The Impact of Payment Reform on Pediatric Craniofacial Fracture Care in Maryland

Abstract

Study Design: Retrospective, quasi-experimental difference-in-differences investigation. Objective: Pediatric craniofacial fractures are often associated with substantial morbidity and consumption of healthcare resources. Maryland’s All Payer Model (APM) represents a unique case study of the health economics surrounding pediatric craniofacial fractures. The APM implemented global hospital budgets to disincentivize low-value care and encourage preventive, community-based efforts. The objective of this study was to investigate how this reform has impacted pediatric craniofacial fracture care in Maryland. Methods: Children (≤18 years) receiving inpatient craniofacial fracture-related care in Maryland between January, 2009 through December, 2016 were investigated. New Jersey was used for comparison. Data were abstracted from the Kid’s Inpatient Database (Healthcare Cost and Utilization Project). Results: Between 2009–2016, 3,655 pediatric patients received inpatient care for craniofacial fractures in Maryland and New Jersey. Prior to APM implementation, around 20% of Maryland patients received care outside of urban teaching hospitals. After APM implementation, less than 6% of patients received care outside of urban teaching hospitals (p = 0.003). Implementation of the APM in Maryland also resulted in fewer pediatric craniofacial fracture admissions than New Jersey, though this only reached borderline significance (adjusted difference-in-differences estimate: −1.1 fewer admissions, 95% confidence interval: −2.1 to 0.0, p = 0.05). Inpatient costs for pediatric craniofacial care and mean did not change post-APM. Conclusions: Maryland’s APM consolidated pediatric craniofacial fracture inpatient care at urban, teaching hospitals. Inpatient costs and lengths of stay did not change after policy implementation, but overall admission rates decreased. Such considerations are important when considering national expansion of global hospital budgeting.

Introduction

Pediatric craniofacial fractures, though relatively rare, are associated with substantial morbidity and high consumption of healthcare resources. These injuries often confer a greater risk for readmission than other traumatic pediatric injury types and require long-term medical follow-up.[1] Children with craniofacial fractures have been repeatedly demonstrated in the literature to impose a large economic burden on healthcare systems.[2] In light of an increasingly cost-conscious healthcare climate, it is important to understand how healthcare payment reform has impacted care for patients presenting with this challenging and expensive, yet largely preventable, injury type.[3]

Maryland is a unique case study of pediatric craniofacial fracture care in a policy-driven, cost-constrained system. In 2014, Maryland implemented the All-Payer model (APM), a payment innovation initiative with a tripartite mission: provision of better healthcare, promotion of population health, and reduction of healthcare costs for all Maryland patients.[4] Specifically, the APM reduced healthcare spending through the Global Budget Revenue program, a policy that caps annual hospital reimbursement.[5] Under this model, every hospital is allocated an annual budget for hospital-based care to control costs.[6] In addition to mandating statewide global hospital budgeting, the APM also disincentivized low-value expenditures such as 30-day hospital readmissions.[7] In this way, the APM created a powerful impetus to reorganize healthcare toward more community-focused, preventative efforts.

The Maryland APM may hold important implications for pediatric craniofacial fracture care. On one hand, greater population health-based efforts may prevent these injuries, which are largely due to falls and avoidable trauma, and may improve longitudinal care coordination for children who do suffer from craniofacial fractures. On the other hand, pediatric craniofacial fractures are resource-intensive injuries, and cost-controlling policies may adversely impact access to craniofacial fracture care for children. As globally-budgeted hospital payment structures are being scaled on a national level, investigation of this payment reform’s impact on pediatric craniofacial fractures is increasingly critical to investigate.[8] This study was therefore undertaken as a retrospective investigation of the impact of Maryland’s APM on the incidence, management, and costs associated with craniofacial fractures in pediatric patients.

Methods

Data Sources

Maryland and New Jersey data were extracted from the Healthcare Cost and Utilization Project Kid’s Inpatient Database (HCUP KID). HCUP KID is the largest publicly-available all-payer database of pediatric inpatient care in the United States. The HCUP KID yields estimates of hospital inpatient stays for pediatric patients younger than 21 years of age, nationwide. The HCUP KID database allows for national and regional investigation of healthcare utilization, access, costs, and outcomes for various pediatric conditions. Pediatric patients admitted for craniofacial fracture-related care were identified using the diagnostic codes listed in Supplementary

Table 1. Demographic and clinical characteristics of children admitted for craniofacial fractures between 2009-2016 in Maryland (n = 1,865) and New Jersey (n = 1,790), stratified by pre- and post-APM timepoints.

. Treatments were determined using procedure codes listed in Supplementary

Table 2. Difference-in-Differences Analysis of Pediatric Craniofacial Admissions in Maryland and New Jersey Between 2009-2016 (n = 3,655).

. It should be noted that HCUP KID diagnosis and procedure codes are recorded using the International Classification of Diseases 9th and 10th revisions.

Study Design: Maryland Versus New Jersey

A retrospective, quasi-experimental difference-in-differences investigation of pediatric patients (18 years or younger) in Maryland and New Jersey was conducted. New Jersey was used as a comparator state. Only residents of each state were included in study analyses. This study was determined to be exempt by the Johns Hopkins University Institutional Review Board. Inclusion criteria included the following: (1) pediatric patient in Maryland or New Jersey (18 years or younger), who (2) received inpatient care for craniofacial fractures between January, 2009 and December, 2016. New Jersey and Maryland are similar in terms of social, economic, and demographic population characteristics, per data from the 2012-2016 American Community Survey published by the United States Census Bureau.[9] Additionally, both Maryland and New Jersey expanded Medicaid under the Affordable Care Act in 2014. However, New Jersey does not mandate statewide global hospital budgeting, in contrast to the APM in Maryland. Thus, comparison of New Jersey and Maryland allowed for the determination of the specific impact of the APM on pediatric craniofacial injuries and care.

Investigations

Two main investigations were conducted: (1) the impact of Maryland’s APM on pediatric craniofacial fracture prevention and (2) the impact of Maryland’s APM on pediatric craniofacial fracture treatment.

To investigate the impact of Maryland’s APM on the prevention of pediatric craniofacial fractures, we sought to determine whether there were changes in admission rates for this injury type pre- and post-implementation of the APM using a quasi-experimental difference-in-differences model.[10] The proportion of all pediatric inpatient admissions that were a result of craniofacial fractures pre- and post-reform in 2014 were tabulated. Patient (demographics, measures of injury severity), hospital, and community-level factors were used to generate risk-adjusted admission rates in both Maryland and New Jersey. These analyses were adjusted for age, sex, race/ethnicity, and All Patients Refined Diagnosis Related Groups (APR DRG) risk of mortality score, as well as a linear quarterly time trend. The APR DRG is a system that classifies hospital inpatients by their admitting diagnosis, severity of illness, and risk of mortality. Clustering at the hospital level was accounted for using cluster-correlated robust estimates of variance. A binary indicator for payment reform status was created to indicate pediatric craniofacial care received in Maryland versus the comparator state (New Jersey), and a second binary indicator variable was created to delineate care that occurred post-2014 versus pre-2014. An interaction term between these 2 indicators served as the difference-in-differences estimator. We assessed parallel trends in outcomes between New Jersey and Maryland prior to 2014 to determine model validity.

Pediatric craniofacial fracture treatment was also investigated pre- versus post-implementation of the APM in Maryland. Operative interventions for craniofacial fractures were tabulated pre- versus post-APM implementation, as were mean costs of care and mean hospital lengths of stay. Pediatric craniofacial fracture care was also assessed at the hospital level. Chi square analyses were used to analyze the distribution of hospital types (urban teaching, urban non-teaching, non-urban) at which pediatric patients received care pre- and post-APM. Maps of pediatric craniofacial fracture-related patient care in Maryland pre- and post-APM were generated using Python 3.7.2 (Python Software Foundation, Wilmington, DE).

Statistical Analyses

All statistical analyses were performed using Stata version 15.0 (StataCorp LLC, College Station, TX). The 2-tailed threshold for significance was set at p values <0.05, with Bonferroni correction for post-hoc analyses. Study design and analyses were constructed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Guidelines and conformed to the Declaration of Helsinki Ethical Principles for Medical Research involving Human Subjects.[11]

Results

All admissions for pediatric craniofacial fractures in Maryland and New Jersey were tabulated between 2009 and 2016. In total, 1,865 pediatric patients in Maryland and 1,790 patients in New Jersey met inclusion criteria. Demographic data for Maryland craniofacial fracture admissions as well as admissions from the comparator state, New Jersey, are reported in Table 1.

Pediatric Craniofacial Fracture Prevention

In Maryland, the proportion of pediatric inpatient admissions that were a result of craniofacial fractures decreased between pre-APM and post-APM time periods (6.8 per 1000 admissions [95% confidence interval: 5.8-7.7] in the pre-APM period to 4.4 per 1000 admissions [95% confidence interval: 3.6-5.2] in the post-APM period). In New Jersey, the proportion of pediatric inpatient admissions that were due to craniofacial fractures also decreased between pre-APM and post-APM time periods (4.5 per 1000 admissions [95% confidence interval: 3.7-5.3] in the pre-APM period to 3.5 per 1000 admissions [95% confidence interval: 2.8-4.2] in the post-reform APM). The adjusted difference-in-difference estimate comparing pre-APM versus post-APM time periods was −1.1 (95% confidence interval: −2.1 to 0.0, p = 0.05; Table 2).

Pediatric Craniofacial Fracture Treatment

Prior to APM implementation, most children received inpatient care for craniofacial fractures at urban teaching hospitals in Maryland. However, around 20% of patients also received care at urban non-teaching hospitals and non-urban hospitals. After APM implementation, less than 6% of patients received care at non-urban hospitals or urban non-teaching hospitals in Maryland (Figure 1; chis quare: p = 0.003). Furthermore, the geographic distribution of hospital-based care for pediatric craniofacial fractures was consolidated at 2 major pediatric trauma centers after APM implementation, despite the fact that the overall geographic distribution of craniofacial fracture patients did not substantially change in the pre- versus post-APM time periods (Figure 2).

In terms of pediatric craniofacial fracture treatment, the average cost of hospitalization (inflated to 2016 dollars) and average length of stay in Maryland did not significantly change between pre- versus post-APM timepoints (

Table 3. Pediatric craniofacial fracture care in Maryland, pre- versus post-APM (n = 1,865).

). The mortality rate in patients admitted for craniofacial fractures also did not significantly differ pre- versus post-APM implementation. Within Maryland, the proportion of admitted patients who received open reduction slightly increased during the post-APM time period (30.3% versus 25.2%, p = 0.04)

Discussion

Pediatric craniofacial fractures are complicated injuries for both patients and health care systems.[12] Surgical management can impact long-term craniofacial and dental development at the fractured segment, requiring longitudinal follow-up.[13] Facial fractures in children also commonly occur in association with head and chest injuries, resulting in considerably higher overall morbidity and longer intensive care unit stays when compared to adult craniofacial fractures.[2,3,14] Resultantly, the average cost of an open reduction, internal fixation procedure for a pediatric craniofacial fracture has been cited to be $84,849, with an average length of stay of 5.3 days.[2] Overall, pediatric craniofacial fractures represent a large economic burden on both healthcare systems and patients; these injuries are estimated to represent $1.2 billion of national healthcare expenditures annually.[2] In light of national movements to limit healthcare spending and to implement cost-controlling payment reform policies such as Maryland’s APM, it is important to investigate the provision and evolution of care for expensive and resource-intensive injuries such as pediatric craniofacial fractures.[15] Such analyses could aid in understanding and evaluating the overall utility of such payment reform measures, and could inform considerations for scale-up of such policies on a national level.

Maryland as a Test Case for National Trends

Maryland’s Global Budget Revenue (GBR) methodology is a central tenet of the APM implemented in 2014.[16] The GBR is a prospective revenue capping system that encourages hospitals to focus on population health.[17] The GBR system presents hospitals with a financial incentive to manage healthcare resources efficiently by providing the highest value of care possible and to slow rates of increase in healthcare costs.

Maryland was the first state to implement such a system on a statewide basis for all its hospitals. While this model has demonstrated success in controlling costs, it is largely unknown how this healthcare payment reform model has impacted care for pediatric craniofacial fractures.[18] This is especially critical to investigate given that in the years since implementation of Maryland’s APM, Vermont, Pennsylvania, and other states have also begun to adopt such globally-budgeted models for accountable care.[19]

Maryland’s All Payer Model Reduced Admissions for Pediatric Craniofacial Fractures

Pediatric craniofacial fractures are most commonly a result of preventable injuries, such as falls, sports injuries, and motor vehicle collisions.[20] The APM was designed to incentivize a fundamental restructuring of care, with a focus on prevention and community services.[21] Official documentation from the Maryland Health Services Cost Review Commission highlights that this model is designed to improve population health through a variety of measures, such as promoting the utilization of primary healthcare services and promoting the creation of safe physical environments.[22] While the proportion of admissions due to craniofacial fractures decreased over time in both New Jersey and Maryland, difference-in-differences analyses demonstrated that Maryland experienced a greater rate of decline than did New Jersey between pre- and post-reform time periods. Since both Maryland and New Jersey expanded Medicaid under the Affordable Care Act (ACA) in 2014, but Maryland also implemented the APM that year, the results from this study suggest that the APM may have resulted in the lower rate of pediatric craniofacial fracture admissions observed in Maryland compared to New Jersey. However, this difference only reached borderline statistical significance after adjusting for demographic factors, injury severity, and linear time trends.

The findings from this study may support the utility of the APM in improving healthcare at the community level by encouraging population health initiatives. For instance, after implementation of the APM, the Maryland Department of Health allocated $3 million to expand access to healthcare in underserved communities.[23] These reflected statewide efforts to align public health and community organizations, in order to help create safer community spaces and greater healthcare accessibility. Furthermore, in the final report of the APM, Maryland authorities found that implementation of the APM incentivized hospitals and providers to place an emphasis on patient education and safety. Given that trauma and fractures are one of the potentially preventable conditions outlined in the official APM agreement signed in 2014 between CMS and the state of Maryland, it is possible that the reduction in inpatient pediatric craniofacial fracture admissions noted in our results was a result of the concerted efforts by Maryland and statewide hospital/healthcare systems to improve metrics in this area.[22]

A limitation of this aspect of the study is that only inpatient admissions were investigated. Thus, it is uncertain whether the same number of patients presented to emergency departments with pediatric craniofacial fractures and a fewer proportion of these patients were admitted for inpatient care, or whether there were fewer craniofacial fractures overall. Either way, fewer admissions for pediatric craniofacial fractures may indicate lesser global severity of injuries that did occur, which could indicate greater community-based preventative efforts. Overall, while the APM may have improved pediatric population health with regard to craniofacial fracture prevention, further investigation is necessary to more comprehensively determine the specific effects of this policy measure. Given that the Centers for Medicare and Medicaid Services has approved an extension of Maryland’s APM, in a new initiative termed the “Total Cost of Care All-Payer Model,” prospective, longitudinal studies of this reform policy can better highlight the specific impacts of such prevention measures on reducing the incidence of craniofacial injuries in children.[24]

Access to Pediatric Craniofacial Fracture Care

As previously mentioned, the APM was implemented in Maryland in the same year that Maryland and many other states across the country, including New Jersey, expanded Medicaid under the ACA. The study results reflect this expansion: in both New Jersey and Maryland, we found that between pre- and post-APM timepoints, the proportion of publicly-insured patients admitted for craniofacial fractures increased, while the proportion of uninsured patients decreased. This trend mirrors patterns seen in other areas of healthcare post-Medicaid expansion, with regard to increased utilization of hospital-based surgical care by publicly-insured patients and by patients who were previously insured but who now received public coverage.[25,26]

It was reassuring to note that implementation of cost-constraining policies under the APM did not appear to hinder any increases in access to care by Medicaid beneficiaries as a result of Medicaid expansion. Pediatric craniofacial fractures can be complicated injuries requiring long-term follow-up care and revision procedures.[1] Publicly-insured children are known to have poorer clinical outcomes than those who have private insurance, which may make their long-term care more expensive.[27] The increase in the number of publicly-insured children receiving craniofacial fracture-related inpatient care post-reform in Maryland suggests that there were provisions within the APM to ensure that hospitals treating children with these injuries were allocated appropriate budgets in order to do so, despite these children’s risk for readmissions/poor outcomes.

It is interesting to observe that, even prior to implementation of APM and Medicaid expansion, Maryland had a greater proportion of publicly-insured children receiving inpatient care for craniofacial fractures (45.4% in Maryland versus 28.0% in New Jersey), despite the fact that the states have similar underlying proportions of Medicaid/Children’s Health Insurance Program beneficiaries. This may be a result of multiple factors, including Maryland’s longstanding policy of all-payer rate setting. Under all-payer rate setting, all payers reimburse at the same rate for procedures, unlike in other states where Medicaid reimburses at rates lower than other payers.[6,28] Additionally, Maryland has a Trauma Physician Services Fund, which helps to offset costs of undercompensated care. These factors may have made Maryland hospitals historically more willing to treat larger volumes of publicly-insured children.

Global Budget Revenues Consolidated Care to Urban Academic Centers

While a substantial proportion of pediatric craniofacial fracture patients were treated at urban academic centers prior to global budgeting, implementation of global budgets in Maryland further consolidated care for these patients at academic hospitals. In fact, despite the fact that the post-APM distribution of patients by zip code in Maryland did not differ substantially from pre-APM distributions, these patients were more likely to receive care at larger urban, academic centers after implementation of the APM. Pediatric craniofacial fractures can be challenging and expensive injury types; urban teaching hospitals likely have greater capacity/budgets to treat these fractures.[29] This reflects trends seen in other fields after implementation of the APM in Maryland as well.[30]

There is a large body of literature demonstrating that treatment of children at high volume tertiary care centers improves outcomes.[31] Pediatric craniofacial fractures are relatively rare, and transferring these patients to tertiary, urban centers may be both clinically and fiscally expedient given that providers in these hospitals have more experience in treating these pathologies. Furthermore, consolidating care at these centers can also facilitate investigation into evidence-based practices for these conditions, helping to improve overall management of these fractures. Thus, the APM may not only serve to enhance care in the short term, but it may also facilitate processes such as research that can improve outcomes in the long run.

In terms of management, an increase in utilization of open reduction was observed across all children hospitalized for craniofacial fractures in Maryland. This may be a result of improved imaging techniques and patient selection for inpatient management of this injury type, rather than a result of payment reform.[2,32] The management of pediatric craniofacial fractures with open versus closed reduction is still a debated topic in the literature, and few guidelines for best practices exist.[2] As mentioned previously, it is a limitation of this study that only inpatient admissions for pediatric craniofacial fractures were considered. Further investigation into all children presenting to the hospital with craniofacial fractures will help to better characterize the impact of payment reform on management strategies for craniofacial fractures.

Other Limitations

Beyond the aforementioned limitations, this study was retrospective in nature, and was therefore dependent on the quality of the database. Thus, study analyses were limited by the variables included in the database. Additionally, capturing appropriate data from the database was dependent on ICD coding specificity, which was limited especially for ICD-9 procedure codes. Some relevant data may not have been tabulated by the codes used in this study, thereby leading to loss of data. However, given that all pediatric craniofacial fracture types were queried, the overall study sample was large enough to withstand the impact of limitations in coding. Also, the same procedure codes were used to evaluate both pre- and post-healthcare reform timepoints, thereby allowing us to form conclusions about overall trends in pediatric craniofacial fractures. Last, the external validity of this study was limited by the fact that only 2 states were investigated, and the expansion of such payment reform to other states may not demonstrate similar patterns/outcomes. However, the data provided by this investigation still provides useful and relevant information with regard to how cost-constrictive payment reform measures such as the APM can impact care for pediatric craniofacial fracture patients. This may be of particular importance as more states decide to adopt such payment reform policies.

Conclusions

Pediatric craniofacial fractures are challenging, resource-intensive injuries. This study investigated trends in inpatient pediatric craniofacial fracture care as a result of Maryland’s unique cost-controlling payment reform measures, implemented under the APM. Implementation of the APM was found to be associated with fewer admissions for pediatric craniofacial fractures and consolidation of fracture care at larger, urban academic hospitals with greater capacity/budgets to handle their more complex presentation. Despite this payment reform’s focus on high-value care, inpatient costs and lengths of stay did not significantly decrease after policy implementation. Further investigation is necessary to determine the specific effects of the APM on the type of care that admitted children receive for their craniofacial fractures, as well as to study outcomes. As more states have conversations to adopt global hospital budgeting, it is important to understand how such payment reform measures can impact care for serious injuries such as pediatric craniofacial fractures.