A, Multivariable analysis of associations of patient-level factors with odds of undergoing early umbilical hernia repair. B, Multivariable analysis of associations of patient-level factors with odds of undergoing early umbilical hernia repair following an emergent or urgent presentation. Early and delayed repair were defined in the context of current guidelines that recommend delaying repair of asymptomatic umbilical hernias until at least age 4 years. Early repair indicates repair performed at 3 years or younger. NA indicates not applicable; OR, odds ratio.
Hospitals are ordered by odds ratio (OR); the hospital number does not necessarily indicate the same hospital across figures. Filled dots indicate statistical outliers. The blue line indicates the line of overall effect.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Hills-Dunlap JL, Melvin P, Graham DA, Kashtan MA, Anandalwar SP, Rangel SJ. Association of Sociodemographic Factors With Adherence to Age-Specific Guidelines for Asymptomatic Umbilical Hernia Repair in Children. JAMA Pediatr. 2019;173(7):640–647. doi:10.1001/jamapediatrics.2019.1061
Are sociodemographic factors associated with adherence to age-specific guidelines for the management of asymptomatic umbilical hernias in children?
In this cohort study of 25 877 patients from 47 children’s hospitals, public insurance, lower income, and female sex were independently associated with umbilical hernia repair earlier than recommended by current guidelines.
Children from disadvantaged socioeconomic backgrounds are at increased risk of undergoing early repair of asymptomatic umbilical hernias, many of which may spontaneously close with further observation.
Current guidelines recommend delaying repair of asymptomatic umbilical hernia in children until after age 4 to 5 years to allow for spontaneous closure.
To examine the association of sociodemographic factors with adherence to age-specific guidelines for asymptomatic umbilical hernia repair in children.
Design, Setting, and Participants
In this multicenter retrospective cohort study, children 17 years and younger who underwent umbilical hernia repair from January 2013 to June 2018 at 47 freestanding children’s hospitals participating in the Pediatric Health Information System database were eligible for study inclusion. Children who underwent multiple procedures, repair of recurrent hernias, or had missing sociodemographic data were excluded.
Early umbilical hernia repair was defined as repair at 3 years or younger. Emergent or urgent presentation was defined as repair performed during the same encounter or within 2 weeks of an emergency department visit, respectively. Patients were categorized by sex, race/ethnicity, insurance type, income quintile, and presence of complex chronic conditions.
Main Outcomes and Measures
Multivariable mixed-effects logistic regression was used to evaluate the association of sociodemographic factors with the odds of early repair after adjusting for emergent or urgent presentation and hospital-level effects.
Of the 25 877 included children, 13 817 (53.4%) were female, 14 143 (54.7%) had public insurance, and the median (interquartile range) age was 5.0 (3.0-6.0) years. Following adjustment, increased odds of early repair was associated with public insurance (public vs commercial insurance: odds ratio [OR], 1.46; 95% CI, 1.36-1.56; P < .001), lower income (lowest vs highest income quintile: OR, 1.48; 95% CI, 1.33-1.65; P < .001), and female sex (female vs male sex: OR, 1.20; 95% CI, 1.13-1.27; P < .001). Children with public insurance in the lowest income quintile had 2.2-fold increased odds of early repair compared with children with commercial insurance in the highest income quintile (OR, 2.15; 95% CI, 1.93-2.40; P < .001). Sociodemographic factors were not associated with increased odds of early repair in the subgroup of children who underwent early repair following emergent or urgent presentation.
Conclusions and Relevance
Public insurance, lower income, and female sex are independently associated with repair of asymptomatic umbilical hernias in children earlier than recommended by current guidelines. These children may be at greater risk of undergoing repair of umbilical hernias that may spontaneously close with further observation.
Umbilical hernia is a common condition in children that affects approximately 800 000 infants in the United States each year.1 The incidence in black children is more than 8-fold higher compared with white children (25% vs 3%) but is highly variable depending on gestational age and degree of African ancestry.2-4 It has been estimated that most children with umbilical hernias will not require surgical repair, as more than 85% of all hernias will spontaneously close by age 4 or 5 years.5-8 Complications, such as incarceration that require emergent or urgent repairs, are relatively uncommon events,9-12 with contemporary data suggesting that less than 2% of all repairs in childhood are associated with an emergency department (ED) presentation.13 Based on these epidemiologic data, current guidelines do not recommend routine surgical repair for children with asymptomatic umbilical hernias until they reach age 4 or 5 years.14-16
In a 2018 multicenter study of umbilical hernia repair including 38 freestanding children’s hospitals,13 nearly 30% of all children were 3 years or younger at the time of surgery, and only 3.4% of all repairs in this age group were associated with an ED presentation. Significant variation in compliance with age-specific guidelines was observed across hospitals, suggesting that differences in local practice and referral patterns or disparities in surgical care may play an important role in the observed variation. Surgical disparities have been attributed to sociodemographic factors, including race/ethnicity and insurance status, the latter of which serves as an indicator of access to care.17-19 However, to our knowledge, the association of patient-level sociodemographic factors with the risk of early umbilical hernia repair within and across hospitals has not been well described. The purpose of this study was to examine the association of sociodemographic factors with adherence to age-specific guidelines for the management of asymptomatic umbilical hernias in children and to explore whether patient-level factors are associated with an increased risk of emergent presentation, which could justify earlier repair for certain at-risk populations.
This was a multicenter retrospective cohort study of children undergoing umbilical hernia repair at children’s hospitals participating in the Pediatric Health Information System (PHIS) database. The PHIS database is an administrative database managed by the Children’s Hospital Association, Lenexa, Kansas, that contains detailed patient care data from 51 freestanding children’s hospitals across the United States. The database includes demographic and payer information, primary and secondary International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM diagnostic and procedural codes, and date-stamped billing data, including Current Procedural Terminology (CPT) codes, for a wide range of clinical services, including diagnostic tests and therapeutic procedures. Data are screened quarterly for accuracy by participating hospitals, the Children’s Hospital Association, and an independent data manager. The Institutional Review Board of Boston Children’s Hospital approved this study. Informed consent was not required because of the deidentified nature of the patient data that were used.
Patients 17 years or younger with a primary diagnosis of any umbilical hernia (ICD-9-CM diagnostic codes 551.1, 552.1, and 553.1; ICD-10-CM diagnostic codes K42.0, K42.1, and K42.9) who underwent umbilical hernia repair (ICD-9-CM procedure code 5349; ICD-10-CM procedure code OWQFOZZ; CPT codes 49580, 49582, 49585, and 49587) between January 1, 2013, and June 30, 2018, were identified. Patients with missing sociodemographic information, those undergoing multiple procedures, inpatients with a preoperative length of stay greater than 2 days, inpatients with a postoperative length of stay greater than 4 days, and those undergoing surgical repair of recurrent umbilical hernia were excluded. Hospitals were excluded if they did not submit sociodemographic data or billing information for identifying ED encounters (Figure 1).
Patients were categorized as having an early umbilical hernia repair if they underwent surgical repair at 3 years or younger. Clinical presentation was categorized as emergent or urgent if repair was performed during the same hospital encounter or within 2 weeks if discharged from the ED without immediate repair, respectively. Urgent presentation was included in this group as an additional surrogate marker of incarcerated or symptomatic disease based on the reported practice of delaying repair 1 to 2 weeks for patients who present incarcerated but undergo successful manual reduction.20 The intent of including urgent presentation was to provide a more liberal definition of repairs preceded by an emergent or urgent presentation with the intent of overestimating, rather than underestimating, the proportion of children undergoing early repair that may have been indicated for incarcerated or symptomatic disease. Sensitivity analyses were performed using different time intervals to define urgent presentation (1-week and 4-week delays between ED presentation and repair) to account for variation in delayed repair timeframes following ED presentation across hospitals. Of note, ICD-based and CPT–based codes indicating complicated umbilical hernias (associated with obstruction or gangrene) were not used for categorizing clinical presentation owing to the subjective nature of these codes, especially when applied to children evaluated in the ED. Predictor variables captured for all patients included sex; race/ethnicity, stratified into white, black, Hispanic, and other; insurance type, stratified into commercial (health maintenance organization, preferred provider organization, and TRICARE), public (in-state and out-of-state Medicaid, Medicare, and Children’s Health Insurance Program), and self-pay/charity; median household income by zip code, stratified into quintiles; and the presence of at least 1 complex chronic condition using the validated PHIS complex chronic conditions flag.21
Univariate, mixed-effects logistic regression models were used to test for associations of patient characteristics and socioeconomic factors with the binary outcome of early repair. Models included random intercepts to control for hospital clustering. Multivariable, mixed-effects logistic regression was used to examine the association of the final covariates with early repair after adjusting for emergent or urgent presentation and hospital-level random effects. Final model covariates were chosen based on a priori hypotheses or a univariate P value less than .05 and included sex, race/ethnicity, insurance type, income quintile, and presence of at least 1 complex chronic condition. The multivariable model was repeated in the subgroup of children with emergent or urgent presentations to examine whether sociodemographic factors were associated with early repair in this higher-acuity cohort. Logistic regression was also used to estimate hospital-specific odds ratios (ORs) of early repair to explore variation in the association of sociodemographic factors with early repair at the hospital level. Hospitals with ORs and 95% CIs that did not include 1.00 were considered to be statistical outliers relative to their peer group of hospitals. All statistical analyses were performed using SAS version 9.4 (SAS Institute), and 2-sided P values less than .05 were considered statistically significant.
The final study cohort included 25 877 children undergoing umbilical hernia repair from 47 children’s hospitals (median [interquartile range] annual case volume, 94 [55-133] per hospital) (Figure 1). The median (interquartile range) age was 5.0 (3.0-6.0) years. Most children were female (13 817 [53.4%]), black (12 734 [49.2%]), and publicly insured (14 143 [54.7%]) (Table). The median (interquartile range) household income was $41 692 ($32 696-$55 223). Overall, 7650 of 25 877 children (29.6%) were 3 years or younger at the time of repair; of these, 286 early repairs (3.7%) were associated with an emergent or urgent presentation (emergent, 209 [2.7%]; urgent, 77 [1.0%]).
Nonwhite race/ethnicity, public insurance, lower income, female sex, and the presence of complex chronic conditions were significantly associated with early umbilical hernia repair (Table). Compared with white race, black race was associated with 25% increased odds of early repair, whereas Hispanic ethnicity was associated with 32% increased odds of early repair. Public insurance was associated with 61% increased odds of early repair and self-pay/charity status with 58% increased odds of early repair compared with commercial insurance. All 4 lower income quintiles were associated with significantly increased odds of early repair compared with the highest income quintile, with the lowest income quintile associated with 83% increased odds of early repair. Female sex was associated with 20% increased odds of early repair compared with male sex (Table).
Public insurance, lower income, and female sex were all associated with significantly increased odds of undergoing early umbilical hernia repair after adjustment for other patient characteristics, socioeconomic factors, emergent or urgent presentation, and hospital-level effects (Figure 2A). Following adjustment, race/ethnicity was no longer associated with increased odds of early repair. Insurance type and income level were additive effects for increased odds of early umbilical hernia repair; children with public insurance in the lowest income quintile had 2.2-fold increased odds of early repair compared with children with commercial insurance in the highest income quintile (OR, 2.15; 95% CI, 1.93-2.40; P < .001). In the subgroup analysis of children who underwent repair associated with an emergent or urgent presentation, no sociodemographic factors were associated with increased odds of early repair (Figure 2B). The results were no different in both sensitivity analyses when urgent presentation was defined as repair within 1 or 4 weeks of an ED presentation.
At the hospital level, no individual hospital or group of hospitals accounted for the overall association of public insurance, lower income, or female sex with the outcome (Figure 3 and Figure 4). Compared with commercial insurance, public insurance was associated with increased odds of early repair at 43 of 47 hospitals (91%). Compared with the highest income quintile, the lowest income quintile was associated with increased odds of early repair at 38 of 45 hospitals (84%). Compared with male sex, female sex was associated with increased odds of early repair at 36 of 47 hospitals (77%).
In this multicenter study of 25 877 patients from 47 children’s hospitals, public insurance, lower income, and female sex were all independently associated with decreased compliance with age-specific guidelines for asymptomatic umbilical hernia repair. Insurance type and income appeared to have additive effects, with public insurance and the lowest income quintile together being associated with a greater than 2-fold increased odds of early repair compared with children with commercial insurance in the highest income quintile. No individual or group of hospitals appeared to account for the overall association of these sociodemographic factors with the increased odds of early repair.
To our knowledge, the present study is the first to examine the association of patient characteristics and socioeconomic factors with age at repair for children undergoing elective, emergent, and urgent umbilical hernia repairs. Differential associations of race/ethnicity, insurance type, and sex with timing of surgery have been previously reported for other pediatric surgical procedures with age-related guidelines.22-24 However, to our knowledge, only one of these studies characterized a paradoxical association similar to the present analysis of markers of lower socioeconomic status with earlier surgical intervention.24 In this study of 2989 children with ureteropelvic junction obstruction (managed expectantly or operatively based on symptoms and disease severity), black race, Hispanic ethnicity, public insurance, and male sex were all found to be independent predictors of earlier age at pyeloplasty. The investigators hypothesized that nonwhite infants may undergo potentially unnecessary surgery at an earlier age owing to concern surrounding their family’s ability to comply with surveillance visits necessary for expectant management.24 Similarly, the results of the present study suggest that nonclinical factors may be associated with potentially unnecessary surgery in children with umbilical hernias that may have spontaneously closed with further observation. In aggregate, these findings align with the growing body of evidence suggesting that nonclinical patient-level and clinician-level factors may influence clinical management decisions, which in turn can result in disparities in care for certain at-risk populations.25,26
The findings of this study may be unexpected in the context of existing data that more often demonstrate an association of markers of lower socioeconomic status with delayed or reduced access to surgical care.22,23,27 The paradoxical increase in risk of early repair associated with public insurance and lower income may be explained by several considerations. Public insurance or lower income may be associated with a higher risk of symptomatic disease, which would justify early repair to avoid the risk of incarceration. However, no association was identified between insurance type or income level and rates of early repair after ED presentation (a surrogate marker of symptomatic disease) among children in this study. Public insurance or lower income may also be associated with increased likelihood of referral for outpatient surgical consultation following ED presentation. In the pediatric population, public insurance has been shown to be independently associated with increased ED use (ie, 1 or more annual ED visits) and high ED reliance (ie, ED visits represent a significant proportion of all health care visits).28,29 Lower income has also been shown to be independently associated with high ED reliance, with children in the lowest income quartile accounting for a disproportionately higher rate of all pediatric ED visits.28-30 Compared with children evaluated in primary care clinics, children evaluated in the ED may be more likely to be referred for outpatient surgical consultation, especially younger children if ED clinicians have concerns over primary care access and a family’s ability to comply with safe expectant management without surgical follow-up.
In the cohort of children undergoing surgical consultation, several considerations may explain why public insurance or lower income may increase a child’s risk of early repair. Parents of lower-income households may be less educated or perceive less control over health care decisions and therefore be less likely to question a surgeon’s recommendation for early repair.31-33 Surgeons may be concerned that a child with public insurance may not have access to routine primary care and therefore lack follow-up for safe expectant management.24,34,35 Early repair may be perceived as the most effective means to protect a child from future complications, especially if the surgeon believes the hernia is unlikely to spontaneously close.
Insurance coverage and other financial incentives may also play a role in decision-making around timing of surgical repair. Public insurers may be more likely to cover the cost associated with an early repair, in contrast to some commercial insurers that may not authorize the repair because they consider it medically unnecessary until a child reaches age 4 or 5 years. Similarly, cost sharing (eg, copayments, high deductibles) may incentivize parents with commercial insurance to continue expectant management, while parents with public insurance with limited or no cost sharing may be incentivized to pursue early repair.36 In addition to increased ED use, differential reimbursement and cost sharing considerations may in part explain the additive association of public insurance and lower income with the risk of early repair.
In addition to socioeconomic factors, female sex was also found to be independently associated with early repair. This association may be driven by both outdated indications for repair in girls as well as cosmetic concerns on behalf of parents. Primary care clinicians may request earlier surgical consultation or surgeons may recommend early repair based on older literature recommending repair in female patients older than 2 years.37,38 However, there is no consensus that female sex increases the risk of short-term or long-term complications and is therefore not recognized as an indication for early repair under current guidelines.14-16,39 The cosmetic appearance of an umbilical hernia in girls may also produce significantly more parental anxiety compared with similar defects in boys. Although cosmetic concerns are unlikely to result in surgical consultation from the ED, primary care clinicians may be compelled by anxious parents to refer a girl with an umbilical hernia for surgical consultation.15,40 Although surgeons are likely to differ in their response to parental anxiety over cosmetic appearance as an indication for early repair, some may offer surgery to alleviate perceived anxiety on behalf of the family.
The findings of this study should be considered within the context of its limitations. The number of children 3 years or younger with umbilical hernias that were managed expectantly was unknown, as only patients undergoing surgical repair were included in this study. As a result, the true overall and hospital-level association of sociodemographic factors with risk of early repair could not be calculated. However, when considering the variability by predictor in the proportions of early repairs performed at the hospital level, these results likely provide a reasonable surrogate for the true association of these predictors with increased risk of early repair. Symptoms and other nonclinical factors that may have influenced a surgeon’s decision to offer an early repair are not available in the PHIS database. As a result, timing of umbilical hernia repair following an ED presentation was used as a surrogate marker to identify children with likely incarcerated or symptomatic disease requiring urgent repair. It is plausible that some children may have been referred to ambulatory surgery clinics for symptomatic hernias and then undergone early elective repair. Similarly, some children may have undergone early repair for incarcerated hernias following urgent referral to an ambulatory surgery clinic rather than the ED. These patients would be misclassified as early elective repairs in this study, resulting in an underestimation of the proportion of repairs performed for incarcerated or symptomatic disease and inaccurate adjustment for disease severity in our multivariable analysis. However, these and other potentially unmeasured or misclassified cases would likely account for only a very small proportion of the children included in this study. Data contained in the PHIS database are retrospectively collected and may be subject to inherent data quality issues, including miscoding or missing data. Additionally, median household income by zip code (aggregate-level data) was used to infer the household incomes of individual patients. However, study participants were categorized into broader income quintiles to account for potential income variability within zip codes.
Public insurance, lower income, and female sex were all found to be independently associated with repair of asymptomatic umbilical hernias in children earlier than recommended by current practice guidelines. Strategies to eliminate disparities based on nonclinical sociodemographic factors may include educational efforts and outreach programs that target not only high-risk children but also the primary care clinicians and surgeons who care for them. Further investigation should explore how existing differential reimbursement mechanisms may inadvertently support the practice of potentially unnecessary surgery in publicly insured children and whether policy changes surrounding insurance coverage may reduce disparities in care for this at-risk cohort.
Accepted for Publication: February 13, 2019.
Corresponding Author: Shawn J. Rangel, MD, MSCE, Department of Surgery, Boston Children’s Hospital, Harvard Medical School, 300 Longwood Ave, Fegan 349, Boston, MA 02115 (firstname.lastname@example.org).
Published Online: May 6, 2019. doi:10.1001/jamapediatrics.2019.1061
Author Contributions: Drs Hills-Dunlap and Rangel had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Hills-Dunlap, Graham, Anandalwar, Rangel.
Acquisition, analysis, or interpretation of data: All authors.Drafting of the manuscript: Hills-Dunlap, Melvin, Rangel.Critical revision of the manuscript for important intellectual content: Hills-Dunlap, Graham, Kashtan, Anandalwar, Rangel.Statistical analysis: Hills-Dunlap, Melvin, Graham, Kashtan, Anandalwar.Obtained funding: Rangel.Administrative, technical, or material support: Anandalwar, Rangel.Study supervision: Rangel.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Hills-Dunlap was supported by grant T32HD075727 from the Eunice Kennedy Shriver National Institute of Child Health & Human Development as part of the Harvard-wide Pediatric Health Services Research Fellowship Program.
Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The content is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health.
Create a personal account or sign in to: