Presented at the American College of Emergency Physicians Scientific Assembly, Seattle, WA, October 7, 2007. Acknowledgment of grants: Emergency Medicine Foundation, NIH R01 HS013920-01.
Trends in Charges and Payments for Nonhospitalized Emergency Department Pediatric Visits, 1996–2003
Article first published online: 27 MAR 2008
© 2008 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 15, Issue 4, pages 347–354, April 2008
How to Cite
Hsia, R. Y., MacIsaac, D., Palm, E. and Baker, L. C. (2008), Trends in Charges and Payments for Nonhospitalized Emergency Department Pediatric Visits, 1996–2003. Academic Emergency Medicine, 15: 347–354. doi: 10.1111/j.1553-2712.2008.00075.x
- Issue published online: 27 MAR 2008
- Article first published online: 27 MAR 2008
- Received December 3, 2007; revision received December 28, 2007; accepted December 29, 2007.
- health care use;
- emergency department utilization;
- health services research;
- health service utilization
Objectives: To compare charges and payments for outpatient pediatric emergency visits across payer groups to provide information on reimbursement trends.
Methods: Total charges and payments for emergency department (ED) visits Medicaid/State Children’s Health Insurance Program (SCHIP), privately insured, and uninsured pediatric patients from 1996 to 2003 using data from the Medical Expenditure Panel Survey. Average charges per visit and average payments per visit were also tracked, using regression analysis to adjust for changes in patient characteristics.
Results: While charges for pediatric ED visits rose over time, payments did not keep pace. This led to a decrease in reimbursement rates from 63% in 1996 to 48% in 2003. For all years, Medicaid/SCHIP visits had the lowest reimbursement rates, reaching 35% in 2003. The proportion of visits from children insured by Medicaid/SCHIP also increased over the period examined. In 2003, after adjustment, charges were $792 per visit from children covered by Medicaid/SCHIP, $913 for visits from uninsured children, and $952 for visits from privately insured children.
Conclusions: Reimbursements for outpatient ED visits in the pediatric population have decreased from the period of 1996 to 2003 in all payer groups: public (Medicaid/SCHIP), private, and the uninsured. Medicaid/SCHIP has consistently paid less per visit than the privately insured and the uninsured. Further research on the effects of these declining reimbursements on the financial viability of ED services for children is warranted.
Emergency departments (EDs) play a number of key roles in the American health care system, from trauma care delivery to the provision of safety-net services. Over the past decade, however, there has been rising concern (as evidenced by the Institute of Medicine report on the future of emergency care) that the financial position of EDs is threatened. While EDs continue to see yearly increases in visits, efforts to contain health care costs have increasingly constrained their budgets.
Previous work has studied reimbursement patterns for adult patients seen in EDs and documents decline over time in the share of ED charges that are ultimately reimbursed.1–3 Pediatric visits are both a significant component of overall ED visits (they constitute one in every four ED visits4 or about 30 million visits annually5) and a particular source of concern about reimbursement. While the percentage of children insured is higher than that of adults,6,7 much of this coverage is due to Medicaid, the State Children’s Health Insurance Program (SCHIP), or other governmental programs that can have important limitations on ED payments. Many Medicaid programs have also attempted to move their enrollees into managed care arrangements,8 some of which have aggressively moved to reduce ED payments over time. This has led to the concern that pediatric visits have a disproportionately negative impact on ED budgets.
Despite these concerns, current information about reimbursement rates for pediatric ED visits is limited. While there have been numerous studies documenting the ED utilization rates of these patients,8–14 we are aware of only one other article that examines ED charges for children, studying these pediatric patients in the state of Colorado.15 We examined trends in charges and reimbursements for ED visits made by the pediatric population nationally between 1996 and 2003. Using national data, we studied amounts billed and payment rates for privately insured, uninsured, and Medicaid/SCHIP patients.
This was a retrospective study using data from the Household and Medical Provider Components of the Medical Expenditure Panel Survey (MEPS). Our study was reviewed by the institutional review board and deemed exempt from informed consent.
Study Setting and Population
We analyzed data from MEPS, collected by the U.S. Agency for Healthcare Research and Quality, for the years 1996 through 2003. The MEPS provides information about health care utilization and expenditures, insurance coverage, medical conditions, and other characteristics for a nationally representative sample of the non-institutionalized U.S. civilian population.16 Person-level response rates fluctuate by year from 1996 to 2003, ranging from 64% to 71%.17
The utilization and expenditure information reported in the MEPS was compiled using a multipart data collection process. Respondents were asked to use a diary to compile information about their health care use, associated charges, and out-of-pocket and insurance payments. These diaries were reviewed by survey personnel, and respondents are also queried about any additional medical care use. Finally, after obtaining permission from respondents, MEPS researchers frequently contacted providers and insurers to verify the information given.
Data Collection and Processing. For our analysis, we abstracted visit-level data on use of EDs and associated charges and payments from the MEPS database. We selected ED visits for those younger than 18 years of age. In the interest of examining visits for which we could clearly identify charges and payments, we excluded those covered under flat-fee arrangements or for which zero charges were reported. We also excluded visits that resulted in hospital admission, because it is often not possible to disaggregate charges and payments for the ED visit from those associated with inpatient care.
We compared visits by uninsured patients, those covered by Medicaid/SCHIP, and those insured privately. We excluded visits covered by Medicare (because only children with end-stage renal disease requiring dialysis or transplant are covered by Medicare) and other public hospital or physician insurance programs, as well as those for which insurance coverage, could not be identified. Finally, we excluded visits by patients for whom we lacked data required in our risk adjustment regressions. Application of these criteria left us with a sample of 10,202 visits.
Outcome Variables. The outcome variables we examined were total charges and total payments per visit. We defined total charges as the aggregate of facility and physician charges. In the MEPS, this was intended to reflect all verified charges for medical care before negotiated discounts or payments for bad debt or free care. Total charges included diagnostic tests, laboratory work, services, and treatment, but not charges for prescription medications.
Total payments represented the sum of all payments to the facility and to providers treating the patient. They included out-of-pocket payments and those made by private insurers, Medicaid, and SCHIP, but did not capture bonuses or other retroactive payment adjustments from third-party payers. We reported payments and charges in 2003 dollars, converted using the Consumer Price Index.
To examine charges and payments separately for uninsured, privately insured, and Medicaid/SCHIP-covered patients, we assigned each ED visit to a payer group using the respondent’s insurance status for month of the visit. The MEPS identified children as privately insured if the reporting individual for that household indicated that the child had private coverage for hospital and physician services and did not indicate Medicare or Medicaid/SCHIP coverage. Individuals were identified as having Medicaid/SCHIP coverage if Medicaid or SCHIP coverage was indicated. MEPS reported Medicaid and SCHIP coverage as a single category due to the difficulties many respondents had in distinguishing the two programs. Some individuals for whom Medicaid or SCHIP coverage was not initially indicated, but for whom answers to further questions suggested that they did in fact have Medicaid or SCHIP coverage, also had Medicaid/SCHIP coverage recorded on the MEPS.18 Finally, individuals were considered uninsured if they did not indicate having coverage from any source or only indicated coverage that did not provide for physician and hospital services (e.g., dental coverage only).
For each group we calculated estimates of total visits nationally, total ED charges, and total payments, using MEPS sample weights that allowed the MEPS sample to represent the population of the United States. These sample weights were also applied in the regression analyses. Based on these estimates, we estimated unpaid charges as the difference between total charges and total payments, as well as the ratio of total payments to total charges.
We also examined average charges and payments per visit, by insurance status. To account for variations in patient characteristics between insurance groups and over time, we focused on estimates of average charges and payments that are adjusted for patient characteristics. To compute adjusted mean charges and payments, we estimated regression models with payments or charges as the dependent variable and a number of patient characteristics as independent variables. These included relevant risk factors and patient demographic characteristics and are listed in Table 1. In addition, the models accounted for clinical characteristics of the visit using indicator variables for the clinical classification code assigned to the visit. To account for the highly skewed distributions of charges and payments, we estimated the regressions using a generalized linear models approach,19 specifying a log transformation of the dependent variable and a Poisson variance structure.
|All||Medicaid/ SCHIP (n = 4,408)||Private (n = 4,485)||Uninsured (n = 1,309)|
|Income (in relation to FPL)|
|Poor (less than 100% FPL)||22.5||51.7||4.4||25.1|
|Near poor (100% to <125% FPL)||6.5||11.2||3.1||8.9|
|Low income (125% to <200% FPL)||17.8||21.5||13.8||25.8|
|Middle income (200% to <400% FPL)||31.0||13.2||41.7||30.6|
|High income (≥400% FPL)||22.3||2.4||36.9||9.7|
|Gender and age (yr)|
|Female age 0–4||16.8||20.3||14.2||19.0|
|Female age 5–9||10.4||11.2||9.7||11.4|
|Female age 10–14||9.6||8.0||10.1||11.2|
|Female age 15–18||7.6||5.4||8.4||9.9|
|Male age 0–4||20.9||27.7||17.6||17.2|
|Male age 5–9||13.3||12.7||14.3||10.0|
|Male age 10–14||12.8||9.9||14.4||13.5|
|Male age 15–18||8.7||4.9||11.2||7.8|
|Mother’s educational attainment|
|High school degree||51.9||50.0||52.1||56.3|
|Master’s or doctorate||4.8||0.7||8.0||1.6|
|Education data missing||0.2||0.3||0.1||0.3|
|Father’s educational attainment|
|High school degree||37.1||24.8||45.1||34.2|
|Master’s or doctorate||5.0||0.1||8.6||1.9|
|Education data missing||0.4||0.3||0.6||0.2|
Using the results of the regression analyses, we estimated adjusted mean charges and adjusted mean payments, holding all independent variables fixed at their sample means. Based on the adjusted mean payment and charge estimates, we computed the adjusted mean payment rate as the ratio of adjusted mean payments to adjusted mean charges.
Study Population Characteristics
We analyzed a total of 10,202 visits that met our inclusion criteria. Excluded visits accounted for slightly less than 15% of the original sample, with the majority of exclusions (approximately 6%) being visits that resulted in admission, which was not statistically different across the three-payer groups. Overall, 54% of our observations were visits that were privately insured, 33% covered by Medicaid/SCHIP, and the remaining 12% were uninsured. Table 1 shows the study population characteristics by payer group.
Using the MEPS sample weights, the visits observed in the sample can be used to construct estimates of the total number of visits nationwide. Our MEPS sample represents an estimated 87.5 million ED visits made by pediatric patients nationwide between 1996 and 2003, approximately 11 million visits per year (Table 2).
|Year||No.||Estimated Visits||Adjusted Charge||Adjusted Payment||Adjusted Reimbursement Ratio (%)|
As a share of all visits, in 1996, Medicaid/SCHIP patients accounted for 29% of visits; this share increased to 41% in 2003. Visits made by uninsured patients moved in the opposite direction, decreasing from 1.6 million (14% of total visits in 1996) to 1.1 million (9% of total visits in 2003), and visits from patients who were privately insured also dropped from 6.4 million (56% of total visits in 1996) to 5.7 million (49% of total visits in 2003).
Total Charges and Payments
Estimated nationwide aggregate charges for pediatric ED visits increased from $7.4 billion in 1996 to $10.2 billion in 2003, an increase of 39%. Total payments, however, remained relatively constant at approximately $4.9 billion per year, an increase of less than 1% over the 8-year period (Figure 1). Because charges increased while payments did not, unpaid charges in 2003 increased 113% relative to 1996, with $5.3 billion in unpaid ED charges for pediatric patients in 2003. Further subgroup analysis of total payments and charges revealed noteworthy differences in the payment rates (Figure 2). Table A1, a table of aggregate payments and charges from which our reimbursement ratios are based for Figure 2, is included as an online Data Supplement, which also includes the regression results in Table A2 (available at http://www.blackwell-synergy.com/doi/abs/10.1111/j.1553-2712.2008.00075.x).
Per-visit Charges and Payments
Because trends in total payment and charge information can be affected by changes in the volume of visits in each insurance group, we performed visit-level charge and payment analyses to better capture the trends within these subgroups. Despite the higher total values for charges and payments, throughout all years of the analysis, Medicaid/SCHIP patients consistently had the lowest absolute and relative adjusted per-visit charges and payments. In 1996, on average, adjusted payments were $230 and charges were $532, for Medicaid/SCHIP patients, a ratio of 43%. There was a general downward trend in the payment rate. In 2003, on average, payments were $277 and charges were $792 per bill, a ratio of 35%. This phenomenon of essentially flat payments with increased charges was similar in the privately insured and uninsured group, with 2003 per-visit adjusted charges for those privately insured of $952 and payments of $554 and $913 charged for the uninsured with $400 paid. Table 2 provides a summary of the charges and payments per insurance group by year.
The reimbursement ratio for all three groups declined. Figure 2 illustrates the ratios for total charges and payments, and Table 2 shows the ratios for per-visit charges and payments. The reimbursement ratios of the total charges and payments differed slightly from the per-visit ratios due to changes in volume of visits by insurance category, as explained earlier. Both the total and the per-visit ratios, however, exhibited the same declining trend.
Adjusted per-visit charges and payments provide a more direct view of how reimbursement rates changed within insurance groups over time. Overall, the adjusted share of per-visit charges paid went from 63% (95% confidence interval [CI] = 59% to 67%) in 1996 to 48% (95 CI = 45% to 51%) in 2003. All three insurance groups experienced comparable relative declines in payment shares: for Medicare, the payment rate declined from 43% (95% CI = 37% to 49%) to 35% (95% CI = 32% to 38%); for private, from 74% (95% CI = 67% to 81%) to 58% (95% CI = 53% to 64%); and for the uninsured, from 52% (95% CI = 43% to 62%) to 44% (95% CI = 32% to 56%).
Our findings document important trends in the reimbursement for emergency care in nonhospitalized pediatric patients. In particular, we found steadily declining payment rates for all three of the insurance groups we studied over the 1996 through 2003 period. In percentage terms, the payment rate for privately insured patients declined the most, but the payment rate for these patients started at the highest level, as well.
In all three cases, the declining payment rate was largely the result of rapidly increasing charges, whereas payments were not keeping pace. This is somewhat difficult to interpret, because increasing charges could come about for two reasons. First, they may have reflected actual increases in the costs that EDs incurred to provide care, such as introduction of new technologies and increased utilization of services such as computed tomography scanners.20,21 Second, it is possible that at least some of the charge increase was simply the result of higher markups. For example, providers may have responded to declining reimbursement rates by increasing their billing amounts to maintain an absolute level of payments in contracts that specify payments as a percentage of charges.22 If the declining payment rate is a reflection of costs of caring for ED patients not being met, these results could have important and concerning implications for the financial health of EDs. If declining payment rates represent unpaid hospital markups, with the underlying costs still being met, the implications for the financial health of EDs are less clear.
This study did not address reasons that payments have not grown as rapidly as charges. There have been substantial pressures from a variety of places to contain costs, for both private insurers and public programs like Medicaid. It seems quite plausible that these contributed to efforts by payers to adopt payment policies that limited payment growth. The related growth in more restrictive health plans, such as health maintenance organizations (HMOs) and Medicaid managed care organizations, may also play a role, because these seem to reimburse at lower rates compared to standard Medicaid.23
In 2003, the payment rate for Medicaid/SCHIP patients was the lowest of the three groups we studied. In our analysis, we examined both the total unpaid bill and the average nonpayment rate. Medicaid/SCHIP scored poorly on both; that is, Medicaid reimbursed less per visit as a percentage of the per-visit charge, but because Medicaid patients account for a large number of ED visits, Medicaid’s responsibility for the total unpaid bill was also substantial.
The fact that Medicaid/SCHIP reimbursed more poorly than the uninsured contrasts with the general perception that uninsured patients are a large source of unpaid ED charges. It poses the question of whether, from a financial perspective only, it may be that EDs would rather, on average, see uninsured than Medicaid-covered patients. While the Emergency Medicine Treatment and Active Labor Act (EMTALA) requires that all patients must be screened in the ED, the relatively low reimbursement rate for Medicaid/SCHIP patients does suggest that continued attention to Medicaid policy toward ED payments is warranted as cost pressures continue to drive discussions of Medicaid payment limitations.
We observed a substantial increase in visits by Medicaid/SCHIP patients as a proportion of visits in our data. Since SCHIP’s creation in 1997, enrollment of children has grown markedly,24 and there is a good deal of evidence showing that the creation of SCHIP decreased the number of uninsured children, particularly over the time period of this study.7 Visit rates for Medicaid-covered children to EDs have also been increasing, which may be related to declines in the ability of Medicaid-covered children to obtain access to non-ED care as Medicaid payment rates outside of EDs have fallen relative to other payers. 5,15,25–32
While Medicaid/SCHIP visits increased over the study period, we also showed that Medicaid/SCHIP patients consistently had lower charges than those who are uninsured and privately insured. Previous studies show higher hospitalization rates and charges for publicly insured and uninsured children compared with privately insured patients.15 One interpretation of this evidence, although not the only interpretation, is that Medicaid/SCHIP and uninsured children may have more limited access to primary care, leading to increased use of EDs for nonemergent, less resource-intensive care. This interpretation is consistent with a number of studies,5,33–40 although not with recent MEPS-based evidence on the use of ED visits for nonurgent conditions by publicly insured children.41 This study has some important strengths, including the use of detailed information on verified charges and payments specific to individual ED visits, as well as demographics, health status, and insurance information.
The MEPS has been criticized for undercounting total ED visits in the adult population, in relation to the National Hospital Ambulatory Medical Care Survey (NHAMCS).1 This has not been previously documented in pediatric populations but, even if it does exist, would not affect our estimates of trends or group comparisons, unless there is differential undercounting over time or across payer groups, which we have not seen suggested. Furthermore, our estimated sample contained 11 million pediatric visits annually, and using recent work showing 40 to 50 million total annual visits,3 we reach a ratio of one pediatric visit per four to five ED visits. This is consistent with NHAMCS data4 (at least on a relative level, which is all that is required for evaluating trends in the reimbursement ratio). Second, the data did not reflect discounts negotiated by insurers or payments not associated with individual visits. In the former case, we believe it does not bias our results since we want to examine total charges that reflect the cost of care before discounts; in the latter case, our payment rates may have been slightly lower than actual.
The ratio of payments to charges for pediatric outpatient ED visits decreased from the period of 1996 through 2003. Payments did not keep pace with charges in all three payer groups: public (Medicaid/SCHIP), private, and the uninsured. Declining payment rates raise concerns about the ability of EDs to recover costs of providing care for children. Further research is warranted on the effects of these decreasing reimbursement rates on the financial health of EDs.
- 3Decreasing reimbursements for outpatient emergency department visits across payer groups from 1996 to 2003. Ann Emerg Med. 2007 Nov 9 [Epub ahead of print]., , .
- 4National hospital ambulatory medical care survey: 2000 emergency department summary. Adv Data Vital Health Stat. 2002; 326:1–30., .
- 5Reimbursement Issues in Pediatric Emergency and Subspecialty Care. Institute of Medicine presentation, Institute of Medicine, Future of Emergency Care in the U.S. Health System Sub-Committee on Pediatric Emergency Care. Washington, DC, Sep 21, 2004..
- 6Health Insurance Coverage: Estimates from the National Health Interview Survey, 2005. Available at: http://www.cdc.gov/nchs/nhis.htm. Accessed Sep 3, 2006., .
- 16Agency for Healthcare Research and Quality. The MEPS Data and Methods Manual. Available at: http://www.meps.ahrq.gov/mepsweb/. Accessed Jan 6, 2008.
- 17Agency for Healthcare Research and Quality. Medical Expenditures Panel Survey: Frequently Asked General Questions. August 2005. Available at: http://www.meps.ahrq.gov/faqs/faq_hc.htm. Accessed Mar 13, 2006.
- 18Agency for Healthcare Research and Quality. Medical Expenditures Panel Survey HC-079: 2003 Full Year Consolidated Data File. Washington, DC: Agency for Healthcare Research and Quality, 2005.
- 30American Academy of Pediatrics Committee on Pediatric Emergency Medicine. Overcrowding Pediatric Emergency Medicine: Overcrowding crisis in our nation’s emergency departments: is our safety net unraveling? Policy Statement. Pediatrics. 2004; 114:878–88.
|ACEM_075_sm_TablesA1 A2.pdf||31K||Supporting info item|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.