Association between high deductible health plan coverage and age at pediatric umbilical hernia repair

Introduction

Of the 60% of children in the USA who have private health insurance coverage, nearly half are enrolled in a high deductible health plan (HDHP).1–3 HDHPs, defined as plans that meet the Internal Revenue Service (IRS)’s deductible threshold for a health savings account-qualified plan, which for family coverage in 2021 was $2800, have become increasingly common in recent years.2 4 HDHPs confer some advantages over traditional health plans, namely, that they are associated with lower premiums and reductions in overall aggregate healthcare costs.5–7 They have also been associated with other beneficial effects, such as reductions in low-acuity emergency department visits and low-value imaging and laboratory services and increases in generic drug use.5 8–10 However, HDHPs also have some negative effects, such as driving down the use of necessary healthcare.11–13 Many studies have reported decreases in the use of both necessary and unnecessary care among adults with HDHPs.13–16 Rigorous quasi-experimental studies have found that HDHPs are associated with lower medication adherence in adults with chronic diseases,17–20 delays in care-seeking for chronic disease-related complications,21 22 and fewer outpatient office visits.8 18 20 However, to date, there has been very limited research into the effects of HDHPs on pediatric healthcare use or child health. Data from the National Health Interview Survey indicate that families with HDHPs are twice as likely as similar families with lower deductible commercial health insurance plans to report delaying or forgoing medical care for their children due to cost in the last year.3 With regard to specific types of healthcare services, there is evidence that HDHPs are not associated with reduced use of well-child visits.23 However, the association between HDHPs and use of pediatric elective surgery has yet to be determined.

Umbilical hernia repair (UHR) is an elective pediatric surgical procedure that is frequently unnecessarily performed in young children, as most pediatric umbilical hernias close spontaneously within the first few years of life.24 Current guidelines recommend careful observation until children are 4 years of age with the goal of also reducing pediatric anesthesia-related risks, especially repeated anesthesia in the setting of hernia recurrence.25 However, nearly 30% of pediatric UHRs are performed in children under 4 years of age.26 As HDHPs and out-of-pocket costs have been shown to motivate patients and families to delay or avoid medical care, patients with HDHPs may be less likely to undergo UHR before 4 years of age. Out-of-pocket cost considerations may also motivate families to delay UHR even beyond 5 years of age, perhaps long after it is recommended by a physician to be performed. It is also possible that physician preferences play a greater role than parental preferences in the timing of UHR, from the decision to refer for surgical consultation to the pursuance and timing of surgery after surgeon consultation. As a result, it may be that out-of-pocket cost considerations have little effect on age at pediatric UHR. This study aimed to evaluate whether children with HDHPs, as compared with other types of commercial health plans, are less likely to undergo UHR before 4 years of age but are more likely to undergo UHR after 5 years of age. We also aimed to evaluate whether children with HDHPs more often undergo UHR toward the end of the calendar year, when their families are more likely to have already met their health plan’s deductible.

Methods

Results

The study cohort consisted of 8601 children. Table 1 shows the characteristics of these children at the time of their UHR, overall and in groups defined by HDHP versus non-HDHP insurance coverage. The median age was 5 years in both groups, and no significant difference was found between the two groups in terms of the outcomes of interest, namely, being less than 4 years of age or being greater than 5 years of age at the time of UHR. However, geographical region, metropolitan area size, and year were all found to be associated with HDHP status. No association was found between having a chronic condition or undergoing a concurrent surgical procedure at the time of UHR and HDHP status.

Table 2 shows that there were no associations between our instrumental variable and patient sex, whether any chronic condition was present, or whether any concurrent surgical procedure was performed. Thus, these characteristics among children undergoing UHR did not vary across metropolitan areas according to these areas’ rates of HDHP coverage among privately insured children. However, some population-level characteristics of metropolitan areas did vary across quartiles of the instrument. Namely, the MSA/year-level median household income among families with children and the percentage of children of non-Hispanic white race/ethnicity both increased with increasing quartiles of the instrument. As expected, the percentage of included children who had HDHP coverage increased with increasing quartiles of the instrument. An F test from the first-stage model also demonstrated the strength of the instrument, as the instrument contributed significantly to the instrumental variable model (F_1,8031=189.49, p<0.001).

Two-stage least squares regression analysis incorporating the instrument demonstrated no significant association between HDHP coverage and the probability of a child undergoing UHR at an age less than 4 years, with the difference in this probability being 2.4% (–13.0% to 17.7%, p=0.76). Similarly, two-stage least squares regression analysis incorporating the same instrument demonstrated no significant association between HDHP coverage and the probability of a child undergoing UHR at an age greater than 5 years, with the difference in this probability being 1.4% (–15.7% to –18.5%, p=0.87). The results of both analyses were similar when they were restricted to children who were 10 years of age or younger at the time of their surgery (table 3).

Furthermore, when CDHPs and HDHPs were evaluated separately, neither type of health plan was associated with the probability of a child undergoing UHR at an age less than 4 years or the probability of a child undergoing UHR at an age greater than 5 years (online supplemental appendix table 3). In a final sensitivity analysis that excluded children who underwent any other surgical procedure concurrent with their UHR, instrumental variable analysis again demonstrated no significant association between HDHP coverage and the probability of a child undergoing UHR at an age less than 4 years or the probability of a child undergoing UHR at an age greater than 5 years (online supplemental appendix table 4).

Although HDHP coverage was not associated with delaying UHR until after 4 years of age, figure 1 shows that children with HDHPs compared with other types of commercial health plans were more likely to undergo UHR in November or December versus earlier in the same calendar year. In regression analysis, we found that children with HDHPs were 12% (95% CI 1% to 23%) more likely than children with other types of commercial health plans to undergo surgery in November or December versus earlier in the same calendar year. This was the case for children who were under 4 years of age at the time of surgery as well as for children older than age 5 years (online supplemental appendix figures S1 and S2).

Request permission

Discussion

This study found that, despite the greater incentive of families with HDHPs compared with traditional commercial health plans to avoid or delay medical care, HDHP coverage was not associated with a lower likelihood of children undergoing UHR at less than 4 years of age or a higher likelihood of children undergoing UHR after 5 years of age. Our results are somewhat inconsistent with studies of use of elective surgeryin adults. For example, a 2019 study by Chhabra et al showed that use of bariatric surgery and cost sharing is inversely related.31 While our analysis did not reveal an association between delaying a common elective surgery in children and HDHPs, we did see results consistent with other studies in terms of the seasonal variation in use of elective surgery. The study by Chhabra et al and others reported an increase in use of elective surgery at the end of the calendar year.31 This is thought to reflect an increase in the number of privately insured persons with met deductibles toward the end of the insurance year who wish to undergo elective procedures before their annual deductible resets. Pediatric procedures, including UHR, are not exempt from this end-of-year uptick. A recent study comparing privately insured to publicly insured children showed that the so-called ‘December effect’ was present for numerous elective pediatric procedures, including UHR, among privately insured children overall.32 Our study was consistent with these results and showed that the end-of-year uptick in UHRs among privately insured children is particularly high among those with HDHPs. This suggests that families’ anticipated cost sharing may influence their decision making about the timing of UHR once a decision is made to pursue surgery but perhaps does not greatly influence decision making about whether a child should undergo surgery versus wait for the hernia to potentially resolve on its own.

Previous studies have identified several patient and family characteristics associated with early asymptomatic UHR in children. In one recent study, lower income, public health insurance, and female sex were found to be associated with hernia repair occurring at an age under 3 years. The association with public health insurance was interpreted as suggesting that the absence of cost sharing in Medicaid may incentivize unnecessary early operations in children with Medicaid coverage, whereas the often high-cost sharing required of privately insured families for surgical care may incentivize waiting.33 While our study did not compare publicly versus privately insured children, we did find that HDHPs, as compared with other types of commercial health plans, were not associated with a lower likelihood of undergoing UHR before 4 years of age or a higher likelihood of undergoing UHR at age 6 years or older. This suggests that families’ high-cost sharing may not be a strong driver of decision making about whether and at what age to pursue UHR in children. Instead, physician practice patterns may play a greater role in the timing of UHR, from the decision to refer for surgical consultation to the pursuance and timing of surgery after surgeon consultation. Primary care providers’ practices surrounding criteria for surgical referral and surgeons’ practices surrounding the timing of surgery vary based on a variety of factors, including differences in training, financial incentives, perceived risk of incarceration or spontaneous closure, and personal responses to parental concerns and anxiety.25 34–36 However, given the risks associated with unnecessary surgery, the risks associated with additional anesthesia in the setting of reoperation for recurrence,25 and the increased rates of readmission and postoperative complications in children who undergo UHR before 4 years of age,26 37 it is critical to provide education with regard to best practice guidelines and continued advocacy for watchful waiting in children under 4 years of age.

This study was limited primarily by a lack of availability of some important data within the Marketscan database. Clinical data such as severity and frequency of symptoms at presentation, ease of reduction, size of the fascial defect, progression over time, and other important clinical details were not available, although these factors likely would not differ across groups defined by commercial health insurance plan type. As in any administrative database, there was also likely some degree of miscoding of diagnoses and procedures. However, per the specifications of the data use agreement, we were unfortunately unable to identify patients to conduct chart review validation. We were also unable to evaluate patient-level socioeconomic status or race/ethnicity and therefore could not investigate how our findings differed according to these factors. While we did not find an association between HDHP status and the avoidance of early UHR, it is possible that there may be such an association among lower income or racial/ethnic minority families, who are often less able to bear the burden of high cost-sharing. We were, however, able to evaluate median income and racial/ethnic distribution at the MSA/year level. At this level, we found that median household income among families with children and the percentage of non-Hispanic white children increased with increasing HDHP prevalence. Since both factors are associated with higher healthcare use in general,38 39 they were used as exogenous variables in our instrumental variable analysis. However, future studies that account for patient-level socioeconomic status and race/ethnicity would be valuable. Unfortunately, we were also not able to assess surgeon preferences or practice patterns in this study because the majority of the included cases (>60%) lacked a surgeon national provider identifier in the database that could be used to track how often a particular surgeon had performed UHR in children under 4 years of age or what proportion of children of different ages who had consulted with a particular surgeon about UHR then proceeded to undergo the procedure. Future studies that evaluate the relative importance of parent, surgeon, and primary care physician preferences on age at pediatric UHR would be valuable. Last, our inclusion criterion requiring MSAs to have no more than 10% missing data on health plan type in any year of the study period resulted in the exclusion of 36 MSAs (representing 2170 patients). Although this is a sizeable number, the remaining included MSAs still represented a large and geographically diverse population across 35 states. We did not use multiple imputation to impute missing health plan type because, unfortunately, the Marketscan database lacks the information needed to accurately impute this, especially employer IDs and data on the types of health plans offered by employers in a given year.

In conclusion, we did not find that HDHP coverage is associated with a lower likelihood of children undergoing UHR before 4 years of age. Thus, our results do not suggest that HDHPs reduce potentially unnecessary surgical procedures or reduce deviation from best practice guidelines for the management of pediatric umbilical hernias. It is possible that future investigations of this or other procedures may yield different results, as the prevalence of HDHPs and the average deductible size in HDHPs have continued to rise in recent years.3 7 40 However, further research should investigate other factors that can be targeted to reduce unnecessary UHRs in young children, in accordance with best practice guidelines.