Prevalence and Trends in Obesity Among Hospitalized Children

This study was a retrospective cohort analysis of children admitted to a single tertiary children’s hospital system between January 1, 2009, and December 31, 2016. Data were extracted from the hospital’s electronic health record (Cerner Corporation; Kansas City, MO). This study was deemed non–human subjects research and was approved by the Children’s Mercy Kansas City Institutional Review Board.

We included all inpatient encounters for children aged 2 to 19 years old hospitalized during the study period, mirroring the age groups reported by the NHANES obesity reports.^1,20  Directly measured weight and height measurements collected from the electronic health record were used to calculate BMI. If height was not recorded for a hospitalization encounter, it was imputed as the mean from other encounters during the same calendar year, if available. Encounters for a patient in a given year were excluded if complete anthropometric data were not obtainable from that calendar year. Outliers for BMI were removed by using the Centers for Disease Control and Prevention (CDC) SAS version 9.4 (SAS Institute Inc, Cary, NC) code for determination of biologically implausible values.²⁰  We excluded the annual encounters for 18 802 children (20 256 patient-years) because of incomplete or biologically implausible anthropometric data (Supplemental Fig 2). When comparing demographic characteristics of the excluded encounters with those of the included encounters, statistically significant, but not clinically meaningful, differences existed (eg, 52% of the included cohort were boys, compared with 53.6% of the excluded cohort) (Supplemental Table 3). De-identified individual participant data will not be made available.

The primary outcome was the annual prevalence of hospitalized children with obesity. We defined obesity according to the current CDC classification system for children of age- and sex-specific BMI percentiles of ≥95%, with CDC subclassifications of class I to class III obesity (class I obesity: ≥95th percentile; class II obesity: ≥120% of the 95th percentile, or BMI ≥ 35; and class III obesity: ≥140% of the 95th percentile, or BMI ≥ 40).^20,21 

We also assessed obesity prevalence by demographic and clinical characteristics, including hospital service line and diagnosis groups. Diagnosis groups were assigned by using the All Patients Refined Diagnosis-Related Groups (3M Corporation, Maplewood, MN) software and grouped into the following service lines: cancer care or hematology, cardiac care, respiratory, orthopedics, transplants, digestive disease, neuroscience, infectious disease, other medicine, and other surgery. We assessed obesity prevalence among the 3 diagnoses groups with the largest absolute number of patients within each service line.

All variables were summarized with frequencies and percentages, then compared across groups (with obesity versus without obesity) by using χ² tests. Obesity prevalence was calculated in 2-year increments (similar to NHANES) and was compared across time by using Cochrane-Armitage trend tests. Obesity prevalence and comparisons across time were calculated overall, by service line, and by diagnosis group. All analyses were completed on the encounter level. All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Inc), and P < .05 was considered statistically significant.

Approximately 1 in 6 (17%) of the 83 329 included hospitalizations were for children with obesity. Prevalence of obesity was highest among children who were 12 to 15 years old (20.2%), among non-Hispanic black (18.9%) or Hispanic (23.9%) children, and among children with public insurance (19.8%; all P < .001) (Table 1). There were no statistically significant differences in obesity prevalence by sex (P = .738).

The proportion of hospitalized children with obesity increased over the study period from 16.5% of hospitalizations in 2009–2010 to 17.3% in 2015–2016 (P = .002) (Fig 1). Proportions of children with class I obesity ranged from 10.3% to 10.7%, not changing significantly over the study period (P = −.273). The prevalence of class II obesity ranged from 3.1% to 3.8%, increasing significantly over the study period (P < .0001). The prevalence of class III obesity ranged from 2.6% to 3.7% but did not reveal a statistically significant change over the study period (P = .504).

The hospital service lines caring for the largest proportion of hospitalized children with obesity included orthopedics (22.1%), infectious disease (20.6%), and neuroscience (18.7%), with statistically significant differences existing between groups (P < .001; Table 2). The prevalence of obesity among children who were hospitalized increased between 2009 and 2016 for patients within only 2 service lines: digestive disease (13.9% in 2009–2010 to 15.8% in 2015–2016; P = .012), and other surgery (17.0% in 2009–2010 to 18.9% in 2015–2016; P = .022), with no statistically significant change in the other groups over time.

Several diagnosis groups within each service line (Table 2) had a disproportionate number of patients with obesity, including skin and soft tissue infections (22.5%), headaches (23.4%), poisonings (21.2%), tonsil and adenoid procedures (22.0%), and hip and lower leg orthopedic procedures (28.6 and 28.6%, respectively).

In this study, we found that ∼1 in 6 children aged 2 to 19 years who were hospitalized had obesity. We observed increases in the annual prevalence of obesity in the inpatient population, mirroring trends in the general US population.¹  Hospitalized patients with obesity were more likely to be adolescents, more likely to be African American and Hispanic, and more likely to have public insurance, characteristics that generally mirror recent NHANES data.¹  This indicates that hospitalized children with obesity are similar to children with obesity in the general population. Obesity prevalence and trends within the hospital population were also similar to estimated regional obesity rates.²²  Our study also included analyses of changes in obesity prevalence among demographic groups over time (2009–2016). The increases in obesity prevalence among particular demographic groups noted within this hospitalized population were similar to those seen in the general population.¹  Our findings are in agreement with previous studies in which it was estimated that 16% to 21% of children who are hospitalized have overweight or obesity, making up a significant portion of the pediatric inpatient population.^15,17–19  Our study adds to those studies by providing a clearer description of the overall pediatric inpatient population over time through use of actual anthropometric measures, as opposed to children with particular diagnoses or through use of administrative data.

From a clinical perspective, 3 service lines cared for a larger proportion of patients with obesity (orthopedics, infectious disease, and neuroscience). We also found that only 2 of 10 service lines (digestive disease and other surgery) experienced statistically significant changes (increases) over the study period. These 2 groups were relatively large and were likely primary drivers of the overall increase in obesity prevalence for the entire cohort. However, smaller (but not statistically significant) increases noted in other service lines might, when taken in aggregate, have contributed to the overall increase in obesity prevalence noted in the children who were hospitalized.

In addition, we identified conditions with a disproportionate number of patients with obesity, including cellulitis, headaches, poisonings, tonsil and adenoid procedures, and orthopedic procedures. Many of the hospital outcomes for these conditions are negatively affected by weight status, including increased risk of postoperative complications (eg, orthopedic and tonsil procedures and adenoidectomy) and treatment failure (eg, cellulitis).^10,23–26  In addition, children with obesity require special attention to issues that affect their safety within the hospital setting.¹⁰  For instance, children with obesity are at increased risk of experiencing medication dosing errors compared with children without obesity.^25,26  The knowledge gained from our findings may help providers and hospitals plan for and implement strategies aimed at improving care and safety for groups of children with higher obesity prevalence within the inpatient population.

As the number of hospitalized children with obesity continues to rise, especially within particular demographic and clinical groups, our findings reveal the need for further research into the special risks that this population of patients faces. Future research should be aimed at identification of disparate outcomes and equity of experiences from the patient perspective for patients with obesity. Our findings provide a starting point for potential areas of focus within the pediatric inpatient population.

Regarding our study, some limitations should be considered. This was a single-center study at a tertiary children’s hospital, limiting generalizability of our results. However, our data are generally reflective of the nationally representative NHANES data and regional population obesity prevalence. There were fewer patients in some groups, which may have introduced variability in the results (eg, minority groups). Additionally, it is possible that weight and height measurements included were not accurate or that All Patients Refined Diagnosis-Related Groups could have been misclassified on the basis of the billing or coding information provided. Finally, it is possible that there was some selection bias in the inclusion of patients; providers might have been more likely to obtain height measurements on patients with obesity, leading to an artificially inflated obesity prevalence.

The prevalence of obesity among children who are hospitalized is rising. Additionally, there exist typologies of patients, service lines, and diagnosis groups with substantially higher than average obesity prevalence. These results reveal the need to focus future research efforts on these disproportionately affected groups, with studies aimed at identification of gaps in the safety and equity of clinical care for children with obesity.