To determine age cutoffs that hospitals without pediatric inpatient beds apply when hospitalizing children.
We conducted a cross-sectional study of patients <25 years old visiting emergency departments in 5 states in 2016 using the Healthcare Cost and Utilization Project State Emergency Department and Inpatient Databases. Hospitals were classified as adult (no pediatric inpatient beds) or pediatric capable (>0 pediatric beds). Referral rates were calculated for each year of life as transfers divided by transfers plus hospitalizations. Two age cutoffs were determined for defining pediatric patients: a specific cutoff (the age at which referral rates were significantly lower than those for younger patients) and an inclusive cutoff (the age at which referral rates differed most from those for younger patients).
Among 389 581 transfers and hospitalizations, 91 967 (23.6%) occurred in adult hospitals. Referral rates at adult hospitals were 86.0% at age 15, 80.6% at age 16, 72.0% at age 17, and 30.5% at age 18. The specific age cutoff was 16 because referral rates were lower than those for ages 0 to 15 (P < .001). The inclusive age cutoff was 18 because the odds ratio for referral was lowest when comparing age 18 to ages 0 to 17.
Children aged <16 years specifically define a population of pediatric patients, as defined by whether an adult hospital would hospitalize instead of transfer from an emergency department. Children aged <18 years inclusively define a population of pediatric patients. These age cutoffs may be used when studying patterns of national acute care for children.
Adult hospitals (those without pediatric inpatient units) are unlikely to have the capability to care for children in the inpatient setting, but they might have the capability to care for adolescents in beds designated for adults.1,2 In research on pediatric care patterns, varying age cutoffs (such as 15, 18, or 21 years) have been used to define pediatric patients without a clear basis for this determination. Age cutoffs have the potential to impact study findings and conclusions because disease risk changes throughout childhood and young adulthood and because outcomes may depend on age. Additionally, adult hospitals’ willingness to provide care is likely to increase as patients transition to adulthood, so studies of pediatric care patterns that include young adults may not isolate pediatric care effects. Understanding the age cutoffs hospitals use to make hospitalization decisions from the emergency department (ED) could provide an empirical basis for defining pediatric patients in health services research.
Methods
This was a cross-sectional study of all ED encounters resulting in transfer or hospitalization among patients <25 years old in Florida, Iowa, Maryland, New York, and Wisconsin from 2016. This diverse group of states was selected because of the particularly high quality in pediatric identifiers and the availability of linkages to hospital bed counts. Encounters were excluded if they were birth encounters, if the primary diagnosis was psychiatric (because patterns of hospitalization differ), or if they occurred in a specialty hospital or a hospital with <50 total (pediatric plus adult) hospitalizations per year. Data were drawn from the Healthcare Cost and Utilization Project State Emergency Department and Inpatient Databases. Hospital bed counts were obtained from the American Hospital Association Annual Survey.
Adult hospitals were defined as having no pediatric inpatient beds, and pediatric-capable hospitals were defined as having any pediatric beds. The main outcome was the referral rate, which was the number of transfers to another hospital divided by transfers plus hospitalizations.3 Referral rates were calculated for each year of age by hospital type.
We determined 2 empirical age cutoffs for defining pediatric hospitalizations by examining data only from adult hospitals. The specific age cutoff was meant to be more specific for pediatric hospitalizations, and the inclusive age cutoff was more sensitive. The specific age cutoff was defined as the age in years that had a statistically lower referral rate than that for all younger patients (eg, age 14 versus ages 0–13). We used 1-sided Fisher’s exact tests with a significance level P < .01. The inclusive cutoff was defined as the age with the lowest odds ratio of referral as compared to that for all younger patients.
Results
We analyzed 389 581 encounters (including hospitalization admissions and transfers), of which 91 967 (23.6%) were to adult hospitals. Referral rates at adult hospitals were 86.0% at age 15, 80.6% at age 16, 72.0% at age 17, and 30.5% at age 18. Referral rates for patients aged <18 were 83.8% at adult hospitals and 17.4% at pediatric-capable hospitals (Fig 1). The specific age cutoff for defining pediatric patients was 16 years (P < .001), whereas the inclusive age cutoff was 18 years (odds ratio of 0.08 for age 18 years as compared to ages 0–17 years).
Referral rates at each age for adult hospitals (those without pediatric inpatient beds) and pediatric-capable hospitals (those with pediatric inpatient beds). Referral rates are calculated as transfers divided by transfers plus hospitalizations. Specific and inclusive age cutoffs are shown.
Referral rates at each age for adult hospitals (those without pediatric inpatient beds) and pediatric-capable hospitals (those with pediatric inpatient beds). Referral rates are calculated as transfers divided by transfers plus hospitalizations. Specific and inclusive age cutoffs are shown.
Discussion
Across 5 states, ED pediatric referral rates from adult hospitals were higher than those at pediatric-capable hospitals. Referral rates begin declining in adult hospitals at age 16, and declines become shallow at age 18. Among children who are not discharged from the ED, decisions to transfer or hospitalize are made at similar rates among children <16 years of age.
Research on health care use in children often includes all patients aged <18 years. Our study suggests that this approach is an inclusive one. However, our data also reveal that adult facilities hospitalize children at ages 16 and 17 at rates higher than those for children <16 years of age. Given that adult hospitals provide inpatient care to patients aged 16 or 17 years, such care may sometimes be viewed as adult care. Investigators specifically studying pediatric care should consider restricting cohorts to those aged <16 rather than those 18 years of age.
Limitations of this study include the self-report of pediatric inpatient bed counts and restriction to available states. We were further unable to assess the reasons hospitals make disposition decisions, which depend on age in addition to other factors, such as diagnosis, complexity, projected resource needs, bed availability, and others.
Conclusions
Children aged <16 years specifically define a cohort of pediatric patients. Children aged 16 or 17 years are transferred at modestly lower rates, and an age cutoff <18 years defines an inclusive population of pediatric patients.
Dr Michelson conceptualized and designed the study, drafted the initial manuscript, and performed the analysis; Dr Neuman collaborated on the study design and reviewed and revised the manuscript; and both authors approved the final manuscript as submitted.
The funder or sponsor funded Dr Michelson’s salary but did not participate directly in the work.
The deidentified analysis data set, in addition to the statistical analysis code, will be made available. The data will be made available after publication to researchers who provide a methodologically sound proposal for use in achieving the goals of the approved proposal. Proposals should be submitted to [email protected].
FUNDING: Supported by award K08HS026503 from the Agency for Healthcare Research and Quality.
Comments