Resource utilization for observation-status stays at children's hospitals

Hospital costs and reimbursement for short-stay inpatient versus observation status hospitalizations for children with medical complexity

BACKGROUND

There is a lack of uniformity across hospitals in applying inpatient versus observation status for short-stay (<48 h) pediatric hospitalizations, with negative financial implications associated with observation. Children with medical complexity (CMC) represent a growing population and incur high costs of care. The financial implications of inpatient and observation status for CMC have not been studied.

OBJECTIVES

To compare costs and reimbursement for short-stay hospitalizations for CMC by inpatient and observation status, overall and stratified by payor.

METHODS

We performed a cohort study of short-stay hospitalizations for CMC from 2016 to 2021 at 10 children's hospitals reporting reimbursement in the Pediatric Health Information System and Revenue Management Program. The primary outcome was the cost coverage ratio (CCR), defined as an encounter's reimbursement divided by the estimated cost.

RESULTS

There were 89,282 encounters included. The median costs per encounter were similar across observation ($5206, IQR $3604-$7484) and inpatient ($6547, IQR $4725-$9349) encounters. For government payors, the median CCR was 0.6 (IQR 0.2-0.9) for observation encounters and 1.2 (IQR 0.8-1.9) for inpatient. For nongovernment payors, the median CCR was 1.6 (IQR 1.3-1.9) for observation and 1.6 (IQR 1.4-2) for inpatient. Government reimbursement was associated with increased risk for financial loss (OR 13.91, 95% CI 7.23, 26.77) and with a median net loss of $985,952 (IQR $389,871-$1,700,041) per hospital annually for observation encounters.

CONCLUSIONS

Government-paid observation encounters for CMC are associated with significant financial loss at children's hospitals. This reimbursement model may pose a threat to children's hospitals' ability to care for CMC.

Pediatric Hospitalizations at Rural and Urban Teaching and Nonteaching Hospitals in the US, 2009-2019

Importance

National analyses suggest that approximately 1 in 5 US hospitals closed their pediatric units between 2008 and 2018. The extent to which pediatric hospitalizations at general hospitals in rural and urban communities decreased during this period is not well understood.

Objective

To describe changes in the number and proportion of pediatric hospitalizations and costs at urban teaching, urban nonteaching, and rural hospitals vs freestanding children's hospitals from 2009 to 2019; to estimate the number and proportion of hospitals providing inpatient pediatric care; and to characterize changes in clinical complexity.

Design, Setting, and Participants

This study is a retrospective cross-sectional analysis of the 2009, 2012, 2016, and 2019 Kids' Inpatient Database, a nationally representative data set of US pediatric hospitalizations among children younger than 18 years. Data were analyzed from February to June 2023.

Exposures

Pediatric hospitalizations were grouped as birth or nonbirth hospitalizations. Hospitals were categorized as freestanding children's hospitals or as rural, urban nonteaching, or urban teaching general hospitals.

Main Outcomes and Measures

The primary outcomes were annual number and proportion of birth and nonbirth hospitalizations and health care costs, changes in the proportion of hospitalizations with complex diagnoses, and estimated number and proportion of hospitals providing pediatric care and associated hospital volumes. Regression analyses were used to compare health care utilization in 2019 vs that in 2009.

Results

The data included 23.2 million (95% CI, 22.7-23.6 million) weighted hospitalizations. From 2009 to 2019, estimated national annual pediatric hospitalizations decreased from 6 425 858 to 5 297 882, as birth hospitalizations decreased by 10.6% (95% CI, 6.1%-15.1%) and nonbirth hospitalizations decreased by 28.9% (95% CI, 21.3%-36.5%). Concurrently, hospitalizations with complex chronic disease diagnoses increased by 45.5% (95% CI, 34.6%-56.4%), and hospitalizations with mental health diagnoses increased by 78.0% (95% CI, 61.6%-94.4%). During this period, the most substantial decreases were in nonbirth hospitalizations at rural hospitals (4-fold decrease from 229 263 to 62 729) and urban nonteaching hospitals (6-fold decrease from 581 320 to 92 118). In 2019, birth hospitalizations occurred at 2666 hospitals. Nonbirth pediatric hospitalizations occurred at 3507 hospitals, including 1256 rural hospitals and 843 urban nonteaching hospitals where the median nonbirth hospitalization volumes were fewer than 25 per year.

Conclusions and Relevance

Between 2009 and 2019, the largest decreases in pediatric hospitalizations occurred at rural and urban nonteaching hospitals. Clinical and policy initiatives to support hospitals with low pediatric volumes may be needed to maintain hospital access and pediatric readiness, particularly in rural communities.

Annual Variation in 30-Day Risk-Adjusted Readmission Rates in U.S. Children's Hospitals

OBJECTIVE

Reducing pediatric readmissions has become a national priority; however, the use of readmission rates as a quality metric remains controversial. The goal of this study was to examine short-term stability and long-term changes in hospital readmission rates.

METHODS

Data from the Pediatric Health Information System were used to compare annual 30-day risk-adjusted readmission rates (RARRs) in 47 US children's hospitals from 2016 to 2017 (short-term) and 2016 to 2019 (long-term). Pearson correlation coefficients and weighted Cohen's Kappa statistics were used to measure correlation and agreement across years for hospital-level RARRs and performance quartiles.

RESULTS

Median (IQR) 30-day RARRs remained stable from 7.7% (7.0-8.3) in 2016 to 7.6% (7.0-8.1) in 2019. Individual hospital RARRs in 2016 were strongly correlated with the same hospital's 2017 rate (R2 = 0.89 [95% confidence interval (CI) 0.80-0.94]) and moderately correlated with those in 2019 (R2 = 0.49 [95%CI 0.23-0.68]). Short-term RARRs (2016 vs 2017) were more highly correlated for medical conditions than surgical conditions, but correlations between long-term medical and surgical RARRs (2016 vs 2019) were similar. Agreement between RARRs was higher when comparing short-term changes (0.73 [95%CI 0.59-0.86]) than long-term changes (0.45 [95%CI 0.27-0.63]). From 2016 to 2019, RARRs increased by ≥1% in 7 (15%) hospitals and decreased by ≥1% in 6 (13%) hospitals. Only 7 (15%) hospitals experienced reductions in RARRs over the short and long-term.

CONCLUSIONS

Hospital-level performance on RARRs remained stable with high agreement over the short-term suggesting stability of readmission measures. There was little evidence of sustained improvement in hospital-level performance over multiple years.

Ambulatory follow-up among publicly insured children discharged from the emergency department

BACKGROUND

While children discharged from the emergency department (ED) are frequently advised to follow up with ambulatory care providers, the extent to which this occurs is unknown. We sought to characterize the proportion of publicly insured children who have an ambulatory visit following ED discharge, identify factors associated with ambulatory follow-up, and evaluate the association of ambulatory follow-up with subsequent hospital-based health care utilization.

METHODS

We performed a cross-sectional study of pediatric (<18 years) encounters during 2019 included in the IBM Watson Medicaid MarketScan claims database from seven U.S. states. Our primary outcome was an ambulatory follow-up visit within 7 days of ED discharge. Secondary outcomes were 7-day ED return visits and hospitalizations. Logistic regression and Cox proportional hazards were used for multivariable modeling.

RESULTS

We included 1,408,406 index ED encounters (median age 5 years, IQR 2-10 years), for which a 7-day ambulatory visit occurred in 280,602 (19.9%). Conditions with the highest proportion of 7-day ambulatory follow-up included seizures (36.4%); allergic, immunologic, and rheumatologic diseases (24.6%); other gastrointestinal diseases (24.5%); and fever (24.1%). Ambulatory follow-up was associated with younger age, Hispanic ethnicity, weekend ED discharge, ambulatory encounters prior to the ED visit, and diagnostic testing performed during the ED encounter. Ambulatory follow-up was inversely associated with Black race and ambulatory care-sensitive or complex chronic conditions. In Cox models, ambulatory follow-up was associated with a higher hazard ratio (HR) of subsequent ED return (HR range 1.32-1.65) visit and hospitalization (HR range 3.10-4.03).

CONCLUSIONS

One-fifth of children discharged from the ED have an ambulatory visit within 7 days, which varied by patient characteristics and diagnoses. Children with ambulatory follow-up have a greater subsequent health care utilization, including subsequent ED visit and/or hospitalization. These findings identify the need to further research the role and costs associated with routine post-ED visit follow-up.

Bacterial resistance to commonly prescribed antibiotics was rare among children hospitalised for clinical dysentery

AIM

Temporal changes in common pathogens that cause clinical dysentery have been described in Europe. We aimed to describe the distribution of pathogens and their antibiotic resistance in hospitalised Israeli children.

METHODS

This study retrospectively studied children hospitalised for clinical dysentery, with or without a positive stool culture, from 1 January 2016 to 31 December 2019.

RESULTS

We diagnosed 137 patients (65% males), with clinical dysentery at a median age of 3.7 (interquartile range 1.5-8.2) years. Stools were cultured in 135 patients (99%), and the results were positive in 101 (76%). These comprised Campylobacter (44%), Shigella sonnei (27%), non-typhoid Salmonella (18%) and enteropathogenic Escherichia coli (12%). Only one of the 44 Campylobacter cultures was resistant to erythromycin and one of the 12 enteropathogenic Escherichia coli cultures was resistant to ceftriaxone. None of the Salmonella and Shigella cultures were resistant to ceftriaxone or erythromycin. We did not find any pathogens that were associated with a typical clinical presentation or laboratory results on admission.

CONCLUSION

The most common pathogen was Campylobacter, in line with recent European trends. Bacterial resistance for commonly prescribed antibiotics was rare, and these findings support the current European recommendations.

Hospitalization Risk Associated With Emergency Department Reasons for Visit and Patient Age: A Retrospective Evaluation of National Emergency Department Survey Data to Help Identify Potentially Avoidable Emergency Department Visits

Background

Patients often present to emergency departments (EDs) with concerns that do not require emergency care. Self-triage and other interventions may help some patients decide whether they should be seen in the ED. Symptoms associated with low risk of hospitalization can be identified in national ED data and can inform the design of interventions to reduce avoidable ED visits.

Methods

We used the National Hospital Ambulatory Medical Care Survey (NHAMCS) data from the United States National Health Care Statistics (NHCS) division of the Centers for Disease Control and Prevention (CDC). The ED datasets from 2011 through 2020 were combined. Primary reasons for ED visit and the binary field for hospital admission from the ED were used to estimate the proportion of ED patients admitted to the hospital for each reason for visit and age category.

Results

There were 221,027 surveyed ED visits during the 10-year data collection with 736 different primary reasons for visit and 23,228 hospitalizations. There were 145 million estimated hospitalizations from 1.37 billion estimated ED visits (10.6%). Inclusion criteria for this study were reasons for visit which had at least 30 ED visits in the sample; there were 396 separate reasons for visit which met this criteria. Of these 396 reasons for visit, 97 had admission percentages less than 2% and another 52 had hospital admissions estimated between 2% and 4%. However, there was a significant increase in hospitalizations within many of the ED reasons for visit in older adults.

Conclusion

Reasons for visit from national ED data can be ranked by hospitalization risk. Low-risk symptoms may help healthcare institutions identify potentially avoidable ED visits. Healthcare systems can use this information to help manage potentially avoidable ED visits with interventions designed to apply to their patient population and healthcare access.

Pediatric emergency departments' readiness for change toward improving suicide prevention: A mixed-methods study with US leaders

Objective

To assess pediatric emergency departments' (PEDs) current suicide prevention practices and climate for change to improve suicide prevention for youth.

Methods

We conducted an explanatory, sequential mixed-methods study. First, we deployed a national, cross-sectional survey of PED leaders identified through publicly available data in Fall 2020, and then we conducted follow-up interviews with those who expressed interest. The survey queried each PED's suicide prevention practices and measured readiness for change to improve suicide prevention practices using questions scored on a 5-point Likert scale. Interviews gathered further, in-depth descriptions of PEDs' practices and culture. Interviews were audio-recorded, transcribed verbatim, and analyzed using a rapid analysis approach.

Results

Of 135 PED directors eligible to complete the survey, 64 responded (response rate 47%). A total of 64% of PEDs had a mental health specialist available 24 hours/day, 7 days/week; 80% reported practicing mental health disposition planning, and 41% reported practicing psychiatric medication management. Altogether 91% of directors agreed or strongly agreed that their PED had a positive culture and 92% agreed/strongly agreed that their PED was ready for change. However, 31% disagreed/strongly disagreed that their PED had tools for evaluation and quality measurement. Resources needed for change (including budget, staffing, training, and facilities) varied across institutions. Interviews with our convenience sample of 21 directors revealed varying suicide prevention practices and confirmed that standardization, evaluation, and quality improvement initiatives were needed at most institutions. Leaders reported a high interest in improving care.

Conclusions

PED leaders reported high motivation to improve suicide prevention services for young people, and reported needing quality improvement infrastructure to monitor and guide improvement.

Pediatric Return Visits to the Emergency Department: The Time to Return Curve

OBJECTIVE

Although 72-hour return visits are a frequently reported metric for pediatric patients discharged from the emergency department (ED), the basis for this metric is not established. Our objective was to statistically derive a cutoff time point for the characterization of pediatric return visits.

METHODS

We performed a retrospective cohort study using data of patients discharged from any of 44 pediatric EDs. We selected the first encounter per patient from January 1 to December 31, 2019, as the index encounter and included the first return visit within 30 days. We constructed a cumulative hazard curve to characterize the timing of return visits and constructed a multivariable adaptive regression spline model to identify a hinge point in return visit presentations. We identified the association between admission for early return visits and admission for late return visits using generalized linear mixed modeling.

RESULTS

Of 1,986,778 index ED discharges, 193,605 (9.7%) ED return visits were included. A double-exponential decay model demonstrated superior fit compared with a single exponential model ( P < 0.0001). Multivariable adaptive regression spline modeling identified a hinge at 7 days. When comparing proportions of return visits leading to hospitalization between early (23.8%) and late (15.1%) return visits, early visits (≤7 days) had higher adjusted odds of hospital admission (adjusted odds ratio, 1.73; 95% confidence interval, 1.69-1.77) relative to late return visits (>7 days). Findings were similar in sensitivity analyses within age subgroups, Census region, and in which the diagnosis (using the Diagnosis and Grouping System) was the same between the index and return visit. Among return visits that occurred within 7 days of the index visit, 46.3% had the same diagnosis grouping in both visits.

CONCLUSIONS

An empirically derived 7-day cutoff may be more appropriate for characterization of pediatric return visits to the ED. Encounters after this period had lower adjusted odds of admission.

Comparison of crosswalk methods for translating ICD-9 to ICD-10 diagnosis codes for child maltreatment

BACKGROUND

A validated source of International Classification of Disease (1CD) 10th revision diagnostic codes to identify child maltreatment has not been developed. Such a reference would be essential for the practical purposes of administrative data-based research and public health surveillance.

OBJECTIVE

To evaluate the validity of diagnosis code classifications for child maltreatment following conversion from ICD 9th edition, clinical modification (ICD-9-CM) to 10th revision.

PARTICIPANTS AND SETTING

Children receiving inpatient or emergency medical care in the United States with ICD-9-CM/ICD-10-CM diagnosis codes for child maltreatment, identified using two large multicenter hospital-based datasets.

METHODS

We evaluated the performance of general equivalence mappings (GEMs) and network-based mappings for previously-validated ICD-9-CM diagnosis codes for child maltreatment in the 2013-2014 PHIS and 2012 KID and resulting ICD-10-CM codes in the 2018-2019 PHIS and 2016 KID datasets.

RESULTS

Of 56 previously-validated ICD-9-CM diagnoses, GEMs identified 15 with a similar proportion of diagnosed children in the KID ICD-9-CM and ICD-10 eras and 14 diagnoses in PHIS. Network-based mapping identified 18 diagnoses with similar proportions in the KID datasets, and 13 diagnoses in PHIS. For six diagnoses, the proportion of children identified in the ICD-10 era using network-based mapping was more than ten times the proportion identified in the ICD-9-CM era.

CONCLUSION

Neither crosswalk method provided consistently reliable conversions, due to both crosswalk methodology and changes introduced by the ICD 10th revision. These findings highlight the need for independent construction and validation of ICD-10-based definitions of child maltreatment as a precursor to administrative data-based research and public health surveillance.

Financial Implications of Short Stay Pediatric Hospitalizations

BACKGROUND

Observation status (OBS) stays incur similar costs to low-acuity, short-stay inpatient (IP) hospitalizations. Despite this, payment for OBS is likely less and may represent a financial liability for children's hospitals. Thus, we described the financial outcomes associated with OBS stays compared to similar IP stays by hospital and payer.

METHODS

We conducted a retrospective cohort study of clinically similar pediatric OBS and IP encounters at 15 hospitals contributing to the revenue management program in 2017. Clinical and demographic characteristics were described. For each hospitalization, the cost coverage ratio (CCR) was calculated by dividing revenue by estimated cost of hospitalization. Differences in CCR were evaluated using Wilcoxon rank sum tests and results were stratified by billing designation and payer. CCR for OBS and IP stays were compared by institution, and the estimated increase in revenue by billing OBS stays as IP was calculated.

RESULTS

OBS was assigned to 70 981 (56.9%) of 124 789 hospitalizations. Use of OBS varied across hospitals (8%-86%). For included hospitalizations, OBS stays were more likely than IP stays to result in financial loss (57.0% vs 35.7%). OBS stays paid by public payer had the lowest median CCR (0.6; interquartile range [IQR], 0.2-0.9). Paying OBS stays at the median IP rates would have increased revenue by $167 million across the 15 hospitals.

CONCLUSIONS

OBS stays were significantly more likely to result in poor financial outcomes than similar IP stays. Costs of hospitalization and billing designations are poorly aligned and represent an opportunity for children's hospitals and payers to restructure payment models.

Machine learning methods applied to triage in emergency services: A systematic review

BACKGROUND

In emergency services is important to accurately assess and classify symptoms, which may be improved with the help of technology. One mechanism that could help and improve predictions from health records or patient flow is machine learning (ML).

AIM

To analyse the effectiveness of ML systems in triage for making predictions at the emergency department in comparison with other triage scales/scores.

METHODS

Following the PRISMA recommendations, a systematic review was conducted using CINAHL, Cochrane, Cuiden, Medline and Scopus databases with the search equation "Machine learning AND triage AND emergency".

RESULTS

Eleven studies were identified. The studies show that the use of ML methods consistently predict important outcomes like mortality, critical care outcomes and admission, and the need for hospitalization in comparison with scales like Emergency Severity Index or others. Among the ML models considered, XGBoost and Deep Neural Networks obtained the highest levels of prediction accuracy, while Logistic Regression performed obtained the worst values.

CONCLUSIONS

Machine learning methods can be a good instrument for helping triage process with the prediction of important emergency variables like mortality or the need for critical care or hospitalization.

Trends and Variation in the Use of Observation Stays at Children's Hospitals

BACKGROUND

Observation status could improve efficiency of healthcare resource use but also might shift financial burdens to patients and hospitals. Although the use of observation stays has increased for adult patient populations, the trends are unknown among hospitalized children.

OBJECTIVE

The goal of this study was to describe recent trends in observation stays for pediatric populations at children's hospitals.

DESIGN, SETTING, AND PARTICIPANTS

Both observation and inpatient stays for all conditions were retrospectively studied using the Pediatric Health Information System database (2010 to 2019).

EXPOSURE, MAIN OUTCOMES, AND MEASURES

Patient type was classified as inpatient or observation status. Main outcomes included annual percentage of observation stays, annual percentage of observation stays having prolonged length of stay (>2 days), and growth rates of observation stays for the 20 most common conditions. Risk adjusted hospital-level use of observation stays was estimated using generalized linear mixed-effects models.

RESULTS

The percentage of observation stays increased from 23.6% in 2010 to 34.3% in 2019 (P < .001), and the percentage of observation stays with prolonged length of stay rose from 1.1% to 4.6% (P < .001). Observation status was expanded among a diverse group of clinical conditions; diabetes mellitus and surgical procedures showed the highest growth rates. Adjusted hospital-level use ranged from 0% to 67% in 2019, indicating considerable variation among hospitals.

CONCLUSION

Based on the increase in observation stays, future studies should explore the appropriateness of observation care related to efficient use of healthcare resources and financial implications for hospitals and patients.

Pediatric patients with dog bites presenting to US children's hospitals

Background

To characterize pediatric dog bite injuries presenting to US children’s hospitals and identify factors associated with clinically significant injuries.

Methods

We performed a multicenter observational study from 26 pediatric hospitals between July 1, 2010, and June 30, 2020, including patients ≤ 18 years with dog bites, consolidating together encounters from patients with multiple encounters within 30 days as a single episode of care. We characterized diagnoses and procedures performed in these patients. We used generalized linear mixed models to identify factors associated with a composite outcome that we term clinically significant injuries (defined as admission, operating room charge, sedation, fractures/dislocations, intracranial/eye injury, skin/soft tissue infection, or in-hospital mortality).

Results

68,833 episodes were included (median age 6.6 years [interquartile range 3.5–10.4 years], 55.5% male) from 67,781 patients. We identified 16,502 patients (24.0%) with clinically significant injuries, including 6653 (9.7%) admitted, 5080 (7.4%) managed in the operating room, 11,685 (17.0%) requiring sedation, 493 (0.7%) with a skull fracture, 32 (0.0%) with a fracture in the neck or trunk, 389 (0.6%) with a fracture of the upper limb, 51 (0.1%) with a fracture in the lower limb, 15 (0.0%) with dislocations, 66 (0.1%) with an intracranial injury and 164 (0.2%) with an injury to the eyeball, 3708 (5.4%) with skin/soft tissue infections, and 5 (0.0%) with in-hospital mortality. In multivariable analysis, younger age (0–4 years, 5–9 years, and 10–14 years relative to 15–18 years), weekday injuries, and an income in the second and third quartiles (relative to the lowest quartile) had higher odds of clinically significant injuries. Black patients (relative to White), Hispanic/Latino ethnicity, and private insurance status (relative to public insurance) had lower odds of clinically important injuries. When evaluating individual components within the composite outcome, most followed broader trends.

Conclusion

Dog bites are an important mechanism of injury encountered in children’s hospitals. Using a composite outcome measure, we identified younger, White, non-Hispanic children at higher risk of clinically significant injuries. Findings with respect to race and ethnicity and dog bite injuries warrant further investigation. Results can be used to identify populations for targeted prevention efforts to reduce severe dog bite injuries.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40621-021-00349-3.

Forecast modeling to identify changes in pediatric emergency department utilization during the COVID-19 pandemic

Objective

To identify trends in pediatric emergency department (ED) utilization following the COVID-19 pandemic.

Methods

We performed a cross-sectional study from 37 geographically diverse US children's hospitals. We included ED encounters between January 1, 2010 and December 31, 2020, transformed into time-series data. We constructed ensemble forecasting models of the most common presenting diagnoses and the most common diagnoses leading to admission, using data from 2010 through 2019. We then compared the most common presenting diagnoses and the most common diagnoses leading to admission in 2020 to the forecasts.

Results

29,787,815 encounters were included, of which 1,913,085 (6.4%) occurred during 2020. ED encounters during 2020 were lower compared to prior years, with a 65.1% decrease in April relative to 2010–2019. In forecasting models, encounters for depression and diabetic ketoacidosis remained within the 95% confidence interval [CI]; fever, bronchiolitis, hyperbilirubinemia, skin/subcutaneous infections and seizures occurred within the 80–95% CI during the portions of 2020, and all other diagnoses (abdominal pain, otitis media, asthma, pneumonia, trauma, upper respiratory tract infections, and urinary tract infections) occurred below the predicted 95% CI.

Conclusion

Pediatric ED utilization has remained low following the COVID-19 pandemic, and below forecasted utilization for most diagnoses. Nearly all conditions demonstrated substantial declines below forecasted rates from the prior decade and which persisted through the end of the year. Some declines in non-communicable diseases may represent unmet healthcare needs among children. Further study is warranted to understand the impact of policies aimed at curbing pandemic disease on children.

Observation Status Stays With Low Resource Use Within Children's Hospitals

BACKGROUND

High costs associated with hospitalization have encouraged reductions in unnecessary encounters. A subset of observation status patients receive minimal interventions and incur low use costs. These patients may contain a cohort that could safely be treated outside of the hospital. Thus, we sought to describe characteristics of low resource use (LRU) observation status hospitalizations and variation in LRU stays across hospitals.

METHODS

We conducted a retrospective cohort study of pediatric observation encounters at 42 hospitals contributing to the Pediatric Health Information System database from January 1, 2019, to December 31, 2019. For each hospitalization, we calculated the use ratio (nonroom costs to total hospitalization cost). We grouped stays into use quartiles with the lowest labeled LRU. We described associations with LRU stays and performed classification and regression tree analyses to identify the combination of characteristics most associated with LRU. Finally, we described the proportion of LRU hospitalizations across hospitals.

RESULTS

We identified 174 315 observation encounters (44 422 LRU). Children <1 year (odds ratio [OR] 3.3; 95% confidence interval [CI] 3.1-3.4), without complex chronic conditions (OR 3.6; 95% CI 3.2-4.0), and those directly admitted (OR 4.2; 95% CI 4.1-4.4) had the greatest odds of experiencing an LRU encounter. Those children with the combination of direct admission, no medical complexity, and a respiratory diagnosis experienced an LRU stay 69.5% of the time. We observed variation in LRU encounters (1%-57% of observation encounters) across hospitals.

CONCLUSIONS

LRU observation encounters are variable across children's hospitals. These stays may include a cohort of patients who could be treated outside of the hospital.

Costs and Reimbursements for Mental Health Hospitalizations at Children's Hospitals

The financial impact of the rising number of pediatric mental health hospitalizations is unknown. Therefore, this study assessed costs, reimbursements, and net profits or losses for 111,705 mental health and non-mental health medical hospitalizations in children's hospitals with use of the Pediatric Health Information System and Revenue Management Program. Average financial margins were calculated as (reimbursement per day) - (cost per day), and they were lowest for mental health hospitalizations ($136/day), next lowest for suicide attempt ($518/day), and highest for other medical hospitalizations ($611/day). For 10 of 17 hospitals, margin per day for mental health hospitalizations was lower than margin per day for other medical hospitalizations. For these 10 hospitals, the total net loss for inpatient and observation status mental health hospitalizations, compared with other medical hospitalizations, was $27 million (median, $2.2 million per hospital). Financial margins were usually lower for mental health vs non-mental health medical hospitalizations.

Observation Encounters and Length of Stay Benchmarking in Children's Hospitals

BACKGROUND AND OBJECTIVES

Length of stay (LOS) is a common benchmarking measure for hospital resource use and quality. Observation status (OBS) is considered an outpatient service despite the use of the same facilities as inpatient status (IP) in most children's hospitals, and LOS calculations often exclude OBS stays. Variability in the use of OBS by hospitals may significantly impact calculated LOS. We sought to determine the impact of including OBS in calculating LOS across children's hospitals.

METHODS

Retrospective cohort study of hospitalized children (age <19 years) in 2017 from the Pediatric Health Information System (Children's Hospital Association, Lenexa, KS). Normal newborns, transfers, deaths, and hospitals not reporting LOS in hours were excluded. Risk-adjusted geometric mean length of stay (RA-LOS) for IP-only and IP plus OBS was calculated and each hospital was ranked by quintile.

RESULTS

In 2017, 45 hospitals and 625 032 hospitalizations met inclusion criteria (IP = 410 731 [65.7%], OBS = 214 301 [34.3%]). Across hospitals, OBS represented 0.0% to 60.3% of total discharges. The RA-LOS (SD) in hours for IP and IP plus OBS was 75.2 (2.6) and 54.3 (2.7), respectively (P < .001). For hospitals reporting OBS, the addition of OBS to IP RA-LOS calculations resulted in a decrease in RA-LOS compared with IP encounters alone. Three-fourths of hospitals changed ≥1 quintile in LOS ranking with the inclusion of OBS.

CONCLUSIONS

Children's hospitals exhibit significant variability in the assignment of OBS to hospitalized patients and inclusion of OBS significantly impacts RA-LOS calculations. Careful consideration should be given to the inclusion of OBS when determining RA-LOS for benchmarking, quality and resource use measurements.

Hospital Observation Status and Readmission Rates

BACKGROUND

In several states, payers penalize hospitals when an inpatient readmission follows an inpatient stay. Observation stays are typically excluded from readmission calculations. Previous studies suggest inconsistent use of observation designations across hospitals. We sought to describe variation in observation stays and examine the impact of inclusion of observation stays on readmission metrics.

METHODS

We conducted a retrospective cohort study of hospitalizations at 50 hospitals contributing to the Pediatric Health Information System database from January 1, 2018, to December 31, 2018. We examined prevalence of observation use across hospitals and described changes to inpatient readmission rates with higher observation use. We described 30-day inpatient-only readmission rates and ranked hospitals against peer institutions. Finally, we included observation encounters into the calculation of readmission rates and evaluated hospitals' change in readmission ranking.

RESULTS

Most hospitals (n = 44; 88%) used observation status, with high variation in use across hospitals (0%-53%). Readmission rate after index inpatient stay (6.8%) was higher than readmission after an index observation stay (4.4%), and higher observation use by hospital was associated with higher inpatient-only readmission rates. When compared with peers, hospital readmission rank changed with observation inclusion (60% moving at least 1 quintile).

CONCLUSIONS

The use of observation status is variable among children's hospitals. Hospitals that more liberally apply observation status perform worse on the current inpatient-to-inpatient readmission metric, and inclusion of observation stays in the calculation of readmission rates significantly affected hospital performance compared with peer institutions. Consideration should be given to include all admission types for readmission rate calculation.

To Be or Not to Be (Inpatient Versus Observation): Improving Admission-Status Assignment

BACKGROUND

Observation care is frequently indistinguishable from inpatient care. However, the financial burden of inappropriate status assignment for hospitals and patients can be large. Increased awareness of the potential for financial hardships experienced by patients because of status designation spurred interest among physicians in this improvement project. The goal was to improve the percentage of appropriate inpatient-status assignments from 76% to 90% in 2 years and eliminate observation assignments for patients with hospitalizations >48 hours.

METHODS

Our multidisciplinary team used the Model for Improvement. Interventions included securing a lead physician advisor to the use-review team, improving the process for status review and adjustment, and creating educational sessions and tools for physicians. Data collected included the percentage of appropriate inpatient assignments, percentage of observation assignments for patients with hospitalizations >48 hours, write-off dollar amount per year from denial of payment due to payer disagreement with inpatient status, and resident physician confidence in assigning status.

RESULTS

Appropriate use of inpatient assignments increased from 76% to 84%. Status assignments remaining in observation >48 hours of hospital length of stay decreased by one-half, from 6% to 3%. The write-off dollar amount increased during the study period but decreased by 19% the following calendar year, 2018. Resident self-reported confidence in status designation increased after educational sessions.

CONCLUSIONS

Careful selection of admission status by educated providers and a system to identify relevant cases for status changes can increase appropriate status assignment and, potentially, positively affect the economic burden placed on patients and hospitals.

Care of Pediatric High-Cost Hospitalizations Across Hospital Types

BACKGROUND

High-cost hospitalizations (HCHs) account for a substantial proportion of pediatric health care expenditures. We aimed to (1) describe the distribution of pediatric HCHs across hospital types caring for children and (2) compare characteristics of pediatric HCHs by hospital type.

METHODS

Cross-sectional analysis of all pediatric hospitalizations in the 2012 Kids' Inpatient Database. HCHs were defined as costs >$40 000 (94th percentile). Hospitals were categorized as children's, small general, and large general.

RESULTS

Approximately 166 000 HCHs were responsible for 50.8% of aggregate hospital costs ($18.1 of $35.7 billion) and were mostly at children's hospitals (65%). Children with an HCH were largely neonates (45%), had public insurance (50%), and had ≥1 chronic condition (74%). A total of 131 children's hospitals cared for a median of 559 HCHs per hospital (interquartile range [IQR]: 355-1153) compared to 76 HCHs per hospital (IQR: 32-151) at 397 large general hospitals and 5 HCHs per hospital (IQR: 2-22) at 3581 small general hospitals. The median annual aggregate cost for HCHs was $60 million (IQR: $36-$135) per children's hospital compared to $6.6 million (IQR: $2-$15) per large general hospital and $300 000 (IQR: $116 000-$1.5 million) per small general hospital. HCHs from children's hospitals encompassed nearly 5 times as many unique clinical conditions as large general hospitals and >30 times as many as small general hospitals.

CONCLUSIONS

Children's hospitals cared for a disproportionate volume, cost, and diversity of HCHs compared to general hospitals. Future studies should characterize the factors driving cost, resources, and reimbursement practices for HCH to ensure the long-term financial viability of the pediatric health care system.

Mother's Own Milk Feeding and Severity of Respiratory Illness in Acutely Ill Children: An Integrative Review

PROBLEM

Breastfed infants experience less severe infections while actively breastfeeding. However, little is known about whether a history of prior breastfeeding affects severity of illness. Therefore, the purpose of this integrative review was to examine the relationship between previous exposure to mother's own milk (MOM) feeding and severity of respiratory infectious illness in infants and children.

ELIGIBILITY CRITERIA

Studies meeting the following criteria were included: human subjects, term birth, ages 0-35 months at time of study, diagnosis of pneumonia, bronchiolitis or croup, MOM feeding, and statistical analyses reporting separate respiratory infectious illness outcomes when combined with other infections.

SAMPLE

Twelve articles met eligibility criteria.

RESULTS

Major findings were inconsistent definitions of both dose and exposure period of breastfeeding and the severity of illness. In particular, the severity of illness measure was limited by the use of proxy variables such as emergency room visits or hospitalizations that lacked reliability and validity. However, given this limitation, the data suggested that exclusive breastfeeding for four to six months was associated with reduced severity of illness as measured by frequency of visits to the primary care provider office, emergency department or hospitalization.

CONCLUSIONS

Future research in this area should incorporate reliable and valid measures of MOM dose and exposure period and severity of illness outcomes in the critically ill child.

IMPLICATIONS

Among many positive outcomes associated with breastfeeding, an additional talking point for encouragement of exclusive breastfeeding for four to six months may be protective against severe respiratory infectious illness after cessation of breastfeeding.

Marked Variation Exists Among Surgeons and Hospitals in the Use of Secondary Cleft Lip Surgery

OBJECTIVE

To identify child-, surgeon-, and hospital-specific factors at the time of primary cleft lip repair that are associated with the use of secondary cleft lip surgery.

DESIGN

Retrospective cohort study.

SETTING

Forty-nine pediatric hospitals.

PARTICIPANTS

Children who underwent cleft lip repair between 1999 and 2015.

MAIN OUTCOME MEASURE

Time from primary cleft lip repair to secondary lip surgery.

RESULTS

By 5 years after primary lip repair, 24.0% of children had undergone a secondary lip surgery. In multivariable analysis, primary lip repair before 3 months had a 1.22-fold increased hazard of secondary surgery (95% confidence interval [CI]: 1.02-1.46) compared to repair at 7 to 12 months of age, and children with multiple congenital anomalies had a 0.77-fold decreased hazard of secondary surgery (95% CI: 0.68-0.87). After adjusting for cleft type, age at repair, presence of multiple congenital anomalies, and procedure volume, there remained substantial variation in secondary surgery use among surgeons and hospitals (P < .01). For children with unilateral cleft lip repaired at 3 to 6 months of age, the predicted proportion of children undergoing secondary surgery within 5 years of primary repair ranged from 4.9% to 21.8% across surgeons and from 4.5% to 24.7% across hospitals.

CONCLUSIONS

There are substantial differences among surgeons and hospitals in the rates of secondary lip surgery. Further work is needed to identify causes for this variation among providers.

Five-Fold Variation Among Surgeons and Hospitals in the Use of Secondary Palate Surgery

OBJECTIVE

To identify child-, surgeon- and hospital-specific factors at the time of primary cleft palate repair that are associated with the use of secondary palate surgery.

DESIGN

Retrospective cohort study.

SETTING

Forty-nine pediatric hospitals.

PARTICIPANTS

Children who underwent cleft palate repair between 1998 and 2015.

MAIN OUTCOME MEASURE

Time from primary cleft palate repair to secondary palate surgery.

RESULTS

By 5 years after the primary palate repair, 27.5% of children had undergone secondary palate surgery. In multivariable analysis, cleft type and age at primary palate repair were both associated with secondary surgery ( P < .01). Children with unilateral cleft lip and palate had a 1.69-fold increased hazard of secondary surgery (95% confidence interval [CI]: 1.54-1.85) compared to children with cleft palate alone. Primary palate repair before 9 months had a 3.99-fold increased hazard of secondary surgery (95% CI: 3.39-4.07) compared to repair at 16 to 24 months of age. After adjusting for cleft type, age at repair, and procedure volume, there remained substantial variation in secondary surgery use among surgeons and hospitals ( P < .01). For children with isolated cleft palate, the predicted proportion of children undergoing secondary surgery within 5 years of primary repair ranged from 8.5% to 46.0% across surgeons and 9.1% to 49.4% across hospitals.

CONCLUSIONS

There are substantial differences among surgeons and hospitals in the rates of secondary palate surgery. Further work is needed to identify causes for this variation among providers and develop interventions to reduce the need for secondary surgery.

Are Caregivers Who Respond to the Child HCAHPS Survey Reflective of All Hospitalized Pediatric Patients?

OBJECTIVES

The Child Hospital Consumer Assessment of Healthcare Providers and Systems (C-HCAHPS) survey was developed to measure satisfaction levels of pediatric inpatients' caregivers. Studies in adults have revealed that certain demographic groups (people of color or who are multiracial and people with public insurance) respond to surveys at decreased rates, contributing to nonresponse bias. Our primary goal was to determine if results from the C-HCAHPS survey accurately reflect the intended population or reveal evidence of nonresponse bias. Our secondary goal was to examine whether demographic or clinical factors were associated with increased satisfaction levels.

METHODS

This was a retrospective cohort study of responses (n = 421) to the C-HCAHPS survey of patients admitted to a tertiary-care pediatric hospital between March 2016 and March 2017. Respondent demographic information was compared with that of all hospital admissions over the same time frame. Satisfaction was defined as "top-box" scores for questions on overall rating and willingness to recommend the hospital.

RESULTS

Caregivers returning surveys were more likely to be white, non-Hispanic, and privately insured (P < .001). Caregivers with the shortest emergency department wait times were more likely to assign top-box scores for global rating (P = .025). We found no differences in satisfaction between race and/or ethnicity, length of stay, insurance payer, or total cost.

CONCLUSIONS

Caregivers who identified with underrepresented minority groups and those without private insurance were less likely to return surveys. Among the surveys received, short emergency department wait time and older age were the only factors measured that were associated with higher satisfaction. Efforts to increase patient satisfaction on the basis of satisfaction scores may exacerbate existing disparities in health care.

Machine Learning-Based Prediction of Clinical Outcomes for Children During Emergency Department Triage

IMPORTANCE

While machine learning approaches may enhance prediction ability, little is known about their utility in emergency department (ED) triage.

OBJECTIVES

To examine the performance of machine learning approaches to predict clinical outcomes and disposition in children in the ED and to compare their performance with conventional triage approaches.

DESIGN, SETTING, AND PARTICIPANTS

Prognostic study of ED data from the National Hospital Ambulatory Medical Care Survey from January 1, 2007, through December 31, 2015. A nationally representative sample of 52 037 children aged 18 years or younger who presented to the ED were included. Data analysis was performed in August 2018.

MAIN OUTCOMES AND MEASURES

The outcomes were critical care (admission to an intensive care unit and/or in-hospital death) and hospitalization (direct hospital admission or transfer). In the training set (70% random sample), using routinely available triage data as predictors (eg, demographic characteristics and vital signs), we derived 4 machine learning-based models: lasso regression, random forest, gradient-boosted decision tree, and deep neural network. In the test set (the remaining 30% of the sample), we measured the models' prediction performance by computing C statistics, prospective prediction results, and decision curves. These machine learning models were built for each outcome and compared with the reference model using the conventional triage classification information.

RESULTS

Of 52 037 eligible ED visits by children (median [interquartile range] age, 6 [2-14] years; 24 929 [48.0%] female), 163 (0.3%) had the critical care outcome and 2352 (4.5%) had the hospitalization outcome. For the critical care prediction, all machine learning approaches had higher discriminative ability compared with the reference model, although the difference was not statistically significant (eg, C statistics of 0.85 [95% CI, 0.78-0.92] for the deep neural network vs 0.78 [95% CI, 0.71-0.85] for the reference; P = .16), and lower number of undertriaged critically ill children in the conventional triage levels 3 to 5 (urgent to nonurgent). For the hospitalization prediction, all machine learning approaches had significantly higher discrimination ability (eg, C statistic, 0.80 [95% CI, 0.78-0.81] for the deep neural network vs 0.73 [95% CI, 0.71-0.75] for the reference; P < .001) and fewer overtriaged children who did not require inpatient management in the conventional triage levels 1 to 3 (immediate to urgent). The decision curve analysis demonstrated a greater net benefit of machine learning models over ranges of clinical thresholds.

CONCLUSIONS AND RELEVANCE

Machine learning-based triage had better discrimination ability to predict clinical outcomes and disposition, with reduction in undertriaging critically ill children and overtriaging children who are less ill.

Pediatric Inpatient-Status Volume and Cost at Children's and Nonchildren's Hospitals in the United States: 2000-2009

OBJECTIVES

The evolving role of children's hospitals (CHs) in the setting of rising health care costs has not been fully explored. We compared pediatric inpatient discharge volumes and costs by hospital type and examined the impact of care complexity and hospital-level factors on costs.

METHODS

A retrospective, cross-sectional study of care between 2000 and 2009 was performed by using the Kids' Inpatient Database. Weighted discharge data were used to generate national estimates for a comparison of inpatient volume, cost, and complexity at CHs and nonchildren's hospitals (NCHs). Linear regression was used to assess how complexity, payer mix, and hospital-level characteristics affected inflation-adjusted costs.

RESULTS

Between 2000 and 2009, the number of discharges per 1000 children increased from 6.3 to 7.7 at CHs and dropped from 55.4 to 53.3 at NCHs. The proportion of discharges at CHs grew by 6.8% between 2006 and 2009 alone. In 2009, CHs were responsible for 12.6% (95% confidence interval: 10.4%-14.9%) of pediatric discharges and 14.7% of major therapeutic procedures, yet they accounted for 23.0% of inpatient costs. Costs per discharge were significantly higher at CHs than at NCHs for all years (P < .001); however, the increase in costs seen over time was not significant. Care complexity increased during the study period at both CHs and NCH, but it could not be used to fully account for the difference in costs.

CONCLUSIONS

National trends reveal a small rise in both the proportion of inpatient discharges and the hospital costs at CHs, with costs being significantly higher at CHs than at NCHs. Research into factors influencing costs and the role of CHs is needed to inform policy and contain costs.

Development of the CIDSS2 Score for Children with Mild Head Trauma without Intracranial Injury

While most children with mild traumatic brain injury (mTBI) without intracranial injury (ICI) can be safely discharged home from the emergency department, many are admitted to the hospital. To support evidence-based practice, we developed a decision tool to help guide hospital admission decisions. This study was a secondary analysis of a prospective study conducted in 25 emergency departments. We included children under 18 years who had Glasgow Coma Scale score 13-15 head injuries and normal computed tomography scans or skull fractures without significant depression. We developed a multi-variable model that identified risk factors for extended inpatient management (EIM; defined as hospitalization for 2 or more nights) for TBI, and used this model to create a clinical risk score. Among 14,323 children with mTBI without ICI, 20% were admitted to the hospital but only 0.76% required EIM for TBI. Key risk factors for EIM included Glasgow Coma Scale score less than 15 (odds ratio [OR] = 8.1; 95% confidence interval [CI] 4.0-16.4 for 13 vs. 15), drug/alcohol Intoxication (OR = 5.1; 95% CI 2.4-10.7), neurological Deficit (OR = 3.1; 95% CI 1.4-6.9), Seizure (OR = 3.7; 95% CI 1.8-7.8), and Skull fracture (odds ratio [OR] 24.5; 95% CI 16.0-37.3). Based on these results, the CIDSS₂ risk score was created. The model C-statistic was 0.86 and performed similarly in children less than (C = 0.86) and greater than or equal to 2 years (C = 0.86). The CIDSS₂ score is a novel tool to help physicians identify the minority of children with mTBI without ICI at increased risk for EIM, thereby potentially aiding hospital admission decisions.

Characterizing Hospitalizations for Pediatric Concussion and Trends in Care

BACKGROUND

Children hospitalized for concussion may be at a higher risk for persistent symptoms, but little is known about this subset of children.

OBJECTIVE

Delineate a cohort of children admitted for concussion, describe care practices received, examine factors associated with prolonged length of stay (LOS) or emergency department (ED) readmission, and investigate changes in care over time.

DESIGN, SETTING

Retrospective analysis of data submitted by 40 pediatric hospitals to the Pediatric Health Information System.

PATIENTS

Children 0 to 17 years old admitted with a primary diagnosis of concussion from 2007 to 2014.

MEASUREMENTS

Descriptive statistics characterized this cohort and care practices delivered, logistic regression identified factors associated with a LOS of =2 days and ED readmission, and trend analyses assessed changes in care over time.

RESULTS

Of the 10,729 children admitted for concussion, 68.7% received intravenous pain or antiemetic medications. Female sex, adolescent age, and having government insurance were all associated (P = .02) with increased odds of LOS = 2 days and ED revisit. Proportions of children receiving intravenous ondansetron (slope = 1.56, P = .001) and ketorolac (slope = 0.61, P < .001) increased over time, and use of neuroimaging (slope = -1.75, P < .001) decreased.

CONCLUSIONS

Although concussions are usually selflimited, hospitalized children often receive intravenous therapies despite an unclear benefit. Factors associated with prolonged LOS and ED revisit were similar to predictors of postconcussive syndrome. Since there has been an increased use of specific therapeutics, prospective evaluation of their relationship with concussion recovery could lay the groundwork for evidenced-based admission criteria and optimize recovery.

Using Length of Stay to Understand Patient Flow for Pediatric Inpatients

Objectives:

Develop and test a new metric to assess meaningful variability in inpatient flow.

Methods:

Using the pediatric administrative dataset, Pediatric Health Information System, that quantifies the length of stay (LOS) in hours, all inpatient and observation encounters with 21 common diagnoses were included from the calendar year 2013 in 38 pediatric hospitals. Two mutually exclusive composite groups based on diagnosis and presence or absence of an ICU hospitalization termed Acute Care Composite (ACC) and ICU Composite (ICUC), respectively, were created. These composites consisted of an observed-to-expected (O/E) LOS as well as an excess LOS percentage (ie, the percent of day beyond expected). Seven-day all-cause risk-adjusted rehospitalizations was used as a balancing measure. The combination of the ACC, the ICUC, and the rehospitalization measures forms this new metric.

Results:

The diagnosis groups in the ACC and the ICUC included 113,768 and 38,400 hospitalizations, respectively. The ACC had a median O/E LOS of 1.0, a median excess LOS percentage of 23.9% and a rehospitalization rate of 1.7%. The ICUC had a median O/E LOS of 1.1, a median excess LOS percentage of 32.3%, and rehospitalization rate of 4.9%. There was no relationship of O/E LOS and rehospitalization for either ACC or ICUC.

Conclusions:

This metric shows variation among hospitals and could allow a pediatric hospital to assess the performance of inpatient flow.

Management of Pediatric Isolated Skull Fractures: A Decision Tree and Cost Analysis on Emergency Department Disposition Strategies

OBJECTIVE

Pediatric isolated skull fractures (ISFs) are common injuries that represent challenging disposition decisions for clinicians. The purpose of this study is to use a decision analysis to compare the clinical and cost-effectiveness of 3 emergency department (ED)-based disposition scenarios for a pediatric patient presenting with ISF.

METHODS

We conducted a cost-effectiveness analysis comparing ED disposition scenarios that included current practice, increased at-home surveillance, and observation unit utilization. Current rates of admission, deterioration after initial diagnosis, and ED return after discharge, as well as cost of observation-only status, were obtained through literature review. Cost calculations using Healthcare Cost and Utilization Project data included total ED cost, admission without complication, and admission with deterioration.

RESULTS

In current practice, 76% of subjects with ISF are admitted and 2.5% of those develop persistent or new symptoms. No patient diagnosed with ISF required neurosurgical intervention. Of those discharged home from the ED, 2.8% return with a new concern with 7.4% having new findings on imaging leading to admission. Total cost per 100 patients by current practice was US $583,587. Increasing at-home surveillance by 20% resulted in a total cost saving of US $113,176 per 100 patients while increasing returns to the ED from less than 1% to 1.1%. Admitting at the current rate to an observation unit resulted in a US $205,395 cost saving per 100 patients.

CONCLUSIONS

Decreased inpatient utilization through home surveillance or observation unit use reduced cost associated with pediatric ISF management without increasing clinical risk owing to the low probability of clinical deterioration after initial diagnosis.

Variation among cleft centres in the use of secondary surgery for children with cleft palate: a retrospective cohort study

OBJECTIVES

To test whether cleft centres vary in their use of secondary cleft palate surgery, also known as revision palate surgery, and if so to identify modifiable hospital- and surgeon-factors that are associated with use of secondary surgery.

DESIGN

Retrospective cohort study.

SETTING

Forty-three paediatric hospitals across the United States.

PATIENTS

Children with cleft lip and palate who underwent primary cleft palate repair from 1999 to 2013.

MAIN OUTCOME MEASURES

Time from primary cleft palate repair to secondary palate surgery.

RESULTS

We identified 4,939 children who underwent primary cleft palate repair. At ten years after primary palate repair, 44% of children had undergone secondary palate surgery. Significant variation existed among hospitals (p<0.001); the proportion of children undergoing secondary surgery by 10 years ranged from 9% to 77% across hospitals. After adjusting for patient demographics, primary palate repair before nine months of age was associated with an increased hazard of secondary palate surgery (initial hazard ratio 6.74, 95% CI 5.30-8.73). Postoperative antibiotics, surgeon procedure volume, and hospital procedure volume were not associated with time to secondary surgery (p>0.05). Of the outcome variation attributable to hospitals and surgeons, between-hospital differences accounted for 59% (p<0.001), while between-surgeon differences accounted for 41% (p<0.001).

CONCLUSIONS

Substantial variation in the hazard of secondary palate surgery exists depending on a child's age at primary palate repair and the hospital and surgeon performing their repair. Performing primary palate repair before nine months of age substantially increases the hazard of secondary surgery. Further research is needed to identify other factors contributing to variation in palate surgery outcomes among hospitals and surgeons.

Observation for isolated traumatic skull fractures in the pediatric population: unnecessary and costly

BACKGROUND

Blunt head trauma accounts for a majority of pediatric trauma admissions. There is a growing subset of these patients with isolated skull fractures, but little evidence guiding their management. We hypothesized that inpatient neurological observation for pediatric patients with isolated skull fractures and normal neurological examinations is unnecessary and costly.

METHODS

We performed a single center 10year retrospective review of all head traumas with isolated traumatic skull fractures and normal neurological examination. Exclusion criteria included: penetrating head trauma, depressed fractures, intracranial hemorrhage, skull base fracture, pneumocephalus, and poly-trauma. In each patient, we analyzed: age, fracture location, loss of consciousness, injury mechanism, Emergency Department (ED) disposition, need for repeat imaging, hospital costs, intracranial hemorrhage, and surgical intervention.

RESULTS

Seventy-one patients presented to our ED with acute isolated skull fractures, 56% were male and 44% were female. Their ages ranged from 1week to 12.4years old. The minority (22.5%) of patients were discharged from the ED following evaluation, whereas 77.5% were admitted for neurological observation. None of the patients required neurosurgical intervention. Age was not associated with repeat imaging or inpatient observation (p=0.7474, p=0.9670). No patients underwent repeat head imaging during their index admission. Repeat imaging was obtained in three previously admitted patients who returned to the ED. Cost analysis revealed a significant difference in total hospital costs between the groups, with an average increase in charges of $4,291.50 for admitted patients (p<0.0001).

CONCLUSION

Pediatric isolated skull fractures are low risk conditions with a low likelihood of complications. Further studies are necessary to change clinical practice, but our research indicates that these patients can be discharged safely from the ED without inpatient observation. This change in practice, additionally, would allow for huge health care dollar savings.

Implementation Analysis of a Nurse-Led Observation Unit

OBJECTIVE

This implementation analysis of a nurse-led observation unit describes the development process and analyzes patients' characteristics, patient satisfaction, and provider perceptions.

BACKGROUND

A nurse-led observation unit was developed to create more inpatient bed capacity and place patients in the clinical area best suited to their needs.

METHODS

Descriptive statistics and content analysis were used for analysis.

RESULTS

The average length of stay of 467 patients was 1.1 days; 68.1% (n = 318) were female. Elective surgery was the most frequent reason for admission. All of the patients rated the observation unit patient feedback survey factors favorably except for noise. All healthcare providers (n = 64) reported that they communicated well with each other and had resources to provide quality care but rated the environment less favorably.

CONCLUSIONS

A nurse-led observation unit was found to be an effective and efficient approach to providing postoperative and postprocedure care, which was generally well received by patients and healthcare providers.

Socioeconomic Status and Hospitalization Costs for Children with Brain and Spinal Cord Injury

OBJECTIVE

To determine if household income is associated with hospitalization costs for severe traumatic brain injury (TBI) and spinal cord injury (SCI).

STUDY DESIGN

Retrospective cohort study of inpatient, nonrehabilitation hospitalizations at 43 freestanding children's hospitals for patients <19 years old with unintentional severe TBI and SCI from 2009-2012. Standardized cost of care for hospitalizations was modeled using mixed-effects methods, adjusting for age, sex, race/ethnicity, primary payer, presence of chronic medical condition, mechanism of injury, injury severity, distance from residence to hospital, and trauma center level. Main exposure was zip code level median annual household income.

RESULTS

There were 1061 patients that met inclusion criteria, 833 with TBI only, 227 with SCI only, and 1 with TBI and SCI. Compared with those with the lowest-income zip codes, patients from the highest-income zip codes were more likely to be older, white (76.7% vs 50.4%), have private insurance (68.9% vs 27.9%), and live closer to the hospital (median distance 26.7 miles vs 81.2 miles). In adjusted models, there was no significant association between zip code level household income and hospitalization costs.

CONCLUSIONS

Children hospitalized with unintentional, severe TBI and SCI showed no difference in standardized hospital costs relative to a patient's home zip code level median annual household income. The association between household income and hospitalization costs may vary by primary diagnosis.

The Scope and Trends of Pediatric Hospitalizations in Texas, 2004-2010

OBJECTIVE

To examine demographics and trends of financial cost and prominent diseases/conditions resulting in inpatient hospitalizations for infants, children, and adolescents in Texas between 2004 and 2010.

METHODS

Longitudinal retrospective cross-sectional study using the Texas Hospital Inpatient Discharge Database, including all pediatric hospitalizations in the state of Texas, 2004 to 2010.

RESULTS

Texas has an average of 591 571 pediatric hospitalizations per year. Birth was the most common reason for hospitalization, representing 64% of all pediatric hospitalization annually in Texas. Respiratory illnesses were the most common discharge diagnosis for hospitalized children ages 1 month to 9 years and demonstrated a 2% decrease over the study period. The rate of hospitalizations for digestive conditions and childbirth also demonstrated a decrease over this time frame: 4.7% and 3.0%, respectively. The rate of mental illness diagnoses increased 2.5% over the time frame and represented the most common discharge diagnosis for children aged 10 to 14. Childbirth was the most common reason for hospitalization for adolescents aged 15 to 17 years. There was no increase in total cost of pediatric hospitalizations over the time period under study.

CONCLUSIONS

After birth, respiratory illnesses represent the most common reason for hospitalization for children (between 1 month and 10 years of age) in Texas. Mental health conditions and childbirth represent the most common reason for hospitalization for young adolescents (10-14 years) and older adolescents (15-17 years), respectively.

Observation-status patients in children's hospitals with and without dedicated observation units in 2011

BACKGROUND

Pediatric observation units (OUs) have demonstrated reductions in lengths of stay (LOS) and costs of care. Hospital-level outcomes across all observation-status stays have not been evaluated in relation to the presence of a dedicated OU in the hospital.

OBJECTIVE

To compare observation-status stay outcomes in hospitals with and without a dedicated OU.

DESIGN

Cross-sectional analysis of hospital administrative data.

METHODS

Observation-status stay outcomes were compared in hospitals with and without a dedicated OU across 4 categories: (1) LOS, (2) standardized costs, (3) conversion to inpatient status, and (4) return care.

SETTING/PATIENTS

Observation-status stays in 31 free-standing children's hospitals contributing observation patient data to the Pediatric Health Information System database, 2011.

RESULTS

Fifty-one percent of the 136,239 observation-status stays in 2011 occurred in 14 hospitals with a dedicated OU; the remainder were in 17 hospitals without. The percentage of observation-status same-day discharges was higher in hospitals with a dedicated OU compared with hospitals without (23.8 vs 22.1, P < 0.001), but risk-adjusted LOS in hours and total standardized costs were similar. Conversion to inpatient status was higher in hospitals with a dedicated OU (11.06%) compared with hospitals without (9.63%, P < 0.01). Adjusted odds of return visits and readmissions were comparable.

CONCLUSIONS

The presence of a dedicated OU appears to have an influence on same-day and morning discharges across all observation-status stays without impacting other hospital-level outcomes. Inclusion of location of care (eg, dedicated OU, inpatient unit, emergency department) in hospital administrative datasets would allow for more meaningful comparisons of models of hospital care.

Trends in observation-prone emergency department visits among Michigan children, 2007-2011

OBJECTIVES

To the best of the authors' knowledge, admission of children under observation status in community hospitals has not been examined. The hypothesis of this study was that there has been an increase in observation charge code use over time and variations in the application of observation charge codes across hospital types.

METHODS

This was a cross-sectional analysis of 5 years (2007 through 2011) of administrative claims data from Michigan residents enrolled in Medicaid, Blue Cross/Blue Shield of Michigan preferred provider organization, and Blue Cross Network health maintenance organization compiled into a single data set. Emergency department (ED) visits to facilities in Michigan made by children (younger than 18 years) were selected. Observation-prone ED visits were identified based on the presence of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. Counts of observation-prone ED visits were determined and descriptive statistics were calculated. Changes over time in the proportion of visits with observation charge codes by hospital type were assessed with chi-square analysis.

RESULTS

The observation-prone ICD-9-CM codes were identified in 881,622 ED visits made by children to 142 Michigan facilities during the 5-year study period. Overall, the vast majority of visits (n = 646,499; 91.0%) with the selected ICD-9-CM codes resulted in discharge from the ED without associated observation or inpatient charge codes. Among the 64,288 visits that resulted in admission for observation or inpatient care, observation charge codes without inpatient charge codes were applied to 22,933 (35.7%) admissions, observation and inpatient charge codes were applied to 4,756 (7.4%) admissions, and inpatient charge codes without observation charge codes were applied to 36,599 (56.9%) admissions. Hospitals with pediatric ED and inpatient services (Type 1 and Type 2 hospitals) had higher proportions of ED visits that went on to admission for observation or inpatient care (15.9 and 10.7%) than hospitals without pediatric ED services (Type 3 and Type 4 hospitals; 7.2 and 3.7%). The proportion of admissions that had observation charge codes for all hospital types increased over time, most prominently among Type 1 and Type 2 hospitals.

CONCLUSIONS

The application of observation charge codes to Michigan children with observation-prone conditions has increased over time across all hospital types. There is a need to evaluate pediatric observation care in diverse settings to compare the effectiveness of different models.

Rates and impact of potentially preventable readmissions at children's hospitals

OBJECTIVE

To assess readmission rates identified by 3M-Potentially Preventable Readmissions software (3M-PPRs) in a national cohort of children's hospitals.

STUDY DESIGN

A total of 1 719 617 hospitalizations for 1 531 828 unique patients in 58 children's hospitals from 2009 to 2011 from the Children's Hospital Association Case-Mix Comparative database were examined. Main outcome measures included rates, diagnoses, and costs of potentially preventable readmissions (PPRs) and all-cause readmissions.

RESULTS

The 7-, 15-, and 30-day rates by 3M-PPRs were 2.5%, 4.1%, and 6.2%, respectively. Corresponding all-cause readmission rates were 5.0%, 8.7%, and 13.3%. At 30 days, 60.6% of all-cause readmissions were considered nonpreventable by 3M-PPRs, more than one-half of which were related to malignancies. The percentage of readmissions rated as potentially preventable was similar at all 3 time intervals. Readmissions after chemotherapy, acute leukemia, and cystic fibrosis were all considered nonpreventable, and at least 80% of readmissions after index admissions for sickle cell crisis, bronchiolitis, ventricular shunt procedures, asthma, and appendectomy were designated potentially preventable. Total costs for all readmissions were $1.7 billion; PPRs accounted for 27.3% of these costs. The most costly readmissions were associated with ventricular shunt procedures ($26.5 million/year), seizures ($15.5 million/year), and sickle cell crisis ($15.0 million/year).

CONCLUSIONS

Rates of PPRs were significantly lower than all-cause readmission rates more than one-half of which were caused by exclusion of malignancies. Annual costs of PPRs, although significant in the aggregate, appear to represent a much smaller cost-savings opportunity for children than for adults. Our study may help guide children's hospitals to focus readmission reduction strategies on areas where the financial vulnerability is greatest based on 3M-PPRs.

Resource utilization of pediatric patients exposed to venom

BACKGROUND AND OBJECTIVE

Treating envenomation with antivenom is costly. Many patients being treated with antivenom are in observation status, a billing designation for patients considered to need care that is less resource-intensive, and less expensive, than inpatient care. Observation status is also associated with lower hospital reimbursements and higher patient cost-sharing. The goal of this study was to examine resource utilization for treatment of envenomation under observation and inpatient status, and to compare patients in observation status receiving antivenom with all other patients in observation status.

METHODS

This was a retrospective study of patients with a primary diagnosis of toxic effect of venom seen during 2009 at 33 freestanding children's hospitals in the Pediatric Health Information System. Data on age, length of stay, adjusted costs (ratio cost to charges), ICU flags, and antivenom utilization were collected. Comparisons were conducted according to admission status (emergency department only, observation status, and inpatient status), and between patients in observation status receiving antivenom and patients in observation status with other diagnoses.

RESULTS

A total of 2755 patients had a primary diagnosis of toxic effect of venom. Of the 335 hospitalized, either under observation (n = 124) or inpatient (n = 211) status, 107 (31.9%) received antivenom. Of those hospitalized patients receiving antivenom, 24 (22.4%) were designated as observation status. Costs were substantially higher for patients who received antivenom and were driven by pharmacy costs (mean cost: $17 665 for observation status, $20 503 for inpatient status). Mean costs for the 47 162 patients in observation status with other diagnoses were $3001 compared with $17 665 for observation-status patients who received antivenom.

CONCLUSIONS

Treatment of envenomation with antivenom represents a high-cost outlier within observation-status hospitalizations. Observation status can have financial consequences for hospitals and patients.

Common and costly hospitalizations for pediatric mental health disorders

BACKGROUND AND OBJECTIVES

Inpatient pediatric mental health is a priority topic for national quality measurement and improvement, but nationally representative data on the patients admitted or their diagnoses are lacking. Our objectives were: to describe pediatric mental health hospitalizations at general medical facilities admitting children nationally; to assess which pediatric mental health diagnoses are frequent and costly at these hospitals; and to examine whether the most frequent diagnoses are similar to those at free-standing children's hospitals.

METHODS

We examined all discharges in 2009 for patients aged 3 to 20 years in the nationally representative Kids' Inpatient Database (KID) and in the Pediatric Health Information System (free-standing children's hospitals). Main outcomes were frequency of International Classification of Diseases, Ninth Revision, Clinical Modification-defined mental health diagnostic groupings (primary and nonprimary diagnosis) and, using KID, resource utilization (defined by diagnostic grouping aggregate annual charges).

RESULTS

Nearly 10% of pediatric hospitalizations nationally were for a primary mental health diagnosis, compared with 3% of hospitalizations at free-standing children's hospitals. Predictors of hospitalizations for a primary mental health problem were older age, male gender, white race, and insurance type. Nationally, the most frequent and costly primary mental health diagnoses were depression (44.1% of all mental health admissions; $1.33 billion), bipolar disorder (18.1%; $702 million), and psychosis (12.1%; $540 million).

CONCLUSIONS

We identified the child mental health inpatient diagnoses with the highest frequency and highest costs as depression, bipolar disorder, and psychosis, with substance abuse an important comorbid diagnosis. These diagnoses can be used as priority conditions for pediatric mental health inpatient quality measurement.