Hospital finances following Connecticut's ban on outpatient facility fees

Abstract

Introduction

In 2017, Connecticut prohibited hospitals from billing “facility fees” for certain services delivered off their main campuses.

Methods

Using Hospital Medicare Cost Reports and the National Academy for State Health Policy Hospital Cost Data (2011-2022), we applied a difference-in-differences model within an event study framework to compare hospital financial outcomes between Connecticut hospitals and matched control hospitals nationwide.

Results

Unadjusted trends suggest small declines in operating margins and outpatient-to-total charges after the ban, alongside an increase in inpatient-to-total charges. In adjusted analyses, operating margins and inpatient-to-total charges did not change significantly, while outpatient-to-total charges declined significantly by 6.91%.

Conclusion

Given the small sample and limited precision, a policy effect cannot be ruled out for several outcomes. Future research evaluating the long-term sustainability and impacts on hospital financials, access, and care quality are warranted.

Introduction

Hospital “facility fees” are operational charges billed by hospitals for outpatient services delivered at hospital-owned facilities, separate from the fees for professional services. Often when a patient receives care in a hospital-owned setting, the bill is “split” into 2 components: one for the clinician and one for the facility.¹ In contrast, freestanding physician offices use a single bill, combining overhead costs and clinician fees. This approach originated from Medicare's traditional billing structure for inpatient care, which paid hospitals and doctors separately. Encouraged, in part, by the potential to collect more revenue from outpatient facility fees, hospitals expanded into and shifted care toward ambulatory settings, and facility fees became a common part of bills, even for some routine office visits.²

Nationwide, outpatient services have grown to account for roughly half of hospital revenue.³ Hospital leaders justify facility fees as essential for covering the higher operational costs that hospitals bear (eg, IT systems, infection control, emergency readiness, etc.).⁴ In addition, they assert that facility fee revenue is needed to subsidize under-reimbursed care and standby capacity for expenses associated with high-cost inpatient services such as trauma care, emergency surgery, and intensive care units (ICUs).^5-7

Outpatient facility fees have become increasingly controversial, however, due to their impact on healthcare spending and patients' bills. Patients receiving care in hospital outpatient settings can face greater cost-sharing when these fees are in play,⁸ while the higher payments covered by insurers feed into higher premiums.⁹ Studies have found that when hospitals acquire physician practices and start billing facility fees, the price of outpatient care rises significantly (more than 14% in one study), with nearly half of that increase directly attributable to hospitals exploiting payment rules like facility fee billing.^8,10-12 An additional study found that higher physician-hospital integration was associated with increased spending, driven almost entirely by higher prices per service (both facility and professional fees), rather than changes in utilization.¹³

Connecticut's 2017 facility fee ban

In 2015, amid reports of patients being charged new facility fees after their longtime physician practices were acquired by hospital systems, Connecticut enacted legislation eliminating certain hospital outpatient facility fees for commercial plans.^14-17 The law prohibited hospitals from charging a facility fee (defined as any outpatient charge intended to compensate a hospital for its operational expenses, separate and distinct from a professional fee) for outpatient evaluation and management (E + M) visits at off-campus clinics. This ban took effect on January 1, 2017, but the law contained a grandfathering clause allowing facility fee billing to continue under insurance contracts in place before the law was passed. The law also imposed transparency and reporting requirements.¹⁸

Data from the Connecticut Office of Health Strategy collected before the ban showed hospitals collected hundreds of millions annually from outpatient facility fees; Yale New Haven Health alone generated an estimated $9.9 million in 1 year from 5-min outpatient visits, charges now prohibited under the law.¹⁹ A fiscal analysis of the bill estimated that the ban would reduce revenue only modestly at UConn Health Center, a government-owned hospital (∼$850 000 in FY 2017). The analysis did not estimate effects on non-governmental hospitals.⁽²⁰⁾

There is no peer-reviewed evidence evaluating the impact of Connecticut's facility fee ban on hospital finances. Using Medicare Cost Reports, we sought to address 2 questions: first, is the facility fee ban associated with changes in Connecticut hospital profitability? Second, do specific financial metrics (eg, lower revenue vs lower costs) drive-associated changes in hospital profitability?

Data and methods

Study design

We use an event study with a difference-in-differences (DID) design to identify the effect of Connecticut's facility fee ban on hospital financial performance. The facility fee ban applied to hospitals licensed in the state of Connecticut; hence our treatment group is Connecticut hospitals subject to the state ban while hospitals in other states serve as our control. We excluded hospitals from Massachusetts, where major insurance carriers reported limiting outpatient facility fee billing in most health-system contracts during the study period, and Maine, which had enacted broader limitations on outpatient facility fee billing in 2005 (Tables S1 and S2). The pre-treatment time period starts in 2011 through 2016, while the post-treatment period runs from 2017 through 2022.

Study data

We analyzed data from Medicare Cost Reports and the National Academy for State Health Policy (NASHP) Hospital Cost Data covering the period 2011-2022, supplemented with community-level demographic and socioeconomic characteristics from the American Community Survey (ACS) 5-year estimates (2011-2022). The Medicare Cost Reports provided detailed hospital-level financial outcomes and facility characteristics.²¹ NASHP Cost Data provided hospital-level financial outcomes for the commercial population. We obtained market-level covariates from the ACS, merging these data by hospital zip code to capture local demographic factors.

Study sample

To reduce confounding and improve the validity of the parallel trends assumption, we applied propensity score matching (PSM) prior to our DID analysis. Matching was conducted at the hospital level using data from the pre-treatment period (calendar year ≤2016) (Methods S1). Specifically, we matched Connecticut hospitals to up to 5 non-Connecticut hospitals using nearest-neighbor matching (with replacement, 0.1 SD caliper) on key baseline characteristics—health-system affiliation, bed size, ownership status, Case-Mix Index (CMI), change-of-ownership—and zip-code–level employer-sponsored and non-group insurance enrollment, and we included each hospital's pre-treatment operating-margin trend. We assessed covariate balance before and after matching using standardized differences and graphical diagnostics (Figure S1).

We then restricted our sample to a balanced panel, where all hospitals must have been present in every year of the data from 2011 to 2022. The final matched sample contained Connecticut hospitals (n = 15) and a reweighted group of matched control hospitals (n = 63). Matching weights derived from the PSM procedure were applied to the entire 2011-2022 panel dataset, and our subsequent DID analysis was conducted using this weighted matched sample.

Statistical analysis

Following PSM, we estimated treatment effects using a DID model within an event study framework with hospital-level panel data spanning from 2011 to 2022. Before estimating the DID, we plotted unadjusted outcome trends for the matched treatment and control groups to visually assess parallel pre-trends. The DID model included hospital fixed effects to control for time-invariant differences across hospitals and year fixed effects to account for secular trends in outcomes across all hospitals. The primary exposure was an interaction term between an indicator for being a Connecticut hospital (treatment group) and an indicator for the post-policy period (2017 onward), representing the estimated effect of Connecticut's facility fee ban (see Methods S1). To account for the matched sampling design, we applied matching weights derived from the PSM procedure. We computed inference using wild cluster bootstrap-t P-values with states as clusters, alongside conventional state-clustered standard errors, to account for the limited number of state clusters and avoid overstating precision. We also conducted a randomization-inference placebo test: re-estimating the DID while, in turn, assigning each non-Connecticut state a placebo treatment starting in 2017, holding the sample and specification fixed. Next, we addressed potential COVID-19 contamination with 2 robustness checks: (1) re-estimating the model using data through 2019 only (pre-pandemic), and (2) re-estimating the full panel excluding 2020 while retaining 2021-2022.

Outcomes and covariates

We examined a range of facility and financial characteristics for hospitals in the treated and control groups. Facility characteristics included whether the hospital was affiliated with a larger health system, the number of licensed beds, and the share of patients covered by employer-sponsored or non-group (individual market) insurance. We also captured ownership status using a binary indicator for for-profit ownership and included the Medicare impact CMI to account for patient complexity and acuity. A binary indicator for change-of-ownership was used to identify hospitals that experienced an ownership transition during the year.

Financial characteristics included overall operating margin and commercial operating margin. Operating margin measures the profitability of a hospital's core services by converting the proportion of operating income per revenue dollars into a ratio comparable across organizations of various sizes. A higher operating margin indicates better financial health and efficiency in delivering healthcare services, reflecting the organization's ability to sustain its core operations profitably.

To assess potential shifts in service delivery between outpatient and inpatient settings, we examined the outpatient charges-to-total charges ratio and the inpatient charges-to-total charges ratio. These metrics capture the proportion of charges derived from outpatient or inpatient services compared to the total services provided. A lower outpatient ratio may indicate a shift toward less or different levels of outpatient care, which could be influenced by changes in a facility fee ban. The logged revenue-to-charge ratio, calculated as net patient revenue divided by total charges, served as a measure to supplement interpretations of outpatient charges. Statistically non-significant differences in the logged revenue-to-charge ratio could suggest that while charges were higher than revenue, they were not different across the treatment and control group.

To assess potential longer-term associations of the ban on organizational financial health, we measured reinvestment in capital purchases (as a percentage of total expenses). Hospitals spending a greater proportion on capital investment may indicate longer-term financial health, in particular, beyond nearer-term metrics like profitability.

To evaluate commercial financial performance specifically, we included several payer-specific outcomes: commercial hospital operating profit/loss (total commercial revenue minus commercial costs), adjusted commercial profit/loss (net of payer mix), commercial breakeven point (total commercial revenue required to cover costs), and commercial payer mix (share of patients covered by commercial insurance). These measures were used to assess whether the facility fee ban had differential impacts on hospitals' profitability from commercially insured patients and their exposure to commercial reimbursement.

Limitations

This study has several limitations. First, the sample size was constrained by the number of hospitals affected by the policy; only 15 Connecticut hospitals were included in the treatment group, which limited statistical power and resulted in wide confidence intervals for several outcomes. Second, the follow-up window ended in 2022 capturing only 5 years post-policy implementation as well as the COVID-19 pandemic. This period is insufficient to observe longer-term effects, including those related to grandfathered contracts, delayed provider responses, or broader market shifts following the COVID-19 pandemic.

Third, while cost reports are a widely adopted source of financial information, they have been shown to report certain data inconsistently and do not provide more granular data on outpatient utilization and visit-level pricing, which could have provided a clearer picture of the mechanisms underlying observed changes. Still, the accessibility of financial data for all hospitals receiving Medicare reimbursement makes cost reports a commonly accepted source of financial data. Some of our analyses use charges instead of revenue. To check that this did not distort results, we tracked the revenue-to-gross-charges ratio over time. Because this ratio moved the same way for treated and control hospitals, it means the “conversion” from charges to revenue was stable and similar across groups—so differences we see in charges should reflect comparable differences in revenue, making charges a reasonable proxy for inpatient and outpatient trends. Finally, Connecticut's hospital market—characterized by a unique payer mix, previous transparency legislation, and specific market dynamics—may limit the generalizability of our findings to other states.

Results

Hospital characteristics

Table S3 compares unadjusted means for facility, state, and financial characteristics between Connecticut hospitals and the matched control group for 2016 and 2022, time periods before and after the 2017 facility fee ban implementation in Connecticut. After matching, there are 15 Connecticut hospitals and 63 similar control hospitals in 2022. Matched characteristics demonstrate the similarity of the treatment and control groups prior to the facility fee ban.

Examining the unadjusted outcomes over time, operating margins declined for Connecticut and matched control hospitals, but trajectories differed. Connecticut hospitals decreased from pre-2017 to near zero by 2019 and turned negative by 2022, whereas matched controls decreased around 2017-2018, increased during 2018-2021, and then regressed in year 2022 (Figure 1). For outpatient-to-total charges, matched controls rose steadily throughout the period, while Connecticut increased through 2016 but flattened and dipped after 2017 before partially increasing—remaining below the matched control trend by 2022 (Figure S2). Consistent with this, inpatient-to-total charges continued to decline among matched controls, but in Connecticut the inpatient share stabilized around 2017 and then increased through the early 2020s (Figure S2).

Figure 1.

Unadjusted trends in operating margins between treated (Connecticut) and matched control facilities. Source: Authors' analysis of Hospital Cost Report data from 2011 to 2022.

Between 2011 and 2022, both control and Connecticut hospitals experienced increases in net patient revenue and total charges. Capital expenditures declined in both groups. Day’s cash on hand decreased for control and Connecticut hospitals. Commercial profitability measures remained stronger for control hospitals, with higher operating profit/loss and margins compared to Connecticut hospitals over time.

Changes in outcomes after Connecticut's facility fee ban in 2017

Adjusted DID results for Connecticut's 2017 facility fee ban are shown in Table 1. After controlling for year and hospital fixed effects and comparing Connecticut hospitals with matched controls, overall operating margin declined but was not statistically significant, and commercial operating margins showed no detectable change—findings consistent with limited sample size reducing precision and our ability to detect modest effects. Event study plots for overall and commercial operating margins (Figure 2) remained stable and parallel pre-ban and showed no clear changes post-ban. Outpatient charges as a share of total charges decreased significantly by 6.91% (P < .01) while inpatient charges as a percentage of total charges increased by 6.24% (P > .1), but was not statistically significant. Event study plots in Figure 3 show that after the ban, outpatient charges steadily decreased, and inpatient charges steadily increased, with both trends stabilizing ∼ 4 to 5 years after the policy went into effect. Capital expenditures relative to total purchases decreased by 14.16% (P < .01) in Connecticut hospitals relative to matched controls. However, after examination of the unadjusted plots (Figure S2), they regress to matched control estimates in the post period. No other financial outcomes significantly changed post-ban (adjusted event study plots reported in Figures S3-S11).

Table 1.

Adjusted difference-in-difference estimates.

	Adjusted difference-in-difference
	Difference (95% CI low–95% CI high)	P	Relative change, %
Operating margin	−2.04 (−5.47, 1.39)	.24	−46.50
Total outpatient-to-total charges	−3.86 (−7.27, −0.44)	.03	−6.91
Capital expenditures to total purchases	−0.90 (−1.51, −0.29)	.003	−14.66
Net patient revenue	$28 444 050.00 (−$39 900 000.00, $96 800 000.00)	.46	6.21
Total charges	$63 317 080.00 (−$191 000 000.00, $317 000 000.00)	.63	4.66
ln revenue-to-gross charges	0.003 (−0.048, 0.054)	.91	−0.29
Total inpatient-to-total charges	2.87 (−1.34, 7.08)	.22	6.24
Days cash	−10.17 (−86.68, 66.34)	.8	−5.55
Commercial hospital operating profit/loss	−$12 800 000.00 (−$35 200 000.00, $9 518 810.00)	.27	−10.76
Payer mix adjusted profit/loss commercial	0.33 (−2.96, 2.31)	.81	2.16
Commercial breakeven	5.79 (−16.61, 28.19)	.62	4.08
Commercial operating profit margin	−2.24 (−7.97, 3.50)	.46	−5.34
Commercial payer mix	1.11 (−1.22, 3.45)	.89	3.04

Source: Authors' analysis of Hospital Cost Report and NASHP data from 2011 to 2022.

Figure 2.

Association of Connecticut's facility fee ban on hospital overall and commercial operating margin and by year, 2011-2022. Source: Authors' analysis of Hospital Cost Report and NASHP data from 2011 to 2022. The figure illustrates the difference in overall and commercial operating margin in Connecticut vs non-Connecticut hospitals. Non-Connecticut hospitals are within states that do not have a facility fee ban.

Figure 3.

Association of Connecticut's facility fee ban on hospital outpatient and inpatient share of total charges and by year, 2011-2022. Source: Authors' analysis of Hospital Cost Report data from 2011 to 2022. The figure illustrates the difference in overall and commercial operating margin in Connecticut vs non-Connecticut hospitals. Non-Connecticut hospitals are within states that do not have a facility fee ban.

Our randomization-inference placebo test—assigning a 2017 “treatment” to each non-Connecticut state in turn—produced no systematic effects on outpatient-to-total charges, indicating the Connecticut result is unlikely to be driven by chance assignment (Figure S12). To address potential COVID-19 contamination, we re-estimated the model using data through 2019 only and, separately, excluding 2020 while retaining 2021-2022. Results were substantively unchanged in both checks (Figures S13 and S14).

Conclusion

Between the narrow scopes of Connecticut's 2017 ban—which applied only to off-campus E + M services, a small portion of outpatient services—and the law's then-broad grandfathering clause, one reasonably may expect such a ban to have minimal effects on hospitals. Consistent with that expectation, over the 5 years following implementation, this study did not find evidence of an association between Connecticut's 2017 facility fee ban and hospital operating margins. However, wide confidence intervals and limited sample size mean a policy effect cannot be ruled out. Unadjusted plots show Connecticut's inpatient share rose relative to matched controls; in adjusted analyses, the inpatient-to-total charges increased by 6.24%, but the change was not statistically significant. Additionally, Connecticut hospitals' outpatient charges had a relative decrease of 6.91% as a portion of total charges.

Together, these observations suggest that the ban may have been associated with hospital billing for certain outpatient services. That is, while narrowly targeted, the law effectively stopped hospitals from billing outpatient facility fees to a noticeable degree. At the same time, it appears that hospitals were able to adapt to these changes such that the law did not affect their operating margins.

This result is consistent with economic theory. While the law changed what hospitals could bill, it did not change their relative bargaining power. Nothing in the law limited the ability of Connecticut hospitals from increasing their charges for and revenue from other items and services, or through other forms of payment. The concern that hospital billing and payment reform initiatives that aim to reduce spending in one area simply result in increased spending in another area, like squeezing one end of a balloon, has been much discussed over the years.^(22-27)

This outcome also tracks anecdotes from other states. Some payers have reportedly negotiated for limitations on outpatient facility fee billing in hospital contracts in a “budget neutral fashion.”²⁸ The state of Indiana recently barred payer-hospital contract terms protecting each other from revenue losses caused by law or regulation, presumably because such revenue guarantees exist in some contracts today.²⁹

Our findings and the related literature should temper any expectations regarding the potential for outpatient facility fee bans alone to lower aggregate health care spending. This does not mean such reforms come without benefit: our research did not address the extent to which Connecticut's 2017 ban may have reduced consumer out-of-pocket spending. What's more, these findings suggest that outpatient facility fee restrictions comparable to Connecticut's can be implemented without affecting, much less imperiling, hospitals' financial conditions. Our findings also do not preclude the possibility that outpatient facility fee bans in tandem with other hospital rate regulations could generate system-level savings. Evidence indicates healthcare providers can become more efficient when faced with declining payment rates.^(30-32)

Important opportunities for further research persist. Analyses of claims data, utilization, and/or hospital-payer contracts could shed further light on how Connecticut hospitals and commercial payers shifted spending from outpatient-to-inpatient care. In addition, Connecticut broadened its facility fee ban in mid-2024 to apply to many on-campus outpatient departments.³³ Future research should evaluate the effects of this expansion––particularly in comparison to this study's estimated effects of the original law––as well as identify whether longer-term effects of the law differ from those identified here.

More broadly, our research examined finances at the hospital level, but Connecticut's law may have affected charges from clinicians working in off-campus settings and revenue and margins of the health systems in which hospitals and clinicians participate. To gain a more complete picture of how this policy affected healthcare provider revenues and financial performance, one needs additional sources of financial data at the health-system level. Such data are not included in the Medicare Cost Reports. Policy opportunity exists to increase financial transparency across affiliated providers. One pathway could be to expand information captured on Medicare Cost Reports, an existing form that hospitals already file annually. Increasing mineable financial data across owned entities within a health system would enable more robust policy evaluations.

This study provides the first analysis of how Connecticut's ban on hospital facility fees was associated with hospital finances. While hospitals' overall profitability did not significantly change after the ban, there was a clear shift in how hospitals billed for services, with outpatient charges decreasing relative to total charges. These changes could occur in tandem if hospitals adjusted billing or negotiated new contracts with insurers to compensate for reduced outpatient revenue through other means.

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

This work was supported by funding from the Gary and Mary West Health Policy Center to Georgetown University (AWD-7776045; C.H.M., principal investigator). The authors of this article are responsible for its content. The ideas and opinions expressed herein are those of the authors, and endorsement by the funder is not intended nor should be inferred.