This cross-sectional study evaluates whether out-of-network billing risk differs by the site of ambulance origination in the US.
Key Points
Question
Is site of ambulance origination (home, incident scene, medical facility) associated with variation in designation of services as out of network?
Findings
In this cross-sectional study of over 2 million ground ambulance services, the prevalence of out-of-network billing was 54.8%, and the probability of out-of-network billing was the highest among ambulances originating from home or the incident scene. Patients’ total financial burden, including cost-sharing and potential balance bills, was $434.70 per service.
Meaning
These findings suggest legislation is needed to protect patients from balance billing following ambulance use, especially when ambulances originate from home, the scene of an incident, or nonhospital medical facilities.
Abstract
Importance
The No Surprises Act implemented in 2022 aims to protect patients from surprise out-of-network (OON) bills, but it does not include ground ambulance services. Understanding ground ambulance OON and balance billing patterns from previous years could guide legislation aimed to protect patients following ground ambulance use.
Objective
To characterize OON billing from ground ambulance services by evaluating whether OON billing risk differs by the site of ambulance origination (home, hospital, nonhospital medical facility, or scene of incident).
Design, Setting, and Participants
Cross-sectional study of the Merative MarketScan dataset between January 1, 2015, and December 31, 2020, using claims-based data from employer-based private health insurance plans in the US. Participants included patients who utilized ground ambulances during the study period. Data were analyzed from June to December 2023.
Exposure
Medical encounter requiring ground ambulance transportation.
Main Outcomes and Measures
Ground ambulance OON billing prevalence was calcuated. Linear probability models adjusted for state-level mixed effects were fit to evaluate OON billing probability across ambulance origins. Secondary outcomes included the allowed payment, patient cost-sharing amounts, and potential balance bills for OON ambulances.
Results
Among 2 031 937 ground ambulance services (1 375 977 unique patients) meeting the inclusion and exclusion criteria, 1 072 791 (52.8%) rides transported men, and the mean (SD) patient age was 41 (18) years. Of all services, 1 113 676 (54.8%) were billed OON. OON billing probabilities for ambulances originating from home or scene were higher by 12.0 percentage points (PP) (95% CI, 11.8-12.2 PP; P < .001 for home; 95% CI, 11.7-12.2 PP; P < .001 for scene) vs those originating from hospitals. Mean (SD) total financial burden, including cost-sharing and potential balance bills per ambulance service, was $434.70 ($415.99) per service billed OON vs $132.21 ($244.92) per service billed in-network.
Conclusions and Relevance
In this cross-sectional study of over 2 million ground ambulance services, ambulances originating from home, the scene of an incident, and nonhospital medical facilities were more likely to result in OON bills. Legislation is needed to protect patients from surprise billing following use of ground ambulances, more than half of which resulted in OON billing. Future legislation should at minimum offer protections for these subsets of patients often calling for an ambulance in urgent or emergent situations.
Introduction
In situations requiring ambulances, patients lack choices for preferred ambulance organization, and ambulances transport patients regardless of ability to pay.1 Thus, these patients are vulnerable to out-of-network (OON) billing.1 The No Surprises Act was introduced to protect patients from surprise bills,2,3 which arise when patients discover services were OON after billing.4 When surprise billing occurs, patients receive balance bills, the difference between OON and in-network bills.3,5,6,7 While the act protects patients from balance billing for emergency care, nonemergency care, and air ambulance, it does not include ground ambulance services.2,8
In the US, more than 3 million emergency department (ED) visits involved ground ambulances for privately insured patients in 2018.7 Additionally, ground ambulances handle interhospital transfers and posthospital stay transportation.9,10 Medicare and Medicaid cover medically necessary ambulance services,11,12 as do private insurance plans to varying degrees (including Anthem, Cigna, Kaiser Permanente, and UnitedHealthcare, among others).13 Per large insurance claims data, nearly 80% of ground ambulance rides resulted in OON billing.1 States and localities regulate ambulance billing at either the ambulance organization or insurance policy level. For instance, Maryland prohibits publicly owned ambulances from balance billing,14 and Colorado limits ground ambulance reimbursements from state-regulated health plans.15 Some states specifically prohibit surprise billing following ground ambulance use (eg, Ohio) or subsidize OON bills.16 When an OON service occurs, insurers could either cover the entire OON bill, pay a standard rate (eg, in-network rate), or refuse to cover the bill.17 Prior studies reported 28% to 71% of emergency ground ambulance rides may result in surprise billing.1,7,18
While prior studies investigated the association of emergency status and ambulance ownership status to OON billing,1,6,18 other important factors associated with risk remain unknown. The context for ground ambulance use varies depending on where the ambulance originates. For example, an ambulance organized by case managers to transport a patient at discharge from a hospital to a rehabilitation facility is different from one transporting a patient from the scene of an incident to the ED. In this study, we examined the association between ground ambulance point of origin and OON billing and the resultant financial burden.
Methods
Data Source and Cohort Selection
The present study was deemed exempt from review by the MedStar Health Research Institute institutional review board and exempt from the need for informed consent because data were deidentified. The study followed applicable recommendations in the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.
We analyzed the 2015 to 2020 Merative MarketScan Commercial Claims and Encounters Database, which covers employer-sponsored plan beneficiaries. Using ground ambulance Healthcare Common Procedure Coding System (HCPCS) codes (eTable 1 in Supplement 1),18 we selected all ground ambulance services billed in the outpatient services tables. We excluded services with missing outcomes or key variables (indicator for in-network service, indicator for in-network billing, ambulance service modifier, and metropolitan statistical area [MSA]). The US Office of Management and Budget classifies geographical regions into 2 types of core-based statistical areas (CBSAs) according to populations: MSA and micropolitan statistical area.19 MSAs are geographic units with 1 or more urbanized areas with a minimum population of 50 000.20
Ambulance service billing requires a modifier detailing the origin and destination.21 For example, the SH modifier signifies that the ambulance originated from the scene (S) and arrived at the hospital (H). These modifiers were used to categorize ambulance origins into 4 groups: hospital, home, scene, or nonhospital medical facility (eTable 2 in Supplement 1). We excluded HCPCS modifiers that begin with I (site of transfer between modes of ambulance, such as helicopter pad for transitioning ground ambulance to air ambulance) or do not specify origin (eg, the GM modifier, signifying more than 1 patient being transferred).21
As MarketScan does not report charges directly billed to patients by organizations, we collected charges reported in the Medicare organization utilization and payment data to calculate potential balance bills. This data source is a reasonable alternative as charges are universal regardless of payer types.22,23 A previous study on OON billing using MarketScan calculated mean charges at the state, year, and service level (ie, Current Procedural Terminology code) to calculate potential balance bills.22 As ground ambulance services are more likely regulated by local jurisdictions than the state,2,7 we used MSA (smallest scale regional variable in MarketScan) instead of state. We merged the zip code–CBSA crosswalk file by the US Department of Housing and Urban Development with the Medicare charge dataset.24 We aggregated charge data at the zip code–year-HCPCS level into the MSA-year-HCPCS level by calculating mean ground ambulance charges weighted by the number of services billed across different zip codes. We merged the MSA-year-HCPCS charge data with the HCPCS-level MarketScan data previously noted.
Variables
We collected patient demographic information (age, sex, full-time employment status, and insurance type), regional variables (geographic region, employer state, and MSA population), financial variables (allowed payment, deductible, coinsurance, and copayment), and service year. The urbanicity of patient residence determined by MSA is associated with presenting to the ED via ambulance.25 We used the 2020 MSA population estimates dataset published by the Bureau of the Census to estimate MSA populations.26 Per the Geographic Areas Reference Manual, regions are categorized into 5 population levels (≥1 000 000, 250 000-999 999, 100 000-249 999, 50 000-99 999, and <50 000).27 Allowed payment is the amount that organizations are eligible for after payers apply pricing guidelines (eg, fee schedules or discounts) to the submitted charge.28 The allowed payment includes cost-sharing by patients (eg, deductible, coinsurance, or copayment), third-party payment (eg, coordination of benefits or other savings), and payment received by the organization excluding patient cost-sharing and third-party payment.28 All monetary values were inflation-adjusted to 2020 US dollars using the Consumer Price Index.29
Our primary outcome variable was OON billing (ie, involving OON payment or OON ambulance organization) following a ground ambulance service. Secondary outcome variables included patient cost-sharing payments and allowed payments by services. We calculated patient cost-sharing as the sum of copayment, coinsurance, and deductible.
For OON ground ambulance services with complete MSA-year-HCPCS charge data (some zip codes could not be matched with MSA using the zip code–CBSA crosswalk file), we calculated a potential balance bill using a method similar to that of Sen et al.22 Potential balance bill was defined as the difference between the mean Medicare charge at the MSA-year-HCPCS level and the MarketScan allowed payment.22 Potential balance bill amounts less than 0 were converted to 0.22 Total financial burden was calculated by adding the potential balance bill and patient cost-sharing payment.22
Our primary explanatory variable was the setting from which the ambulance originated (hospital, home, incident scene, or nonhospital medical facility). Our secondary explanatory variable was the emergency status of ambulances given that the pressing nature of emergencies can limit service choices.18 The HCPCS codes A0427, A0429, and A0433 were labeled emergency, while A0426 and A0428 were labeled nonemergency (eTable 1 in Supplement 1).
Statistical Analysis
We stratified our cohort into OON and in-network services. We calculated descriptive statistics for patient demographics, regional variables, financial variables, and service year. We used 2-sided t-tests and χ2 tests to compare continuous and categorical variables between the 2 cohorts, respectively.
OON billing was our primary binary outcome. OON billing prevalence was calculated across ambulance origins and emergency status. To estimate the effect of ambulance origin and emergency status on OON billing, we fit a linear probability model adjusted for sex, age, employment status, insurance type, region, MSA population levels, advanced life support service, year, and state-level variations using mixed-effects estimation. We calculated OON billing probability measured in percentage points (PP) across ambulance origins. We also fit another linear probability model with the same covariates and interaction terms between emergency status and ambulance origin.
Many ambulance services are governed by regional policies, including some that protect patients from ground ambulance surprise billing.7,30 We created a heat map to assess the regional variability of OON billing prevalence.
To examine the association between ground ambulance origin and patient cost-sharing burden, we compared the mean allowed payment and patient cost-sharing between OON services vs in-network services. We also performed a subanalysis to compare the magnitude of cost-sharing in emergency vs nonemergency services. Finally, using data with complete MSA-year-HCPCS charge data, we calculated the mean patient cost-sharing, potential balance bills, and total financial burden for OON ground ambulance services. We used an α of .05 for statistical significance. We used R software version 4.2.0 (R Project for Statistical Computing) for statistical analyses. Data were analyzed from June to December 2023.
Results
Among 2 031 937 ground ambulance services (1 375 977 unique patients) meeting the inclusion or exclusion criteria (eFigure in Supplement 1), 1 072 791 (52.8%) rides transported men, and the mean (SD) patient age was 41 (18) years. Distributions of variables are reported for excluded ground ambulance services and the primary analysis cohort in eTable 3 in Supplement 1. Of all services, 1 113 676 (54.8%) were billed OON (Table 1). Emergency status and ambulance originating from the scene, home, or nonhospital medical facility were associated with higher OON billing (emergency status: χ21 = 32 457; P < .001; ambulance origin: χ23 = 39 206; P < .001). OON billing prevalence was the highest for exclusive provider organization (EPO) plans and lowest for consumer-driven and high-deductible health plans. Both patient cost-sharing (mean [SD], $157.38 [$264.49] vs $129.93 [$239.26]; difference, $27.45; 95% CI, $26.75-$28.14; P < .001) and allowed payments (mean [SD], $792.49 [$530.65] vs $669.61 [$480.68]; difference, $122.88; 95% CI, $121.48 to $124.27; P < .001) were higher among OON services. Most nonemergency ambulances originated from the hospital, while most emergency ambulances originated from home and scene (eTable 4 in Supplement 1). The Figure shows variations in OON billing prevalence across states (numerical values and descriptive statistics of prevalence are available in eTable 5 in Supplement 1).
Table 1. Characteristics of Ground Ambulance Services.
| Variable | No. (%) | P valuea | ||
|---|---|---|---|---|
| Total (N = 2 031 937) | In-network (n = 918 261) | Out-of-network (n = 1 113 676) | ||
| Sex | ||||
| Female | 959 146 (47.2) | 432 690 (45.1) | 526 456 (54.9) | .03 |
| Male | 1 072 791 (52.8) | 485 571 (45.3) | 587 220 (54.7) | |
| Age, mean (SD), y | 41 (18) | 41 (18) | 40 (18) | <.001 |
| Full time employment | 1 367 024 (67.3) | 614 713 (45.0) | 752 311 (55.0) | <.001 |
| Year | ||||
| 2015 | 366 039 (18.0) | 160 051 (43.7) | 205 988 (56.3) | <.001 |
| 2016 | 370 292 (18.2) | 171 190 (46.2) | 199 102 (53.8) | |
| 2017 | 329 710 (16.2) | 159 404 (48.3) | 170 306 (51.7) | |
| 2018 | 377 777 (18.6) | 161 121 (42.6) | 216 656 (57.4) | |
| 2019 | 325 229 (16.0) | 145 461 (44.7) | 179 768 (55.3) | |
| 2020 | 262 890 (12.9) | 121 034 (46.0) | 141 856 (54.0) | |
| Region | ||||
| North East | 354 603 (17.5) | 158 241 (44.6) | 196 362 (55.4) | <.001 |
| North Central | 382 627 (18.8) | 181 809 (47.5) | 200 818 (52.5) | |
| South | 1 008 040 (49.6) | 471 565 (46.8) | 536 475 (53.2) | |
| West | 283 776 (14.0) | 105 242 (37.1) | 178 534 (62.9) | |
| Unknown | 2891 (0.1) | 1404 (48.6) | 1487 (51.4) | |
| MSA population | ||||
| ≥1 000 000 | 1 208 608 (59.5) | 512 274 (42.4) | 696 334 (57.6) | <.001 |
| 250 000-999 999 | 414 244 (20.4) | 189 035 (45.6) | 225 209 (54.4) | |
| 100 000-249 999 | 129 792 (6.4) | 65 925 (50.8) | 63 867 (49.2) | |
| 50 000-99 999 | 7666 (0.4) | 3754 (49.0) | 3912 (51.0) | |
| <50 000 | 271 627 (13.4) | 147 273 (54.2) | 124 354 (45.8) | |
| Insurance type | ||||
| PPO | 1 015 792 (50.0) | 443 255 (43.6) | 572 537 (56.4) | <.001 |
| HMO | 261 101 (12.8) | 113 642 (43.5) | 147 459 (56.5) | |
| EPO | 20 748 (1.0) | 4488 (21.6) | 16 260 (78.4) | |
| POS | 216 639 (10.7) | 76 488 (35.3) | 140 151 (64.7) | |
| CDHP/HDHP | 433 457 (21.3) | 246 443 (56.9) | 187 014 (43.1) | |
| Other/unknown | 84 200 (4.1) | 33 945 (40.3) | 50 255 (59.7) | |
| Emergency | 1 673 130 (82.3) | 707 375 (42.3) | 965 755 (57.7) | <.001 |
| ALS service | 1 178 273 (58.0) | 521 471 (44.3) | 656 802 (55.7) | <.001 |
| Origin of ambulance | ||||
| Hospital | 436 047 (21.5) | 253 014 (58.0) | 183 033 (42.0) | <.001 |
| Home | 726 132 (35.7) | 299 193 (41.2) | 426 939 (58.8) | |
| Scene | 698 454 (34.4) | 285 365 (40.9) | 413 089 (59.1) | |
| Nonhospital medical facility | 171 304 (8.4) | 80 689 (47.1) | 90 615 (52.9) | |
| Financial variables, mean (SD), US $b | ||||
| Allowed amount | 736.96 (512.34) | 669.61 (480.68) | 792.49 (530.65) | <.001 |
| Patient cost-sharing | 144.97 (253.77) | 129.93 (239.26) | 157.38 (264.49) | <.001 |
| Deductible | 90.91 (246.19) | 82.43 (232.52) | 97.91 (256.70) | <.001 |
| Coinsurance | 48.92 (91.95) | 42.92 (81.08) | 53.86 (99.76) | <.001 |
| Copayment | 5.14 (29.20) | 4.58 (26.50) | 5.60 (31.24) | <.001 |
Abbreviations: ALS, advanced life support; CDHP, consumer-driven health plan; EPO, exclusive provider organization; HDHP, high deductible health plan; HMO, health maintenance organization; MSA, metropolitan statistical area; POS, point-of-service; PPO, preferred provider organization.
Descriptive statistics were calculated to compare distributions of variables between services with vs without out-of-network billing using 2-sided t-tests for continuous variables and χ2 tests for categorical variables.
All monetary values were adjusted to 2020 US dollars. Patient cost-sharing was calculated as the sum of coinsurance, deductible, and copayment.
Figure. Prevalence of Out-of-Network Billing Among Ground Ambulance Services for Privately Insured Patients.
A total of 79 205 services with unknown states and 51 services in Puerto Rico were not mapped in this figure.
When stratified by the ambulance origin and emergency status, emergency ground ambulances from home had the highest prevalence of OON billing (414 559 ambulances [59.3%]) while nonemergency rides from the hospital had the lowest (110 493 ambulances [40.2%]) (Table 2). In the model without interaction terms (full regression result in eTable 6 in Supplement 1), ambulances originating from home or scene had higher OON billing probability than from a hospital (12.0 PP; 95% CI, 11.8-12.2 PP; P < .001 for home; 12.0 PP; 95% CI, 11.7-12.2 PP; P < .001 for scene) (Table 3). With the model including interaction terms, emergency status increased the OON billing probability the most when the ambulance service originated from home (12.9 PP; 95% CI, 12.7-13.2 PP; P < .001) (Table 3, full regression result in eTable 7 in Supplement 1).
Table 2. Prevalence of Out-of-Network Billing in Emergency vs Nonemergency Services Across Different Ground Ambulance Origins.
| Ambulance origin | Out-of-network billing prevalence, No. (%) | ||
|---|---|---|---|
| All services (N = 1 113 676) | Nonemergency services (n = 147 921) | Emergency services (n = 965 755) | |
| Hospital | 183 033 (42.0) | 110 493 (40.2) | 72 540 (45.1) |
| Home | 426 939 (58.8) | 12 380 (46.4) | 414 559 (59.3) |
| Scene | 413 089 (59.1) | 1226 (49.4) | 411 863 (59.2) |
| Nonhospital medical facility | 90 615 (52.9) | 23 822 (43.7) | 66 793 (57.2) |
Table 3. Probability of Out-of-Network Billing in Emergency vs Nonemergency Services Across Different Ground Ambulance Originsa.
| Ambulance origin | Out-of-network billing probability (95% CI), percentage points | |||||
|---|---|---|---|---|---|---|
| All servicesb | P value | Nonemergency servicesc | P value | Emergency servicesd | P value | |
| Hospital | Reference | NA | Reference | NA | Reference | NA |
| Home | 12.0 (11.8-12.2) | <.001 | 9.1 (8.5-9.7) | <.001 | 12.9 (12.7-13.2) | <.001 |
| Scene | 12.0 (11.7-12.2) | <.001 | 10.3 (8.5-12.2) | <.001 | 12.8 (12.5-13.0) | <.001 |
| Nonhospital medical facility | 7.3 (7.0-7.6) | <.001 | 5.5 (5.1-6.0) | <.001 | 8.7 (8.3-9.1) | <.001 |
Abbreviation: NA, not applicable.
All linear probability models were adjusted for emergency status, advanced life support service, patient sex, patient age, metropolitan statistical area populations, employment status, insurance type, year, and state-level mixed-effects.
This model compares adjusted probability of out-of-network billing among all services across ambulance origins with hospital as the reference. Full model results are reported in eTable 6 in Supplement 1.
This model compares adjusted probability of out-of-network billing among nonemergency services across ambulance origins with hospital as the reference. Full model results are reported in eTable 7 in Supplement 1.
This model compares adjusted probability of out-of-network billing among emergency services across ambulance origins with hospital as the reference. Full model results are reported in eTable 7 in Supplement 1.
Across all ambulance origins, mean patient cost-sharing payments were larger in ambulances billed OON vs in-network (eTable 8 in Supplement 1). Services with complete MSA-year-HCPCS charge data (1 748 806 services [86.0%]) were used to calculate potential balance bills (eFigure in Supplement 1). OON billing prevalence was similar between services with complete charge data and all services (56.3% vs 54.8%). The mean (SD) potential balance bill was $274.70 ($349.32) (Table 4). Mean (SD) potential balance bills were the highest ($342.71 [$252.94]) for OON ground ambulance services originating from home in a nonemergent manner. The mean (SD) total financial burden was estimated to be $434.70 ($415.99) for OON services vs $132.21 ($244.92) for in-network mean cost-sharing amount. Mean total financial burden was higher in emergency vs nonemergency situations across all ambulance origins.
Table 4. Potential Balance Bills and Patient Cost Sharing Amounts Among Ambulance Services With Complete Charge Dataa.
| Service type and ambulance origin | Network status, mean (SD), US $ | |||
|---|---|---|---|---|
| In-network, observed patient cost-sharing | Out-of-network | |||
| Observed patient cost-sharing | Potential balance bill | Total potential financial burden | ||
| Pooled | ||||
| Total | 132.21 (244.92) | 160.00 (268.43) | 274.70 (349.32) | 434.70 (415.99) |
| Hospital | 78.04 (174.19) | 117.70 (231.33) | 263.93 (342.82) | 381.63 (389.08) |
| Home | 134.16 (247.23) | 149.24 (259.93) | 280.36 (349.30) | 429.60 (414.26) |
| Scene | 186.78 (287.15) | 196.20 (291.74) | 274.36 (356.79) | 470.56 (432.54) |
| Nonhospital medical facility | 91.33 (206.44) | 121.03 (235.58) | 269.00 (324.36) | 390.02 (375.41) |
| Nonemergency | ||||
| Total | 52.92 (133.02) | 87.09 (201.73) | 269.02 (327.47) | 356.11 (361.77) |
| Hospital | 61.24 (139.80) | 102.61 (216.81) | 256.99 (342.23) | 359.60 (382.08) |
| Home | 18.61 (90.55) | 25.23 (97.86) | 342.71 (252.94) | 367.94 (258.01) |
| Scene | 137.11 (245.64) | 143.30 (237.81) | 282.25 (366.44) | 425.55 (405.87) |
| Nonhospital medical facility | 19.46 (88.69) | 39.16 (137.89) | 290.47 (274.35) | 329.64 (293.86) |
| Emergency | ||||
| Total | 155.79 (264.84) | 170.44 (275.09) | 275.51 (352.33) | 445.96 (421.99) |
| Hospital | 114.80 (228.04) | 143.28 (252.02) | 275.72 (343.48) | 419.00 (397.91) |
| Home | 139.47 (250.83) | 152.37 (261.97) | 278.78 (351.25) | 431.15 (417.34) |
| Scene | 186.98 (287.29) | 196.34 (291.86) | 274.34 (356.76) | 470.68 (432.61) |
| Nonhospital medical facility | 133.06 (240.94) | 147.61 (253.89) | 262.02 (338.73) | 409.64 (396.35) |
Potential balance bill was calculated as the difference between the estimated mean charge data at the metropolitan statistical area-year-Healthcare Common Procedure Coding System level reported in the Medicare Provider Utilization and Payment Data and the allowed amount observed in the MarketScan data. Total potential financial burden was calculated as the sum of the potential balance bill and the observed patient cost-sharing amount. All monetary values were inflation-adjusted to 2020 US dollars.
Discussion
Our study of over 2 million ground ambulance services from a major payer database revealed the following. First, over half of ground ambulances resulted in OON billing. Second, ambulances originating from a hospital had a lower OON billing probability compared with those from home, scene, and a nonhospital medical facility. This association was consistently present after adjusting for emergency status and its interaction with ambulance origins. Thus, we believe the association between ambulance origination and OON billing is an important factor independent of emergency status. Third, although ambulances originating from hospitals had a lower OON billing likelihood, patients transported from a hospital may be exposed to more financial burden through higher cost-sharing payments. Finally, patients could face an extra $275 per service when balance billed.
While insurers may pay the entire OON bill, not leading to surprise bills,1,18 the risk of surprise billing remains for OON ground ambulances. Using national insurance claims data, Adler et al1 reported that among OON ground ambulance services, one-third of emergency transports (33%) and almost half of nonemergency transports (46%) resulted in potential surprise billing. A national poll conducted by Morning Consult in January 2022 reported that 1 in 6 surprise bills were due to OON ambulance rides.31 Half of OON ambulance rides were associated with ED encounters with surprise billing.6 As more than 3 million ground ambulances are utilized by privately insured patients each year,7 we estimate the extra patient cost-sharing for OON vs in-network services to be nearly $84 million per year for just privately insured patients. If balance billing occurs, this financial burden will be substantially higher.
Policies for many ground ambulances are governed locally or regionally.2,7 More than 60% of emergency ground ambulances are provided by local government agencies and fire departments.7 Some ambulance corps even financially rely on balance billing.2 Thus, policymakers may have excluded ground ambulances out of concern for interfering with local policies.2
Availability of resources to navigate network status at the ambulance origin may influence OON billing and patient cost-sharing. For example, the lower probability and patient cost-sharing amounts of OON billing for ambulance services originating from a hospital may be because of hospital-owned services32 that account for 8% of all emergency ground ambulances.7 If a hospital is in-network, their ambulance services are more likely in-network. In addition, in the hospital setting, case managers may contact different transportation organizations to identify the cheapest option.33 Therefore, for scheduled patient transfers from a hospital, finding an affordable ambulance company in contract with the patient’s insurance plan may be feasible.34 Ambulances originating from home and scene had the highest OON billing probability and cost-sharing amounts. This may be because ambulances from the scene or home are more likely to be owned by for-profit companies that remain OON and primarily respond to 911 dispatches.34 Insurance type is also associated with the prevalence of OON billing; for example, EPO plans were associated with the highest likelihood of OON billing, likely because they encourage patients to use in-network organizations by foregoing OON service coverage.28 Thus, understanding ground ambulance utilization patterns from different origins may elucidate varying risks of OON billing and financial burden.
Due to concerns for surprise billing following ground ambulance use, in November 2021, the Ground Ambulance and Patient Billing (GAPB) Advisory Committee was established to make recommendations to protect consumers.35 The GAPB Advisory Committee has a mandate to make the ground ambulance billing process more transparent to help consumers make informed decisions.35,36 Without federal regulations, protecting patients from these situations depends on state and local legislators. As of November 2021, only 10 states provide some protection against ground ambulance surprise billing.37 For example, Maryland law only prohibits ambulance surprise billing for services by volunteer fire department, volunteer rescue squad, or state-authorized organizations, leaving patients exposed to surprise billing risk from many other ambulance services.30 New York protects enrollees in health maintenance organizations, preferred provider organizations, and EPOs from surprise billing in ground ambulances, except for interfacility transfers.30 Neither of these state laws protect enrollees in self-funded plans (ie, employers provide health benefits directly to employees) that account for two-thirds of workers in the US.30,38
High OON billing prevalence regardless of ambulance origin and emergency status necessitates additional protections for patients. Expanding the No Surprises Act to regulate ground ambulance billing at the federal level could be a comprehensive solution. While the act aims to protect patients from costs beyond in-network amounts, ground ambulances—especially during emergencies—are mostly OON,1,18 limiting the utility of using in-network price as the benchmark. Capping ground ambulance charges to a multiple of standardized Medicare-allowed payments is an alternative billing regulation approach.34 However, since some ambulance companies’ financial solvency depends on collecting balance bills, this should be done thoughtfully to avoid disrupting the locally-regulated ground ambulance market. Finally, while mandating the transparency of ambulance services’ network status may better allow patients to pick in-network options, this would not help patients using ambulance services in emergent situations. Further discussion among all stakeholders is needed to identify a feasible and comprehensive policy solution.
Limitations
This study had several limitations. First, because MarketScan does not report charges or whether patients were aware of the network status of services, we were unable to determine surprise and balance billing with absolute certainty. This led us to choose OON billing as our primary outcome. Second, we excluded services with key missing variables. While the emergency status of ground ambulances, a known risk factor for OON billing,1 did not differ between the excluded services and the analysis cohort, some variables differed, suggesting that missing data may have biased our findings. Third, because MarketScan does not report the distance traveled during an ambulance encounter, we did not include mileage HCPCS codes since they are billed per mile; this suggests that the financial burden that we report is an underestimate, as it includes only costs per service rather than per encounter. Fourth, because MarketScan does not report charge data, we could not calculate exact balance bills.5,28 As an alternative, we used Medicare charge data as described in a previous study22 and created MSA-year-HCPCS charge data to capture regional and local variabilities in charges. Although we could not link MarketScan data to specific ambulance organizations, which introduces potential inaccuracies in charge estimates at the organization level, our approach reflects the experience of a typical patient within an MSA region. Fifth, since MarketScan only includes privately insured patients,5,28 Medicare and Medicaid enrollees were not included in our study. These populations, however, are better protected from balance billing by federal regulations.39,40 Also, nearly half of ambulance services in our study were billed in Southern states, which may be overrepresented and skew our nationwide analyses. In contrast, Puerto Rico only billed 51 services, suggesting that our state-level analysis may be more robust for some states than others. Additionally, our findings depend on the accuracy of data entries. Fortunately, MarketScan has comprehensive, high-quality coding with complete payment information and 99% appropriate International Classification of Diseases, Ninth Revision or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes.41
Conclusions
In this cohort of privately insured patients using ground ambulance services, we characterized services that had higher OON billing probability and cost-sharing payments. Ambulance services that originate from the scene, home, and nonhospital medical facilities are more likely to result in OON billing and, if balance billing occurs, high total financial burden. These factors can be important for future legislation at the federal or state levels to better protect patients from receiving surprise bills following use of ground ambulance services.
eTable 1. Ambulance Healthcare Common Procedure Coding System Codes Used for Cohort Selection
eTable 2. HCPCS Code Modifiers Used to Determine Places of Ground Ambulance Origins
eFigure. Inclusion/Exclusion Criteria Flowchart of the Study Cohort Selection
eTable 3. Distribution of Variables for Excluded Ground Ambulance Services and Primary Analysis Cohort
eTable 4. Distribution of Ground Ambulance Services Across Emergency Status and Points of Origin
eTable 5. Prevalence of Out-of-Network Billing Among Ground Ambulance Services Across Different States
eTable 6. Results for the Full Regression Analysis Without Interaction Terms on Probability of Out-of-Network Billing in Ground Ambulance Services
eTable 7. Results for the Full Regression Analysis With Interaction Terms on Probability of Out-of-Network Billing in Ground Ambulance Services
eTable 8. Differences in Mean Total Allowed Amounts and Mean Patient Cost-Sharing Amounts Between Services With vs. Without Out-of-Network Billing Across Ground Ambulance Origins
Data Sharing Statement
References
- 1.Adler L, Ly B, Duffy E, Hannick K, Hall M, Trish E. Ground ambulance billing and prices differ by ownership structure. Health Aff (Millwood). 2023;42(2):227-236. doi: 10.1377/hlthaff.2022.00738 [DOI] [PubMed] [Google Scholar]
- 2.Chhabra KR, Fuse Brown E, Ryan AM. No more surprises - new legislation on out-of-network billing. N Engl J Med. 2021;384(15):1381-1383. doi: 10.1056/NEJMp2035905 [DOI] [PubMed] [Google Scholar]
- 3.Pallone F Jr. H.R.3630 - No Surprises Act. 116th Congress (2019-2020) . Accessed March 5, 2022. https://www.congress.gov/bill/116th-congress/house-bill/3630
- 4.Consumer Financial Protection Bureau. What is a “surprise medical bill” and what should I know about the No Surprises Act? Accessed February 25, 2023. https://www.consumerfinance.gov/ask-cfpb/what-is-a-surprise-medical-bill-and-what-should-i-know-about-the-no-surprises-act-en-2123/
- 5.Long C, Zhang G, Sanghavi KK, Qiu C, Means KR, Jr., Giladi AM. Surprise out-of-network bills for hand and upper extremity trauma patients. J Hand Surg Am. 2022;47(12):1230.E1-1230.E17. doi: 10.1016/j.jhsa.2021.09.015 [DOI] [PubMed] [Google Scholar]
- 6.Garmon C, Chartock B. One in five inpatient emergency department cases may lead to surprise bills. Health Aff (Millwood). 2017;36(1):177-181. doi: 10.1377/hlthaff.2016.0970 [DOI] [PubMed] [Google Scholar]
- 7.Peterson-KFF Health System Tracker . Ground ambulance rides and potential for surprise billing. Accessed March 5, 2022. https://www.healthsystemtracker.org/brief/ground-ambulance-rides-and-potential-for-surprise-billing/
- 8.Richman B, Hall M, Schulman K. The no surprises act and informed financial consent. N Engl J Med. 2021;385(15):1348-1351. doi: 10.1056/NEJMp2109971 [DOI] [PubMed] [Google Scholar]
- 9.Maryland Institute for Emergency Medical Services Systems . Interhospital transfer resource manual. Accessed July 29, 2022. https://miemss.org/home/Portals/0/Docs/Guidelines_Protocols/Interhospital-Transfer-Manual-20211216.pdf?ver=2021-12-16-131351-387
- 10.Munjal KG, Shastry S, Chapin H, et al. Retrospective cohort study of rates of return emergency department visits among patients transported home by ambulance. J Emerg Med. 2020;59(1):147-152. doi: 10.1016/j.jemermed.2020.04.043 [DOI] [PubMed] [Google Scholar]
- 11.Centers for Medicare and Medicaid Services . Let Medicaid give you a ride. Accessed December 16, 2023. https://www.cms.gov/medicare-medicaid-coordination/fraud-prevention/medicaid-integrity-education/downloads/nemt-factsheet.pdf
- 12.Medicare.gov . Ambulance services. Accessed December 16, 2023. https://www.medicare.gov/coverage/ambulance-services
- 13.Rivelli E. Does insurance cover ambulance rides? Forbes Advisor . Accessed December 16, 2023. https://www.forbes.com/advisor/health-insurance/insurance-for-ambulance-rides/
- 14.Maryland Attorney General . The no surprises act. Accessed December 16, 2023. https://www.marylandattorneygeneral.gov/Pages/CPD/HEAU/NSA.aspx
- 15.Colorado Division of Insurance . Federal no surprises act/Colorado out-of-network billing. Accessed December 16, 2023. https://doi.colorado.gov/insurance-products/health-insurance/health-insurance-initiatives/federal-no-surprises-act/colorado
- 16.The Ohio Department of Insurance . Surprise billing FAQs. Accessed December 16, 2023. https://insurance.ohio.gov/consumers/surprise-billing/resources/01-surprise-billing-faqs
- 17.Cooper Z, Nguyen H, Shekita N, Morton FS. Out-of-network billing and negotiated payments for hospital-based physicians. Health Aff (Millwood). 2020;39(1):24-32. doi: 10.1377/hlthaff.2019.00507 [DOI] [PubMed] [Google Scholar]
- 18.Chhabra KR, McGuire K, Sheetz KH, Scott JW, Nuliyalu U, Ryan AM. Most patients undergoing ground and air ambulance transportation receive sizable out-of-network bills. Health Aff (Millwood). 2020;39(5):777-782. doi: 10.1377/hlthaff.2019.01484 [DOI] [PubMed] [Google Scholar]
- 19.Baker ES. Analyzing OMB classification of regions: three case studies. U.S. Bureau of Labor Statistics . Accessed March 5, 2022. https://www.bls.gov/opub/mlr/2016/beyond-bls/analyzing-omb-classification-of-regions-three-case-studies.htm
- 20.United States Census Bureau . Metropolitan and Micropolitan. Accessed March 6, 2022. https://www.census.gov/programs-surveys/metro-micro/about.html
- 21.Centers for Medicare and Medicaid Services . Origin and destination codes specific to ambulance service claims and emergency triage, treat, and transport (ET3) model claims. Accessed July 26, 2022. https://www.cms.gov/files/document/origin-and-destination-codes-specific-ambulance-service-claims-and-emergency-triage-treat-and.pdf
- 22.Sen AP, Meiselbach MK, Wang Y, Eisenberg MD, Anderson GF. Frequency and costs of out-of-network bills for outpatient laboratory services among privately insured patients. JAMA Intern Med. 2021;181(6):834-841. doi: 10.1001/jamainternmed.2021.1422 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bai G, Anderson GF. Extreme markup: the fifty US hospitals with the highest charge-to-cost ratios. Health Aff (Millwood). 2015;34(6):922-928. doi: 10.1377/hlthaff.2014.1414 [DOI] [PubMed] [Google Scholar]
- 24.Office of Policy Development and Researc h. Accessed March 25, 2023. https://www.huduser.gov/portal/datasets/usps_crosswalk.html
- 25.Meisel ZF, Pines JM, Polsky D, Metlay JP, Neuman MD, Branas CC. Variations in ambulance use in the United States: the role of health insurance. Acad Emerg Med. 2011;18(10):1036-1044. doi: 10.1111/j.1553-2712.2011.01163.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.United States Census Bureau . Metropolitan and Micropolitan Statistical Areas Population Totals and Components of Change, 2020-2021. Accessed July 29, 2022. https://www.census.gov/data/datasets/time-series/demo/popest/2020s-total-metro-and-micro-statistical-areas.html
- 27.U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census . The Geographic areas reference manual. Accessed April 15, 2022. https://www2.census.gov/geo/pdfs/reference/GARM/GARMcont.pdf
- 28.Truven Health Analytics . Commercial claims and encounters Medicare supplemental. Accessed March 22, 2022. https://theclearcenter.org/wp-content/uploads/2020/01/IBM-MarketScan-User-Guide.pdf
- 29.U.S. Bureau of Labor Statistics Mid-Atlantic Information Office . Consumer price index historical tables for U.S. city average. Accessed February 13, 2023. https://www.bls.gov/regions/mid-atlantic/data/consumerpriceindexhistorical_us_table.htm
- 30.The Commonwealth Fund . State balance-billing protections. Accessed April 15, 2022. https://www.commonwealthfund.org/node/27021
- 31.Galvin G. Surprise medical billing is now banned. The Public is still concerned about unexpected ER charges. morning consult. Accessed February 25, 2023. https://morningconsult.com/2022/01/13/surprise-billing-polling-data-feature/
- 32.Ragone M. Evolution or revolution: EMS industry faces difficult changes. JEMS. 2012;37(2):34-39. [DOI] [PubMed] [Google Scholar]
- 33.Miller K. A case manager’s perspective of patient transport. Case Manager. 2004;(Spec No):S2-S4. doi: 10.1016/j.casemgr.2003.08.011 [DOI] [PubMed] [Google Scholar]
- 34.Adler L, Fiedler M, Ginsburg PB, et al. State Approaches to mitigating surprise out-of-network billing. 2019. Accessed January 11, 2024. https://www.brookings.edu/articles/state-approaches-to-mitigating-surprise-out-of-network-billing/
- 35.U.S. Department of Health and Human Services . Biden-Harris administration announces steps to protect consumers and improve ground ambulance billing practices through no surprises act implementation. Accessed January 14, 2022. https://www.cms.gov/files/document/gapb-press-release-hhsgov-nov-19-2021.pdf-0
- 36.U.S. Centers for Medicare and Medicaid Services . Advisory committee on ground ambulance and patient billing (GAPB). Accessed July 29, 2022. https://www.cms.gov/regulations-guidance/advisory-committees/advisory-committee-ground-ambulance-and-patient-billing-gapb
- 37.The Commonwealth Fund . Protecting consumers from surprise ambulance bills. Accessed July 29, 2022. https://www.commonwealthfund.org/blog/2021/protecting-consumers-surprise-ambulance-bills
- 38.Kaiser Family Foundation . Employer Health Benefits Survey, Section 10: Plan Funding. 2020. Accessed June 14, 2022. https://www.kff.org/report-section/ehbs-2020-section-10-plan-funding/
- 39.U.S. Centers for Medicare and Medicaid Services . HHS announces rule to protect consumers from surprise medical bills. Accessed August 16, 2022. https://www.cms.gov/newsroom/press-releases/hhs-announces-rule-protect-consumers-surprise-medical-bills
- 40.Boccuti C. Paying a visit to the doctor: current financial protections for Medicare patients when receiving physician services. Kaiser Family Foundation. November 30, 2016. Accessed August 16, 2022. https://www.kff.org/medicare/issue-brief/paying-a-visit-to-the-doctor-current-financial-protections-for-medicare-patients-when-receiving-physician-services/
- 41.Adamson DM, Chang S, Hansen LG. Health research data for the real world: the MarketScan databases. 2006. Accessed January 11, 2024. http://patientprivacyrights.org/wp-content/uploads/2011/06/Thomson-Medstat-white-paper.pdf
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Ambulance Healthcare Common Procedure Coding System Codes Used for Cohort Selection
eTable 2. HCPCS Code Modifiers Used to Determine Places of Ground Ambulance Origins
eFigure. Inclusion/Exclusion Criteria Flowchart of the Study Cohort Selection
eTable 3. Distribution of Variables for Excluded Ground Ambulance Services and Primary Analysis Cohort
eTable 4. Distribution of Ground Ambulance Services Across Emergency Status and Points of Origin
eTable 5. Prevalence of Out-of-Network Billing Among Ground Ambulance Services Across Different States
eTable 6. Results for the Full Regression Analysis Without Interaction Terms on Probability of Out-of-Network Billing in Ground Ambulance Services
eTable 7. Results for the Full Regression Analysis With Interaction Terms on Probability of Out-of-Network Billing in Ground Ambulance Services
eTable 8. Differences in Mean Total Allowed Amounts and Mean Patient Cost-Sharing Amounts Between Services With vs. Without Out-of-Network Billing Across Ground Ambulance Origins
Data Sharing Statement