This study documented the clinical and economic burden associated with IPD, non-invasive ACP, and AOM in hospitalized children in the US, specifically an observed in-hospital mortality rate of 0.6%, median LOS of 4 days, median hospital cost of $4,240 per admission, and median hospital loss (negative margin) of $63 per admission. These findings are particularly notable given that the patient cohort primarily included non-invasive ACP and AOM, which is typically less severe than IPD but far more prevalent [1].
There were 28 patients with an outcome of in-hospital mortality. Risk factors for in-hospital mortality were prior 90-day admission, ICU admission, and having a condition of moderate or high risk for PD. These data indicate the critical importance of PD prevention efforts, primarily vaccination, particularly in patients with factors that put them at greater risk for PD, including chronic and immunocompromising conditions.
Patient-related variables associated with longer LOS, higher hospital costs per admission, and greater negative hospital margins per admission included IPD, prior 90-day admission, ICU admission, and having risk conditions that put them at greater risk for PD. Other studies have also identified associations between underlying comorbid conditions and high costs in patients with PD [17]. Age showed a mixed pattern across these outcomes. Compared with the 5–17 year age group, children < 2 years of age had a longer LOS while the 2–4 year age group trended towards a significantly shorter LOS. Hospital costs per admission were also highest in the youngest age group, but not significantly different from costs in the 5–17 year reference age group, which had significantly higher costs than the 2–4 year age group. Negative margins (hospital loss) were highest in children 5–17 years of age, although the difference was not significantly different from the other age groups. Other studies have also reported higher costs in the youngest age group of children with pneumonia [18]. Possible reasons behind the high costs and negative margins (hospital loss) in the 5–17 year age group remain unclear. One study found that children with comorbidities were older at the time of IPD diagnosis [19], providing a possible explanation for higher costs in this age group. Additional studies will be required to explore this potential association. Other variables that impacted economic outcomes were type of insurance and characteristics of the hospital including urban location, teaching status, and US census region.
This study differs from most other reports on PD in the US pediatric population by reporting hospital costs per admission, rather than costs estimated from Markov models of simulated clinical events [18, 20–23]. These differences in study design make comparisons among studies difficult. The most closely comparable study is a recent study by Hu and colleagues in US children < 18 years based on administrative claims data which reported payments rather than costs [24]. In children with non-invasive ACP, Hu et al. [24] reported mean payments of $15,041 per inpatient admission for the Medicaid-insured children and $34,653 for the commercially insured children compared with estimated mean hospital costs of $15,765 for non-invasive ACP/AOM and $31,837 for IPD in the study reported here. An earlier study using 2008 to 2014 data from a commercial claims database reported a cost of $10,963 per hospitalized pneumonia episode for patients of all ages (including 31% of cases in patients < 18 years of age) but did not analyze pediatric cost data separately [18]. These three studies show the high costs associated with PD in children, including those with non-invasive PD. Our study extends these data by showing that, on average, hospitals lose an adjusted $3,298 for each case of PD in children with no underlying risk factors and up to $7,221 per case in immunocompromised children (high risk PD category). To the best of our knowledge, our study is the first to report on the hospital margins associated with the treatment of pediatric PD in US hospitals, thereby providing additional evidence of the financial impact of PD on the healthcare systems.
Although positive lab cultures were reported for only 41 patients, over half of these patients (n = 23) had antibiotic-resistant S. pneumoniae. This finding is consistent with other reports of high rates of AMR in S. pneumoniae isolates from pediatric and adult patients in the US [2, 11, 12, 25, 26]; increases in resistance are of particular concern for non-vaccine serotypes [10]. Although AMR status was not a significant predictor of outcomes in multivariable analyses, the number of AMR patients included in these analyses may have been too few to detect a significant association.
The high clinical and economic burden and AMR rates associated with PD in children support the need for expanded pneumococcal vaccination efforts. In a prospective study examining IPD in children at 8 children's hospitals in the US from 2014 to 2017, approximately three-quarters (76.1%) of IPD cases were caused by non-PCV13 S. pneumoniae serotypes and these serotypes are more common in patients with underlying risk factors [27], a variable identified here as a key risk factor for poor outcomes. Models suggest that the introduction of effective vaccines protecting against additional S. pneumoniae serotypes may play a key role in reducing not only overall disease burden [7, 20, 28–30], but also AMR S. pneumoniae infections [10]. In addition to higher valency vaccines, more widespread vaccination may also diminish PD impact in children. Pneumococcal vaccination rates in US children are generally high, but some populations, including uninsured children and children living below the poverty level, have suboptimal coverage [31]. Our data also support the importance of protection in older children with risk factors for PD. Targeted vaccination campaigns may help address these at-risk populations.
Study limitations include low representation of pediatric patients from certain geographic regions (Supplementary Table 1), and a small number of patients with culture results. Selection bias may have affected AMR rates, as more severely ill patients are more likely to have cultures performed. AMR rates may also have been affected by variations in breakpoints and susceptibility tests used by different facilities. This study was intentionally designed to focus on ICD10 codes as a determinant of PD. Blood and respiratory cultures are not routinely performed in children with suspected pneumonia or PD and provide limited information concerning a causative pathogen [32, 33]. This study did not evaluate S. pneumoniae serotypes or vaccination status. The study period did not span the COVID-19 pandemic period; rates of IPD [34] and AMR [35] were both affected by pandemic-related factors. In the US, AOM is a leading cause of physician office visits and antibiotic prescriptions in children, and it is primarily managed in outpatient settings [36]. The AOM episodes in this study of hospitalized patients likely reflected complicated AOM which is less common and therefore the outcome estimates for AOM may be underestimates.
In conclusion, this study documents a substantial burden of IPD, non-invasive ACP, and AOM, particularly non-invasive ACP, in hospitalized children in the US. Populations at high risk for poor outcomes include young children (< 2 years) and those with risk conditions that put them at greater risk for PD. The significant impact of pediatric IPD, ACP, and AOM on patients and health systems highlights the need for widespread prevention efforts including pneumococcal vaccination and antimicrobial stewardship programs.