https://orcid.org/0000-0002-4327-4832
Insured patients who receive out-of-network care may receive a “balance bill” for the difference between the practitioner’s charge and their insurer’s contracted rate. In 2017, California banned balance billing for anesthesia care. This study examined the association between California’s law and subsequent payments for anesthesia care. The authors hypothesized that, after the law’s implementation, there would be no change in in-network payment amounts, and that out-of-network payment amounts and the portion of claims occurring out-of-network would decline.
The study used average, quarterly, California county-level payment data (2013 to 2020) derived from a claims database of commercially insured patients. Using a difference-in-differences approach, the change was estimated in payment amounts for intraoperative or intrapartum anesthesia care, along with the portion of claims occurring out-of-network, after the law’s implementation. The comparison group was office visit payments, expected to be unaffected by the law. The authors prespecified that they would refer to differences of 10% or greater as policy significant.
The sample consisted of 43,728 procedure code-county-quarter-network combinations aggregated from 4,599,936 claims. The law’s implementation was associated with a significant 13.6% decline in payments for out-of-network anesthesia care (95% CI, –16.5 to –10.6%; P < 0.001), translating to an average $108 decrease across all procedures (95% CI, –$149 to –$64). There was a statistically significant 3.0% increase in payments for in-network anesthesia care (95% CI, 0.9 to 5.1%; P = 0.007), translating to an average $87 increase (95% CI, $64 to $110), which may be notable in some circumstances but did not meet the study threshold for identifying a change as policy significant. There was a nonstatistically significant increase in the portion of claims occurring out-of-network (10.0%, 95% CI, –4.1 to 24.2%; P = 0.155).
California’s balance billing law was associated with significant declines in out-of-network anesthesia payments in the first 3 yr after implementation. There were mixed statistical and policy significant results for in-network payments and the proportion of out-of-network claims.
“Balance billing” for out-of-network care has gained significant attention by policymakers and medical specialty stakeholders, culminating in the federal “No Surprises Act” passed in December 2020
Previous analyses of state-specific legislation were limited by short follow-up periods or unique policy characteristics, leaving few generalizable data to help understand the potential long-term impact of “balance billing” legislation nationally
The authors estimated the association between the implementation of California’s “balance billing” legislation in 2017 with changes to in-network payment amounts, out-of-network payment amounts, and out-of-network proportion of claims across 2013 to 2020 for common anesthesia care
Across 4,599,936 claims, California’s balance billing law was associated with a 13.6% decline in out-of-network payments for anesthesia care, translating to an average $108 decrease across all procedures
In addition, there was a 3.0% increase in payments for in-network care that may have questionable policy significance
Finally, a 10.0% increase in the proportion of claims that were out-of-network did not meet the statistical significance threshold
Some patients who receive out-of-network care receive “balance bills” for the difference between what their healthcare practitioner charged and what their insurer paid. Although in some cases patients may willingly go out-of-network to receive care from a specific practitioner (i.e., a specific surgeon), concerns can arise when a patient unknowingly receives care from an out-of-network practitioner.1–4
The issue of “surprise balance billing” has received substantial attention from policymakers. In December 2020, the U.S. Congress passed the No Surprises Act to address the issue, and several states adopted similar legislation.5 These laws generally have two common components. First, they commonly require practitioners to hold patients harmless for any “balance-billed” amount. Second, they commonly require the insurer and practitioner to negotiate the payment amount. The laws can differ in how they address disputes in these negotiations. For example, New York’s law requires binding arbitration and establishes a benchmark payment based on the 80th percentile of billed charges, while California’s law requires insurance plans to pay a standard amount: the greater of the local average in-network contracted rate or 125% of the Medicare reimbursement rate.6
The varying ways in which these laws set payments for out-of-network care may also affect prices for in-network care, which has raised concerns for policymakers and stakeholders. For example, some have argued that New York’s law resulted in medical price inflation because the guidance given to arbitrators was overly generous to out-of-network practitioners, weakening insurers’ bargaining power.7,8 By contrast, some have criticized California’s law for partly tying payments for out-of-network care to a multiple of Medicare rates, potentially strengthening insurers’ bargaining power and leading to lower in-network payments.9 The Texas Medical Association brought similar concerns forward in a lawsuit, claiming that insurers would have undue leverage because the No Surprises Act sets the out-of-network payment to the median in-network rate.10 A federal judge recently sided with the Texas Medical Association,11 resulting in changes to the factors considered when assigning the out-of-network payment amount via the dispute resolution process.12
In this study, we examined the association of California’s law, known as Assembly Bill 72 (hereafter referred to as “the law”), with payments for in-network anesthesia care. An early study examined in-network anesthesia per-unit prices in states adopting surprise billing legislation and reported that in-network per-unit prices declined in California by 10.8% in the 6 months after the law’s implementation.7 Because many provider contracts cover periods longer than 6 months, tracking payments for a longer duration is essential for identifying any effects of the law. Using a large dataset of average payments derived from insurance claims for commercially insured patients that extended through the end of 2020, we assessed the extent to which the law’s implementation was associated with lower payments for in-network anesthesia care. We hypothesized that, after the implementation of the law, there would be no change in in-network payment amounts for anesthesia care, a decline in out-of-network payment amounts for anesthesia care, and a decline in the portion of anesthesia claims occurring out-of-network.
Methods
Background on California Assembly Bill 72
California’s balance billing law went into effect on July 1, 2017, and prohibited balance bills for patients in fully insured health plans who received nonemergency care at in-network hospitals, ambulatory surgery centers, laboratories, and radiology or imaging centers. Although the law encompassed all fully insured health insurance plans regulated by California’s Department of Managed Health Care and Department of Insurance, it did not apply to “self-insured” insurance plans regulated by the federal Department of Labor under the Employee Retirement Income Security Act law. Approximately 6 million, or 15%, of Californians are enrolled in self-insured plans.13 In addition, the law did not apply to emergency care—in particular, care provided by emergency medicine physicians—because balance billing in this context was previously banned in California in 2009.14–16
Data Sources and Study Cohort
Our study used payment data derived from the Merative MarketScan Commercial Database (Ann Arbor, Michigan), an administrative claims database of commercially insured individuals. This dataset provides a convenience sample of individuals (i.e., primary subscribers and dependents) nationwide who are younger than 65 yr and mostly enrolled in large employer-sponsored health insurance plans. We obtained data on utilization and payments for 20 selected procedure (Current Procedural Terminology) codes: 15 codes for anesthesia care (00142, 00160, 00170, 00300, 00400, 00670, 00790, 00840, 00952, 01400, 01480, 01630, 01810, 01961, 01967) and 5 codes pertaining to outpatient physician office visits (99201 to 99205), commonly known as “evaluation and management” codes (see table 1 for further descriptions of procedural codes). These codes are referred to as “procedure codes” by the American Medical Association but represent a wide variety of medical services; in this study, the procedure codes signified intraoperative or intrapartum anesthesia care and outpatient clinic care. We chose the given set of anesthesia-related codes because these represented the most commonly billed codes after excluding codes for upper (00740) and lower (00810) endoscopy. We excluded those two codes because they were discontinued and replaced with five new codes (00731, 00732, 00811, 00812, 00813) in 2018, shortly after implementation of the law; the association of the law with payments for endoscopy would therefore be difficult to disentangle from the change in these codes. We chose physician office visit codes, restricted to those filed in the outpatient setting, to create a comparison group of payments for common medical encounters that we expect would not have been affected by the law. We reasoned that they could be used to represent general trends in payments for medical care during the study period unrelated to the law. Although in theory the law could apply to balance billing for outpatient physician office visits, balance billing rarely occurs in this setting.17,18
Procedure Code Definitions and Average Payments Before California Balance Billing Law Implementation
For each of these procedure codes, we then obtained a dataset consisting of (a) the number of claims billed by quarter and (b) the average amount paid to practitioners for the code at the county level for each California county and for each calendar quarter between 2013 and 2020. The dataset was stratified by practitioner network status on claims (in-network, out-of-network, and missing), with network status provided by the MarketScan Commercial Claims Database and derived from raw adjudicated claims-level data from individual health insurance payers. The dataset excluded information from plans classified as health maintenance organizations because reimbursements for those organizations are typically paid on a per-member basis rather than for individual services or procedures.
The initial dataset consisted of 53,475 procedure code-county-quarter-network combinations. We excluded 9,747 observations where network status was missing, resulting in a final dataset of 43,728 procedure code-county-quarter-network combinations. These data represented 4,599,936 aggregated claims.
This study was exempted from Institutional Review Board approval given that all data were de-identified. A prespecified analysis plan defining the study sample, set of associated variables, and data analysis plan was written, date-stamped, and recorded in the investigators’ files before data were accessed.
Exposure
The independent variable of interest was whether the law was in effect for the given procedure code in each given quarter of the study period. The law took effect in July 2017, so we defined anesthesia codes as being potentially affected by the law starting in the third quarter of 2017 and as not affected if earlier.
Outcomes
Our primary outcome of interest was the average payment for in-network anesthesia care. This was defined as the total amount paid to the practitioner, which was the sum of payments made by the patient’s insurer as well as any cost-sharing (e.g., copayments) by the patient. Secondary outcomes were average amounts paid by insurers for out-of-network anesthesia care and the incidence of out-of-network billing.
Other Covariates
We collected several continuous variables to adjust for potential time-varying county-level factors that could affect insurer payments. These covariates included annual county-level total population, median income, and share of the population that was white, male, or age 65 or older. We obtained annual population and income data at the county level from the U.S. Census Bureau (Washington, DC).19–21 Finally, we included county-, quarter-, and procedure-level fixed effects as described in Statistical Analysis.
Statistical Analysis
We estimated the association between implementation of the law and payments for anesthesia care using a difference-in-differences approach. Although our analyses included several controls for observable characteristics that could affect payments (i.e., changes in county-level demographic factors over time), a straightforward comparison of payments for anesthesia care before and after the law could still be biased by unmeasured confounders or secular time trends. The difference-in-differences approach was chosen to attempt to reduce confounding further. The difference-in-differences approach is frequently used in policy analysis and has been used in the previous literature examining the effects of balance billing.7 With the use of this approach, the first difference compares changes in payments for anesthesia procedure codes before and after the implementation of the law. However, this simple before-and-after comparison could still be subject to confounding from secular time trends, such as general changes in payments over time. Thus, the second step of a standard difference-in-differences approach involves the use of a comparison group to control for secular time trends. In this case, we used payments for physician office visits, which we expect would not have been associated with the law, to infer what the payment trend for anesthesia procedure codes would have been absent the law.
We implemented the difference-in-differences approach using a multivariable linear regression model in which the dependent variable was the natural log of the average payment for a given procedure in a given county and quarter. We chose to calculate the natural log of the average payment because, with this transformation, the regression results provide the percentage change in payment associated with the law, which is more easily understood across payments for different types of anesthesia procedures that are reimbursed at different absolute amounts. Our primary independent variable was an indicator representing whether the procedure code was potentially affected by the law. This variable was set equal to 1 for anesthesia-specific procedures in the third quarter of 2017 or later, and zero otherwise. We adjusted for county-, quarter-, and procedure-level fixed effects. We clustered standard errors at the procedure level,22 weighted our estimates by the frequency of the number of claims underlying the payment data associated with each of the procedure code-county-quarter-network combinations, and stratified the analysis by in-network and out-of-network payments. A similar approach was used to assess the association between the law and the incidence of out-of-network billing, except that, instead of average payments, the dependent variable was the percentage of claims that were paid as out-of-network.
Finally, to examine the magnitudes of in-network and out-of-network payments before and after the law, we also estimated the same models but set the dependent variable as the absolute dollar value of the payments (rather than the natural log as discussed earlier in Statistical Analysis). The result of this additional analysis provided the estimated average dollar change in payments across all 15 anesthesia procedures.
To provide a more detailed analysis of quarter-by-quarter changes in payments before and after implementation of the law, we also estimated an event-study model. This model incorporated coefficients for the interaction between indicators representing the lead or lag time for each quarter relative to the reference quarter (i.e., the quarter immediately before enactment of the law) and an indicator for whether the payment was for an anesthesia procedure, with adjustment for the covariates and fixed effects included in the model described earlier in Statistical Analysis.
All dollar values were adjusted to year 2021 dollars using the Consumer Price Index.23 To provide an a priori specified framework for discussion of the significance of study results for policy, for this study we chose to define a policy significant change as a change of at least 10% in either payments to practitioners or incidence of out-of-network billing, based on the estimated magnitude of the effect of this law reported elsewhere.7,16 Although policymakers and stakeholders may find results larger or smaller than this to be of interest, we hope that clarity of terminology will aid in providing a clear discussion. Power analyses were not conducted given that we used data from the full population of MarketScan beneficiaries in California. All analyses were performed using STATA 17.0 (College Station, Texas).
Sensitivity Analyses
We conducted several prespecified sensitivity analyses. First, we assessed for changes in in-network and out-of-network anesthesia payments over time after accounting for linear time trends at the California county level that may have differed from overall state trends. This analysis was similar to the primary analysis described in Statistical Analysis, with the addition of indicator variables for the interaction between county and year representing county-specific factors that may vary through time (e.g., year-over-year demographic shifts). Second, while our primary analysis used physician office visits in California as a comparison group because we expect that those procedure codes were not affected by balance billing reform, other studies7 have modeled this relationship by using anesthesia payments in other states as controls. We therefore used our difference-in-differences approach to estimate changes in in-network and out-of-network anesthesia payments for all states with balance billing reform implemented during our study period (12 states: California, New York, Florida, Connecticut, Oregon, New Jersey, New Hampshire, Colorado, Maine, New Mexico, Texas, and Washington) compared to states that did not have balance billing reform (i.e., all other states, excluding Illinois and Maryland, which enacted balance billing reform before our study period).5,7 Third, to replicate the shortened study period in the previous study,7 we repeated these analyses with our study period restricted to 2013 to 2017: we conducted this analysis first by setting California physician office procedure codes as the comparison group, and then as a separate second analysis by setting anesthesia payments in all non–balance billing reform states as the comparison group. Fourth, we assessed changes in our outcomes over time after aggregating procedures at the California state level, rather than at the county level. Finally, we assessed our outcomes accounting for the transition period between when the law was passed in the third quarter of 2016 and when it was enacted in 2017, by setting our exposure period to start in the third quarter of 2016 rather than 2017, because there may have been anticipatory changes in payments during this transition period.
We also conducted three post hoc sensitivity analyses. For the first analysis, we replicated the primary analysis but started the observation period in 2015 rather than 2013 given that our data met the parallel trends assumption most closely during that period. For the second analysis, in response to reviewer feedback, we accounted for payments with missing network status (which were excluded from the primary analysis) by reassigning them to out-of-network status. For the third analysis, also in response to reviewer feedback, we recalculated change in payments using the median (rather than the average) payment by procedure code-county-quarter-network combination.
Results
Our final sample was derived from 4,599,936 claims across all 58 counties in California that were aggregated to 43,728 combinations of procedure code-county-quarter-network (anesthesia: in-network 20,529, out-of-network 8,680; and physician office visits: in-network 8,749, out-of-network 5,770). We observed the mean payment for each combination.
Average payments for anesthesia procedures ranged from approximately $500 to $2,500 in the quarter immediately preceding the law, while payments for physician office visits ranged from approximately $70 to $300 (table 1). Plots depicting parallel trends before the California law implementation in in-network and out-of-network payments for anesthesia care and physician office visits are available in Supplementary Figure 1 (https://links.lww.com/ALN/D204).
Figure 1 shows the results of difference-in-differences analysis estimating the change in payments for in-network and out-of-network anesthesia care after the law implementation in the third quarter of 2017, after adjusting for county, quarter, and procedure fixed effects, as well as county-level demographic characteristics. The law was associated with a statistically significant decrease in payments for out-of-network anesthesia care (–13.6%, 95% CI, –16.5 to –10.6%; P < 0.001), which translated to an average $108 decrease in out-of-network payments (95% CI, –$149 to –$64; Supplementary Table 2, https://links.lww.com/ALN/D207). The law was associated with a statistically significant increase in payments for in-network anesthesia care (3.0%, 95% CI, 0.9 to 5.1%; P = 0.007), translating to an average $87 increase in in-network payments (95% CI, $64 to $110); however, this did not meet the threshold we prespecified to warrant referring to a result as policy significant. Finally, the law was not associated with a statistically significant change in the incidence of out-of-network billing (10.0%, 95% CI –4.1 to 24.2%; P = 0.155).
Percent change in payment amounts and portion of claims occurring out-of-network for anesthesia care after implementation of California Balance Billing Law. Adjusted percent change in in-network and out-of-network payments (in green), and the proportion of claims occurring out-of-network (in blue), for anesthesia care in the period after implementation of California’s balance billing law (after the third quarter of 2017) compared to the period before. Percentage change in payments (depicted as green columns) translated to an average $87 increase in in-network payments (95% CI, $64 to $110) and an average $108 decrease in out-of-network payments (95% CI, –$149 to –$64). Estimates were calculated using a difference-in-differences approach, weighted by the frequency of the number of claims underlying the payment data associated with each of the procedure code-county-quarter-network combinations, and adjusted for county, procedure, and quarter fixed effects, as well as county-level demographics. Anesthesia care was defined using the 15 most common Current Procedural Terminology codes representing intraoperative or intrapartum anesthesia care. Error bars represent 95% CI and were calculated using robust standard errors clustered at the procedure level. All dollar values presented are standardized to 2021 dollars. *P < 0.05.
Percent change in payment amounts and portion of claims occurring out-of-network for anesthesia care after implementation of California Balance Billing Law. Adjusted percent change in in-network and out-of-network payments (in green), and the proportion of claims occurring out-of-network (in blue), for anesthesia care in the period after implementation of California’s balance billing law (after the third quarter of 2017) compared to the period before. Percentage change in payments (depicted as green columns) translated to an average $87 increase in in-network payments (95% CI, $64 to $110) and an average $108 decrease in out-of-network payments (95% CI, –$149 to –$64). Estimates were calculated using a difference-in-differences approach, weighted by the frequency of the number of claims underlying the payment data associated with each of the procedure code-county-quarter-network combinations, and adjusted for county, procedure, and quarter fixed effects, as well as county-level demographics. Anesthesia care was defined using the 15 most common Current Procedural Terminology codes representing intraoperative or intrapartum anesthesia care. Error bars represent 95% CI and were calculated using robust standard errors clustered at the procedure level. All dollar values presented are standardized to 2021 dollars. *P < 0.05.
Figure 2 depicts results from our event study analysis, showing trends in estimated payments for in-network and out-of-network anesthesia care by quarter relative to the quarter immediately before the law was enacted (i.e., the second quarter of 2017) after accounting for secular trends in medical payments. Out-of-network payments generally decreased over time, particularly in the postlaw period, while in-network payments generally retained their magnitude with a slight downtrend by the end of 2020. Estimated coefficients for this model are available in Supplementary Table 6 (https://links.lww.com/ALN/D211), with reference average payment amounts by procedure code in the reference quarter available in Supplementary Table 7 (https://links.lww.com/ALN/D212). As shown in figure 2, in-network payment amounts for anesthesia care increased from the fourth quarter of 2014 to the first quarter of 2015 (i.e., quarter –11 to –10). We conducted additional post hoc analyses to determine whether a particular procedure code or county was associated with this increase, but these analyses did not identify a clear driver of the trend. Therefore, we added a post hoc sensitivity analysis (see table 2) redefining the study period to start in 2015 rather than 2013.
Association between California Balance Billing Law and Change in Outcomes, Primary and Alternative Models
Adjusted changes in average payment amounts through time for in-network and out-of-network anesthesia care, before and after California’s balance billing law. Adjusted differences in in-network and out-of-network payments for anesthesia care relative to the quarter immediately before balance billing reform was enacted on July 1, 2017 (i.e., payment amounts are relative to those in the second quarter of 2017). The y-axis is rescaled to the adjusted payment amount for the second quarter of 2017 (see Supplementary Table 6, https://links.lww.com/ALN/D211, and Supplementary Table 7, https://links.lww.com/ALN/D212, for adjusted coefficients depicted in this figure, and for estimated dollar amounts paid for anesthesia care by Current Procedural Terminology code in the second quarter of 2017). Payments were estimated using a difference-in-differences approach that accounted for secular time trends in payments (using physician office visit payments as the comparison group), adjusted for county, procedure, and quarter fixed effects and annual county demographic characteristics, and weighted by frequency of the number of claims underlying the payment data associated with each of the procedure code-county-quarter-network combinations. Error bars represent 95% CI and were calculated using robust standard errors clustered at the procedure level. The vertical line delineates pre– and post–balance billing law reform. All dollar values presented are standardized to 2021 dollars.
Adjusted changes in average payment amounts through time for in-network and out-of-network anesthesia care, before and after California’s balance billing law. Adjusted differences in in-network and out-of-network payments for anesthesia care relative to the quarter immediately before balance billing reform was enacted on July 1, 2017 (i.e., payment amounts are relative to those in the second quarter of 2017). The y-axis is rescaled to the adjusted payment amount for the second quarter of 2017 (see Supplementary Table 6, https://links.lww.com/ALN/D211, and Supplementary Table 7, https://links.lww.com/ALN/D212, for adjusted coefficients depicted in this figure, and for estimated dollar amounts paid for anesthesia care by Current Procedural Terminology code in the second quarter of 2017). Payments were estimated using a difference-in-differences approach that accounted for secular time trends in payments (using physician office visit payments as the comparison group), adjusted for county, procedure, and quarter fixed effects and annual county demographic characteristics, and weighted by frequency of the number of claims underlying the payment data associated with each of the procedure code-county-quarter-network combinations. Error bars represent 95% CI and were calculated using robust standard errors clustered at the procedure level. The vertical line delineates pre– and post–balance billing law reform. All dollar values presented are standardized to 2021 dollars.
Our primary analysis of in-network payments was robust to additional sensitivity analyses, all finding slight increases or nonsignificant decreases. Sensitivity analyses with the study period ending in 2017 found variable results with changes in out-of-network payments and portion of claims occurring out-of-network. It is unclear the extent to which sensitivity analyses using other states as comparisons meet the parallel trends assumption required for a difference-in-differences analysis as prelaw trends for states that passed balance billing legislation appear to diverge from states that did not (Supplementary Figure 2, https://links.lww.com/ALN/D205). Of note, the sensitivity analysis with the study period starting in 2015 instead of 2013 estimated a slight downtrend—rather than uptrend—in in-network payment amounts that did not meet our definition of statistical or policy significance (–1.0%, 95% CI, –3.5 to 1.5%; P = 0.416). Findings from this sensitivity analysis also diverged in calculation of the incidence of out-of-network billing, estimating a decrease in this practice rather than an increase (–10.0%, 95% CI, –21.8 to 0.4%; P = 0.058), which met our definition of policy significance but did not achieve statistical significance. Regression coefficients for primary and secondary outcome variables for our primary analysis as well as selected sensitivity analyses are presented in table 2, and detailed descriptions and estimates for all sensitivity analyses are presented in Supplementary Tables 1 to 5 (https://links.lww.com/ALN/D206, https://links.lww.com/ALN/D207, https://links.lww.com/ALN/D208, https://links.lww.com/ALN/D209, https://links.lww.com/ALN/D210).
Discussion
On January 1, 2022, the U.S. Congress implemented the No Surprises Act, banning surprise balance billing at the federal level for emergency and nonemergency care received from an out-of-network facility or provider. Patients who receive such out-of-network care are protected from surprise balance bills. Instead, the out-of-network practitioner and insurer must come to an agreement on the final payment, which is based either on the state’s own laws regarding balance billing payments, or—if state laws do not exist or apply—on a federal benchmark of median in-network price. The federal benchmark used for independent dispute resolution has been challenged by medical professional advocacy groups, and final rulings have recently been revised to account for additional factors, including patient acuity, teaching hospital status, and hospital case-mix.12 California’s experience with balance billing reform via Assembly Bill 72 provides several years of insight into how using a payment benchmark may affect market dynamics between insurers and healthcare providers. California’s benchmark—the greater of the local average in-network payment or 125% of the Medicare rate—is also similar in design to the federal benchmark in contrast to those of other states such as New York, which uses a payment based on the 80th percentile of billed charges.
We found associations between California’s balance billing law and in-network payments, but these were smaller than the threshold that we had prespecified to support referring to a result as policy significant. Our primary and most sensitivity analyses found that there was a statistically significant increase in in-network payment amounts in the 3 yr after implementation of the law, mostly ranging from 2 to 6% increases (with one sensitivity analysis estimating a 1% decrease), but none of these estimates exceeded the 10% threshold for policy significance that we prespecified for this analysis. We chose a 10% threshold for defining policy significance based on the magnitude of estimates from other studies on balance billing in California,7,16 as well as our reasoning that a 10% change in payment amounts associated with this law would be noteworthy for policymakers, payers, and providers. However, had we chosen a lower threshold for policy significance, some of our conclusions on in-network payment amounts would have more broadly encompassed stakeholders for whom smaller changes in payment amounts may be consequential. We also found that California’s law was associated with a statistically and policy significant approximate 14% decrease in payment amounts for out-of-network care. Finally, we found mixed results on the occurrence of out-of-network billing, with our primary model and most sensitivity analyses showing an approximate 10% increase in out-of-network billing, and one sensitivity analysis showing an approximate 10% decrease—trends that may be meaningful in the real world but did not reach statistical significance.
We examined the association of California’s law with in-network payments in multiple ways during the 3 yr after implementation: most of our analyses showed that the implementation of the law corresponded with a slight increase for in-network payments. Our quarter-level estimates did show a general downtrend in in-network payments, raising the possibility that California’s balance billing law may have a delayed effect on in-network payment amounts. Overall, our findings suggest that—despite concerns to the contrary by medical professional stakeholders—California’s law did not meaningfully lower in-network payments in the 3 yr after implementation. Further, they provide some, albeit indirect, insight into how regulation may affect payments in the private insurance market. Several medical lobby groups9,24 predicted that in-network payments would decrease because insurers would have undue leverage when negotiating contracts and rates with physician groups. In the broader context of the No Surprises Act, those in other medical specialties (e.g., radiology) also predicted that in-network payments may decrease due to changes in bargaining power dynamics.25 However, multiple factors influence negotiations between insurers and physician groups and may have attenuated the potential impact of California’s law on in-network payments for anesthesia care. First, insurers are constrained by network adequacy laws that dictate provider-to-enrollee ratios and maximum travel times.26 Second, the degree of local network consolidation may limit insurers’ negotiating power: for example, if only one hospital group exists in a community, market forces and state requirements may limit the ability of the insurer to exclude the provider from a network. Other studies have found that network consolidation has indeed increased in California during our study period, particularly in rural counties.27 In particular, by limiting payments for out-of-network care, California’s law may have encouraged anesthesia providers to consolidate in the hope of negotiating higher in-network rates. Third, increasing demand for anesthesia care—as case volume increases, particularly in non–operating room anesthesia care, or as the number of available anesthesiologists decreases—may offset any change in insurers’ negotiating power due to balance billing legislation.28,29 Fourth, the value of existing relationships between anesthesiologist groups and surgeons, hospitals, and/or insurers may outweigh potential cost-savings from switching to a different anesthesiologist group.
Previous studies of balance billing in anesthesia care evaluated anesthesia unit prices, which are conversion factors used in anesthesiology to standardize payments by procedural complexity, length of time, and patient comorbidities. Those studies have found that out-of-network prices increased after balance billing reform in New York,8 and that in-network prices decreased after reform in California.7 Our dataset provided payment amounts instead of prices, so our analysis also focused on payment amounts. However, this difference is unlikely to explain why our findings diverged from previous work, given that we accounted for procedural and county-level fixed effects and because it is unlikely that surgical times would have dramatically changed during the study period.30,31 Rather, there are several potential other explanations. New York’s implementation of balance billing reform advised arbiters to settle out-of-network payments at the 80th percentile of billed charges, which are usually substantially higher than both average in-network payments and 1.25 times Medicare rates. This benchmark may have incentivized groups to remain out-of-network to collect higher payments and may have led to overall out-of-network price inflation. Regarding the previous study on California prices,7 there were several differences that may explain why our findings diverge. First, California’s law took effect on July 1, 2017, and the previous study’s end date was December 31, 2017, so the study was only able to assess very short-term impacts. Insurance contracts may cover multiple years, so the effects of California’s law may not have manifested before the end of that study’s observational period. Second, to minimize confounding, the previous study used a comparison group of states that had no balance billing protections. Therefore, a key assumption of the previous study is that trends in medical prices outside California serve as a proxy for what prices in California would be absent the passage of its balance billing law. Because factors determining medical prices—such as the degree of consolidation among healthcare practitioners—can vary across states, it is not clear that this assumption would necessarily hold. There is some evidence from the study itself that—relative to control states—prices for anesthesia care were already falling in advance of California’s law. Our sensitivity analysis replicated the approach of using other states as controls and found that trends in in-network payments for states with and without balance billing reform were indeed not parallel before implementation of balance billing reform. Finally, by ignoring intrastate correlation in prices, the study likely overstated the statistical significance of its findings.
Our study should be viewed considering its limitations. First, as a retrospective study, it is subject to potential residual confounding and we are unable to draw causal conclusions. We used a difference-in-differences approach that we expect will minimize these unobserved confounders. Second, our study period was limited to 3 yr after implementation of the law and included the initial phase of the COVID-19 pandemic; it is possible that more significant changes in payments or insurance contracts would be delayed beyond the end of our study period and would not be reflected in our data, or that a change in surgical volume due to the pandemic may have impacted our findings. Third, our data did not allow us to identify self-insured plans governed under the Employee Retirement Income Security Act, which would not have been subject to California’s law; however, only 15% of Californians are enrolled in self-insured plans,13 so potential variability in payments for these plans may not have significantly skewed our findings. Finally, our data did not include all commercial claims in California, so the results of our study may not apply to all patients with commercial insurance; however, the MarketScan database is a substantial repository of commercial claims that we believe still provides useful information for studying this policy’s impact on a large population.
Conclusions
In summary, we found that California’s balance billing law was associated with (1) a statistically significant increase in in-network payment amounts that did not meet the threshold for policy significance, (2) statistically and policy significant declines in out-of-network payment amounts, and (3) increases in the occurrence of out-of-network billing that did not meet the threshold for statistical significance, all during the 3 yr after implementation.
Research Support
This study was supported by grant T32 GM 089626 from the National Institutes of Health (Bethesda, Maryland, to Dr. Dixit) and grant K08DA042314 from the National Institute on Drug Abuse (Bethesda, Maryland, to Dr. Sun) . The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Competing Interests
Dr. Sun is on the advisory board of Lucid Lane, LLC (Los Altos, California) and reports receiving consulting fees for matters involving healthcare providers and health insurers unrelated to this work. Dr. Heavner is a Managing Principal at Analysis Group in Los Angeles, California. Analysis Group is a consulting firm with expertise in economics, finance, healthcare analytics, and strategy. Analysis Group’s clients include healthcare providers, health insurers, manufacturers, and distributors of pharmaceutical products, and others who may have an interest in the subject matter of this article. However, neither Analysis Group nor Dr. Heavner received any compensation for this study, nor did Analysis Group provide any funding for this study. In addition, the views presented in this article are those of the authors and do not represent any opinions or positions of Analysis Group or its affiliated companies. Dr. Baker reports receiving consulting fees from Blue Shield of California (Oakland, California), UnitedHealthcare (Edina, Minnesotra), Anthem (Indianapolis, Indiana), Kaiser Permanente (Oakland, California), Cedars-Sinai Medical Center (Los Angeles, California), Dignity Health (San Francisco, California), Allcare IPA (Modesto, California), Geisinger Health (Danville, Pennsylvania), Analysis Group, Cornerstone Research (San Francisco, California), other healthcare providers, and makers and distributors of pharmaceutical products. Dr. Dixit declares no competing interests.
Supplemental Digital Content
Supplementary Figure 1, https://links.lww.com/ALN/D204
Supplementary Figure 2, https://links.lww.com/ALN/D205
Supplementary Table 1, https://links.lww.com/ALN/D206
Supplementary Table 2, https://links.lww.com/ALN/D207
Supplementary Table 3, https://links.lww.com/ALN/D208
Supplementary Table 4, https://links.lww.com/ALN/D209
Supplementary Table 5, https://links.lww.com/ALN/D210
Supplementary Table 6, https://links.lww.com/ALN/D211
Supplementary Table 7, https://links.lww.com/ALN/D21
