly those related to racial and ethnic disparities and structural racism.

A Pediatric Hospital Medicine Primer for Performing Research Using Administrative Data

Provider- and claims-focused administrative databases are powerful tools for conducting health services research, and these studies often have good generalizability owing to diversity of hospitals from which samples are derived. In this research methods article, we describe administrative data and how available provider- and claims-focused administrative databases can be used to conduct health services research. We describe common observational study designs using administrative data and provide real-world examples. We highlight the strengths and weaknesses of studies conducted using administrative data and describe methodological considerations to reduce bias and improve the rigor of observational studies using administrative data. Finally, we provide guidance on the types of study questions suitable for observational study designs using administrative data.