State Laws and Nonprofit Hospital Community Benefit Spending

Abstract

Objective:

To determine the association of state laws on nonprofit hospital community benefit spending.

Design:

We used multivariate models to estimate the association between different types of state-level community benefit laws and nonprofit hospital community benefit spending from tax filings.

Setting:

All 50 US states.

Participants:

A total of 2421 nonprofit short-term acute care hospital organizations that filled an internal revenue service Form 990 and Schedule H for calendar during years 2009–2015.

Results:

Between 2009 and 2015, short-term acute care hospitals spent an average of $46 billion per year in total, or $20 million per hospital on community benefit activities. Exposure to a state-level community benefit law of any type was associated with an $8.42 (95% confidence interval: 1.20–15.64) per $1000 of total operating expense greater community benefit spending. Spending amounts and patterns varied on the basis of the type of community benefit law and hospital urbanicity.

Conclusions:

State laws are associated with nonprofit hospital community benefit spending. Policy makers can use community benefit laws to increase nonprofit hospital engagement with public health.

Every year, nonprofit hospitals benefit from a federal income tax exemption. In 2002, the Government Accountability Office estimated the value of this tax exemption at approximately $12.6 billion.¹ More recently, Rosenbaum et al estimated the value at $24.6 billion, and Herring et al estimated that the value of the tax exemption amounted to 5.9% of total hospital expenses.^2,3 Since 1969, the internal revenue service (IRS) community benefit (CB) standard has required nonprofit hospitals to provide CB in order to qualify for tax exemptions.⁴ However, prior to the Patient Protection and Affordable Care Act (ACA), neither the IRS nor any other federal entity specified what CB should entail on minimum spending amounts. As a result, states were left to interpret the federal CB standard and have developed varied approaches to increase nonprofit hospital accountability for provision of CB.⁵ The ACA aimed to address this inconsistency with nationwide CB requirements and reporting standards.^6,7 However, wide variation in CB spending remains across the nation.²

There is no consensus methodology for determining an “appropriate” level of nonprofit hospital spending on CB activities needed to justify the nonprofit tax exemption, though a handful have been suggested.^5,8 The IRS has never revoked nonprofit status due to low CB spending. In addition, appropriate CB spending allocation likely looks different depending on community characteristics, for example, well-insured populations may value spending on education or research, while uninsured populations may need financial assistance to remain the primary focus.⁵ There may be other factors that influence CB spending including hospital financial performance, governing board composition, and mission, as well as regulations and laws. A better understanding of the impact of state CB laws can aid policy makers in developing or further refining CB laws at the state or national level.

A handful of studies have examined the impact of state-level CB law on nonprofit hospital CB spending and provision of community-oriented services.^9–15 A recent study by Singh et al⁹ found that a requirement for hospitals to conduct a community health needs assessment was associated with higher CB spending and that reporting requirements were positively associated with total CB spending and direct patient care services but negatively associated with spending on community health activities. They also found that the existence of a state-level minimum CB spending requirement was associated with larger spending on community health initiatives and lower spending on direct patient care services and total CB.

In another study, Ginn and Moseley¹¹ examined the impact of state CB laws on the provision of CBs by comparing nonprofit hospitals with their for-profit counterparts. They found that hospitals in states with CB laws reported more community-oriented activities than hospitals in states without CB laws but that this effect was also observed among for-profit hospitals.

In this article, we investigate hospital responses to state-level hospital CB spending and reporting laws and whether the presence of these laws is associated with hospital CB spending. Our analysis builds upon previous work by differentiating between the types of state laws, examining more years, including more complete data, and also examining the extent to which a hospital’s responses to state requirements are associated with urban or rural status and financial flexibility of the institution. If these are connected, then hospitals with less financial flexibility such as rural or safety net hospitals may not respond as strongly to a state law.¹⁶

Federal Standards and Community Benefit

In the 1990s and 2000s, several interest groups expressed concern over whether the behavior of nonprofit hospitals resulted in benefits to the community sufficient to justify their tax exemption.¹⁷ Starting in 2009, the IRS sought to improve transparency and standardize CB reporting at the national level by requiring nonprofit hospitals to complete Schedule H in addition to Form 990 to keep their tax exempt nonprofit status.⁷ Schedule H collects data on expenses and revenue related to the provision of financial assistance and other CBs and provides insight into hospital activities and improves accountability. With Schedule H, the IRS did not establish any concurrent requirements or standards of provision beyond basic reporting.^18,19

This changed in 2012 with ACA’s Section 9007, which introduced new requirements intended to make nonprofit hospitals more community and public health focused. This ACA provision requires nonprofit hospitals to complete triennial community health needs assessments, to develop an implementation strategy, to establish written financial policies that include standards for billing and collection, and to place limits on charges for medical care provided to financially qualified patients.⁷ While questions remain on their precise interpretation,^6,20 the new rules clearly reflect a shift in focus away from purely charity care-based services (the demand for which decreased in states that expanded Medicaid under the ACA^21–23) toward CB investment on activities that benefit the public beyond the provision of direct medical care.^24,25

Such a shift in spending has the potential to radically impact public health—according to one estimate, if nonprofit hospitals shifted even 20% of their CB spending to population-oriented efforts, $2.2 billion in additional funds would be made available for prevention each year.²⁶ This figure is more than double the budget of the ACA’s Prevention and Public Health Fund.²⁷

State-Level Community Benefits and Relevant Legislation

Only 7 states specify a CB spending threshold minimum (Figure).⁶ Wide differences exist in the specificity and stringency of these minimums. For example, some states require nonprofit hospitals to develop a community health improvement plan (ie, implementation strategy), now a federal requirement under the ACA. However, all of those states also have reporting requirements. Only 2 states, Delaware and Illinois, have legislation on charity care spending. Twenty-six states, however, have laws that are related to certificate of need applications and require charity care spending be reported as a condition of certificate of need. Twenty-one states require nonprofit hospitals to report their CB spending activities to a state entity without some kind of minimum threshold requirement (Figure).²⁸ The lack of systematic accountability creates variability in how nonprofit hospitals spend CB dollars and motivated the introduction of additional reporting requirements.^10,29

FIGURE.

State Community Benefit Laws, 2013

Methods

Hospital community benefit data

Nonprofit hospital CB spending and expense data were collected from the IRS Form 990 Schedule H for calendar years 2009–2015. We analyzed 16 315 tax records from 2421 nonprofit hospital organizations. Of those, 2195 hospital organizations submitted 11 671 forms for unique hospital facilities. The remaining 226 hospital organizations submitted 4418 forms that reported activities taking place at multiple facilities within the organizational entity. The Form 990 and Schedule H data came from the IRS statistics of income data. Schedule H contains all information on nonprofit hospital CB spending and expenses by category (Table 1).¹⁹

TABLE 1.

Hospital Community Benefit Categories in IRS Form 990 Schedule H

Spending Type	Schedule H Categories	Description
Financial assistance at cost	Financial assistance at cost	Spending for direct patient care covered by financial assistance policies
Medicaid and other means-tested government programs	Medicaid losses Other means-tested government programs	Unreimbursed expenses for Medicaid charges Unreimbursed expenses for charges under other government means-tested programs
Flexible community benefit spending	Community health improvement services and community benefit operations	Hospital-subsidized activities and programs to support community health improvement (including community health needs assessments, community planning, etc)
	Health professions education	Medical education and certification programs made available to professionals not directly employed by the hospital
	Subsidized health services	Medical services that are needed by the community but not financially beneficial to the hospital (eg, neonatal intensive care units)
	Research	Both internally funded research conducted by the hospital and costs related to research funded by another entity
	Cash and in-kind contributions for community benefit	Hospital’s support of other community benefit activities
	Community-building activities	While not part of the community benefit table outlined by the IRS, these services are often difficult to distinguish from traditional community benefits; these are activities designed to support the community’s infrastructure for public health activities and we believe could be reasonably argued to be community benefits. Nowhere does the IRS strictly define community benefits as those items in Table 7 of the Schedule H.

Abbreviation: IRS, internal revenue service.

The dependent variable in our analysis is net CB spending as a proportion of total operating expenses. We constructed 4 dependent variables: (1) overall CB, (2) financial assistance provided at cost, (3) Medicaid and other government means–tested programs, and (4) “flexible” spending (eg, those focused on public and population health, rather than direct care delivery) (see Table 1 for details). Our designation of flexible spending is intended to capture CB spending categories that are distinct from traditional charity care and represent spending where the hospital may have more discretion.

Net CB spending, defined as gross spending less revenue, is used in this analysis. We opted to use the net amount because offsetting revenue should be considered when justifying a tax exemption. For the same reason, if revenue from CB activities exceeded spending on CBs, net spending was recoded to equal zero. We report spending per $1000 of hospital total operating expenses throughout.

Multihospital systems with a group exemption submit a consolidated Form 990 to the IRS, which includes multiple hospital facilities across states or in areas with vastly different local characteristics. To perform useful local-level analyses, data from hospital systems were mapped to individual facilities licensed by Centers for Medicare & Medicaid Services. Using Centers for Medicare & Medicaid Services Hospital Cost Report Information System data for 2009 to 2015, the IRS data were allocated proportionally to hospital facilities using their operating expenses.

This allocation approach requires an assumption be made that relative operating expenses are representative of CB spending. To test the validity of these assumptions, we performed the same allocation process using gross Medicaid expenses that are available in both the IRS Form 990 and the Medicare cost report data. We found that our estimated value and the true value correlated well among short-term acute care hospitals (r = 0.77). In addition, we ran a sensitivity analysis that excluded hospitals’ systems that filed a consolidated report, and the results were consistent with our main models (see Supplemental Digital Content Appendix A7, available at http://links.lww.com/JPHMP/A536). Based on these results, we have confidence in our allocation approach and have used it previously.³⁰

Conceptual model and covariates

Our primary independent variables were state laws on CB reporting or minimum thresholds and were based on The Hilltop Institute’s 2013 “Community Benefit State Law Profiles.”²⁸ We confirmed The Hilltop Institute’s categorization and identified the laws’ years of implementation by examining relevant codes and statutes.

We categorized states as having (a) no CB law, (b) a nonpublic reporting law, (c) a public reporting law, and/or (d) a CB spending threshold law (Figure). Seven states have exceptions to these laws based on number of beds, certificate of need laws, county population, or location. Hospitals in these states were coded in accordance with those exceptions.

The presence of a state reporting requirement indicates that the state requires nonprofit hospitals to report some aspect of CB to a regulatory entity, or that nonprofit hospitals have entered a voluntary agreement to do so.²⁸ The presence of a state threshold requirement indicates that nonprofit hospitals in that state are required to meet some form of minimum spending. We considered states to have “any” CB law if they had any type of reporting or threshold requirements. During the time period of our analysis, there were no changes in state CB laws.

If nonprofit hospitals’ decisions are driven at least in part by their charitable mission, we hypothesized that those with more funding available would spend that money on their community. To control for this, we included 3 “financial flexibility” variables in our model: the amount of available unrestricted grants as a proportion of total expenses, the degree of Medicaid dependency, and the rural status of hospitals. We used unrestricted grants because restricted grants may be restricted to CBs spending (eg, charity care) and thus encompassed in our dependent variable.

High levels of unrestricted grants and low proportions of Medicaid patients generally imply greater financial freedom. Hospitals were considered Medicaid-dependent if they were in the 75th percentile of Medicaid revenue relative to total expenses prior to 2009 in their hospital referral region.

Rural hospitals may have less financial flexibility than their urban counterparts depending on their local market. Rural hospitals are also distinct in many other ways and frequently have fewer proximate competitors, may be the only local source of CB funding, and see patient populations with different needs. Rurality was added to the model to control for the unique institutional and market characteristics of rural hospitals. Following previous literature, rural hospitals are defined as short-term acute care hospitals that are either located in a ZIP code with a rural-urban commuting area designation of 4 or greater, an area that is not a metropolitan service area, or designated as a critical access hospital.³¹ We also estimated models stratified by urban and rural status.

The other controls for hospital characteristics include membership in a multifacility health system that submitted a consolidated Form 990; system ownership and teaching status defined by Council of Teaching Hospital membership, both collected from the American Hospital Association’s Annual Survey of Hospitals; hospital size as measured by certified number of beds at the facility; and the number of unprofitable and profitable services provided drawn from the provider of service data set. Larger hospitals may behave differently, for example, they may be able to leverage economies of scale or provide more specialized services. Teaching status and profitable/unprofitable services provided were used to account for hospitals that are more likely to serve sicker and more complex patients. These variables control for the greater intensity of direct medical care at these facilities and their resulting higher per patient cost of care.³² Hospital variables were updated yearly in our data set.

County health uninsurance rates and median income from the US Census³³ as well as unemployment rates from the Bureau of Labor Statistics³⁴ were included in the model to control for local uncompensated care demand and economic conditions. These measures are hypothesized to be negatively associated with hospital profits and the amount and type of CB spending. Finally, we included Medicaid expansion state, post-Medicaid expansion, and year fixed effects to account for time-varying factors associated with CB spending patterns (eg, ACA provision, change in IRS laws, inflation, and the insurance coverage of dependents younger than 26 years).The summary statistics of all independent variables used in this analysis are shown in Table 2.

TABLE 2.

Covariate Summary Statistics for Acute Care Hospitals, 2009–2015^a

	Mean in States Without Community Benefit Reporting Law (N = 6555)	Mean in States With Community Benefit Reporting Law (N = 9074)	Difference	P
Percent Medicaid-dependent	24.85% (0.54%)	22.38 (0.44%)	− 2.48 (0.69%)	<.001
Unrestricted grants/total expenses	0.078 (0.025)	0.076 (0.024)	− 0.002(0.036)	.948
Hospital bed size
<25 beds	9.12% (0.36%)	7.48% (0.28%)	− 1.64% (0.44%)	<.001
25–126 beds	44.74% (0.61%)	40.96% (0.52%)	− 3.78% (0.80%)	<.001
127–287 beds	21.50% (0.51%)	26.14% (0.46%)	4.65% (0.69%)	<.001
288+ beds	24.64% (0.53%)	25.41% (0.46%)	0.78% (0.70%)	.270
Percent multireporters	29.31% (0.56%)	26.63% (0.46%)	− 2.68% (0.88%)	<.001
Percent system owned	64.88% (0.59%)	66.78% (0.49%)	1.90% (0.77%)	.013
Number of unprofitable services offered	1.2462(0.018)	1.4982(0.017)	0.2520 (0.025)	<.001
Number of profitable services offered	1.8857(0.021)	2.1957(0.018)	0.3100(0.027)	<.001
Percent teaching hospitals	6.41% (0.30%)	7.59% (0.28%)	1.17% (0.42%)	.0050
Local characteristics
County uninsured rate	14.81% (0.07%)	14.93% (0.06%)	0.13% (0.09%)	.158
County unemployment rate	7.50% (0.04%)	7.91% (0.03%)	0.41% (0.04%)	<.001
Rural hospital	52.62% (0.62%)	41.97% (0.52%)	− 10.65% (0.80%)	<.001

^aValues within parentheses are standard errors.

Analytic methods

We estimated the association between state CB laws and nonprofit hospital CB spending using ordinary least squares with robust state-clustered standard errors. We estimated a specification with an indicator variable for any type of law and a separate specification that included indicators for each type of law. We also estimated these models stratified by urban and rural status to examine differences by hospital location.

Since our independent variables were at the state level, we opted to use robust state-clustered standard errors to account for autocorrelation in the data. Alternatively, we could have opted to include robust clustered standard errors at the hospital level since our unit of analysis was hospital CB spending. We ran a sensitivity analysis with robust hospital-clustered standard errors (see Supplemental Digital Content Appendix A8, available at http://links.lww.com/JPHMP/A536) and our point estimates were identical with tighter confidence intervals compared with our main model. The more conservative results using state-level robust clustered standard errors gave us additional confidence in our approach.

Results

Between 2009 and 2015, hospitals spent an average of $50 billion per year in total CBs. The hospitals in our sample (short-term acute care hospitals) spent a total of $46 billion per year and averaged $20 million per hospital per year. A total of $27 billion per year went to uncompensated care and other financial assistance. The remaining $19 billion went to alternate, community-focused benefits. The average share of total operating expenses going to CBs was 8.3%. Of this, spending on uncompensated care and other financial assistance accounted for 5.7% of total expenses, whereas spending on flexible CBs accounted for the remaining 2.6%.

Multivariate results, adjusted for all covariates included in the conceptual model, are displayed in Table 3. (See full results in Supplemental Digital Content Appendices A1–A6, available at http://links.lww.com/JPHMP/A536.) In our full sample, the presence of any CB state law (ie, a reporting law and/or a total CB spending threshold law) was associated with significantly higher total and flexible CB spending and not associated with financial assistance at cost and Medicaid and other means-tested government program CB spending. Our analysis by law type suggests that public reporting laws without a threshold was a driver of increased total and flexible spending.

TABLE 3.

Association of Hospital Community Benefit Laws and Hospital Community Benefit Spending per $1000 of Total Expenses, 2009–2015^a

	All Short-Term Acute Care Hospitals (Unique Hospitals = 2733)
Variables	All Community Benefit Categories	Financial Assistance at Cost	Medicaid and Other Means-Tested Government Programs	Flexible Community Benefit Spendingb
Any state community benefit law (reference: no state community benefit law)
Any community benefit reporting law	8.419 (1.202–15.637)	−0.117 (−4.338 to 4.103)	3.721 (−2.375 to 9.818)	4.815 (0.862–8.768)
State community benefit laws (reference: no state community benefit law)
Nonpublic reporting, without threshold	5.726 (−2.684 to 14.136)	0.158 (−4.017 to 4.333)	1.778 (−5.201 to 8.758)	3.789(−1.497 to 9.075)
Nonpublic reporting, with threshold	8.217 (−5.003 to 21.437)	7.025 (2.168–11.881)	0.744 (−6.172 to 7.661)	0.448 (−5.055 to 5.951)
Public reporting, without threshold	13.492 (4.648–22.335)	−3.490 (−10.094to 3.115)	9.733 (−0.890 to 20.356)	7.248(1.132–13.364)
Public reporting, with threshold	5.936 (−8.878 to 20.749)	2.282 (−9.864 to 14.428)	− 1.490 (−7.578 to 4.598)	5.143 (−2.044 to 12.331)
Urban Hospitals (Unique Hospitals = 1446)
Any state community benefit law (reference: no state community benefit law)
Any community benefit reporting law	9.684 (2.688–16.679)	−0.833 (−5.849 to 4.182)	6.043 (−1.348 to 13.434)	4.474 (−0.945 to 9.892)
State community benefit laws (reference: no state community benefit law)
Nonpublic reporting, without threshold	8.953 (0.865–17.042)	0.287 (−4.330 to 4.905)	8.307 (0.770–15.843)	0.359 (−4.625 to 5.344)
Nonpublic reporting, with threshold	6.866 (−2.951 to 16.683)	5.550 (0.043–11.057)	0.836 (−5.428 to 7.101)	0.479 (−4.829 to 5.787)
Public reporting, without threshold	12.468 (4.988–19.948)	−4.796(−11.595 to 2.003)	9.849 (−3.556 to 23.254)	7.415 (−2.036 to 16.865)
Public reporting, with threshold	7.511 (−6.567 to 21.589)	0.773 (−10.659 to 12.205)	− 2.289(−12.528 to 7.949)	9.027 (−2.716 to 20.770)
Rural Hospitals (Unique Hospitals = 1287)
Any state community benefit law (reference: no state community benefit law)
Any community benefit reporting law	6.529 (−2.061 to 15.119)	0.631 (−3.233 to 4.495)	0.920 (−5.470 to 7.311)	4.978 (−0.245 to 10.200)
State community benefit laws (reference: no state community benefit law)
Nonpublic reporting, without threshold	4.058 (−6.206 to 14.322)	0.170 (−4.014 to 4.354)	− 3.257 (−10.052 to 3.538)	7.145 (0.047–14.244)
Nonpublic reporting, with threshold	9.302 (−7.968 to 26.571)	8.116(3.590–12.642)	0.540(−10.397 to 11.477)	0.646 (−4.510 to 5.801)
Public reporting, without threshold	12.658 (−2.351 to 27.667)	− 1.064 (−7.677 to 5.548)	8.626 (−2.693 to 19.945)	5.096 (0.111–10.082)
Public reporting, with threshold	1.741 (−11.827 to 15.309)	2.461 (−8.098 to 13.020)	− 0.196 (−5.734 to 5.342)	− 0.524 (−4.747 to 3.699)

^aEstimates reflect spending ($) per $1000 of total hospital expenses. Values within parentheses indicate 95% confidence intervals. Clustered by hospital referral region. Adjusted for variables In Table 2. Bold values indicate statistical significance.

^bCommunity-building activities, community health improvement services, cash and in-kind contributions for community benefits, research, subsidized health services, and health professions education.

Urban versus rural hospitals

Our analysis of urban and rural hospitals resulted in distinct statistical associations among urban versus rural hospitals. For urban hospitals, the spending trends were similar to the full sample but more pronounced. As with the full sample, urban hospitals exposed to any CB law had greater total CB spending. Our analysis by law type for urban hospitals suggests that this greater spending was driven by nonpublic and public reporting laws without thresholds. Interestingly, when these laws were paired with threshold laws, they were not associated with greater total CB spending at a statistically significant level.

For rural hospitals, exposure to any CB law was not associated with statistically significant associations with either total CB spending or any CB spending subcategory. Our analysis by law type for rural hospitals did reveal some spending shifts but none that resulted in greater total CB spending.

Discussion

We find that state laws influence hospital CB spending but that different types of laws influence total CB spending and spending patterns in different ways. For urban hospitals, the correlation between reporting laws and CB spending suggests that community benefit reporting laws have meaningful and positive impacts on hospital behavior. Reporting laws (both public and nonpublic) were associated with increased total CB spending when not coupled with a threshold law. It may be that accountability associated with reporting laws creates stronger incentives for spending across CB categories.

When coupled with a threshold requirement, however, reporting laws were not associated with higher total CB spending. For nonpublic reporting laws that included a threshold requirement, we found statistically significant higher spending for the CB category financial assistance at cost but again, not with total CB spending. State CB laws that include threshold requirements vary, but the majority of those threshold laws focus on financial assistance at cost. It may be that threshold requirements create a specific incentive to meet the target threshold but that they simultaneously reduce the incentive to spend more on CB categories that are not part of the threshold requirement. This finding highlights the need for caution and careful consideration of the appropriate level of CB minimum spending thresholds were such a law to be implemented more broadly.

Our findings also suggest that urban and rural nonprofit hospitals may respond to CB laws in unique ways. Our subanalysis of rural hospitals did not reveal any statistically significant relationships between CB laws and total CB spending. Instead, we found unique changes to CB spending subcategories for rural hospitals. Rural hospitals have different relationships with their communities, serve different patient populations, existing in different market conditions (frequently with little or no local competition), and may have less financial flexibility than their urban counterparts.^16,35 It is likely that some combinations of these factors contribute to rural hospitals’ unique responses to CB laws.

There are 2 key implications of these findings. The first is that policy makers wishing to increase the role played by nonprofit hospitals in impacting public and population health could implement reporting and threshold laws similar to those examined here. We provide evidence of how such laws incentivize hospitals and how hospitals might respond to those policies. The ACA introduced a federal requirement that nonprofit hospitals report to the Department of the Treasury on community health needs and plans to address these needs in addition to the reporting and transparency requirements created by the revised IRS Form 990 and Schedule H. However, this is part of an incremental approach to CB spending and it is widely anticipated that additional rules and regulations will be issued in the future.

Second, as expectations around hospital community benefits are clarified, policy makers considering CB requirements should take into consideration what is most beneficial for communities, as well as the capacity of individual hospitals. While there is a push for more of a population and public health focus, for some communities a traditional focus on the provision of charity care may be the most beneficial use of hospital CB resources. Future policies should consider the needs of the community in combination with the financial ability of local hospitals to fund such needs. In addition, future research should explore how well hospitals are aligning their CB activities with community need and to what extent hospital CB activities are having population health impacts.

There are limitations to the interpretation of our results. First, it is possible that unobserved phenomena correlated with both state laws and CB spending are responsible for the associations we see, and that should be taken into consideration when assessing the validity and generalizability of our findings. For example, some states may have a more community-focused “character,” resulting in a greater interest in both hospital behavior (leading, in turn, to reporting requirements) and hospital administrators and boards who are more likely to spend on community benefits regardless of accountability laws.

Second, our analysis includes hospitals that reported as part of larger systems and to get facility-level CB spending amounts, we allocated systems-level CB spending to individual hospitals based on operating expenses. This allocation is just an estimate and true CB spending at individual hospitals certainly varies. However, we maintain that for research purposes, our allocation approach and estimates allow for more valid and generalizable analysis across all nonprofit hospitals.

Conclusion

As public and population health continue to hold a central role in health care discussions, interest in and scrutiny of nonprofit hospitals’ role in this area will continue to grow. As large beneficiaries of charitable tax exemption, nonprofit hospitals should prove their role as guardians of the public’s health and be held accountable for providing these services to the best of their ability. Hospital CB investment in and collaboration with public health represents the greatest opportunity for health services and public health systems integration in decades. Additional research is critical to ensuring that hospital CB investment results in the greatest benefit to public health.

Supplementary Material

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (http://www.JPHMP.com).

Implications for Policy & Practice.

Policy makers wishing to increase the involvement of nonprofit hospitals in public health through community benefit activities should bear in mind the following key findings:

• First, nonprofit hospital total community benefit spending and spending patterns were associated with state laws but that these associations were influenced by the type of law, hospital characteristics, and community factors.

• Second, for hospitals that have financial flexibility, reporting laws appear to be associated with higher total CB spending, but this association no longer held when reporting laws were paired with minimum spending threshold laws.

• Third, appropriate community benefit spending for a given nonprofit hospital is a function of both the needs of that hospital’s community and the capacity of the individual hospital.