Innovation in the Safety Net: Integrating Community Health Centers Through Accountable Care

ABSTRACT

BACKGROUND

Safety net primary care providers, including as community health centers, have long been isolated from mainstream health care providers. Current delivery system reforms such as Accountable Care Organizations (ACOs) may either reinforce the isolation of these providers or may spur new integration of safety net providers.

OBJECTIVE

This study examines the extent of community health center involvement in ACOs, as well as how and why ACOs are partnering with these safety net primary care providers.

DESIGN

Mixed methods study pairing the cross-sectional National Survey of ACOs (conducted 2012 to 2013), followed by in-depth, qualitative interviews with a subset of ACOs that include community health centers (conducted 2013).

PARTICIPANTS

One hundred and seventy-three ACOs completed the National Survey of ACOs. Executives from 18 ACOs that include health centers participated in in-depth interviews, along with leadership at eight community health centers participating in ACOs.

MAIN MEASURES

Key survey measures include ACO organizational characteristics, care management and quality improvement capabilities. Qualitative interviews used a semi-structured interview guide. Interviews were recorded and transcribed, then coded for thematic content using NVivo software.

KEY RESULTS

Overall, 28% of ACOs include a community health center (CHC). ACOs with CHCs are similar to those without CHCs in organizational structure, care management and quality improvement capabilities. Qualitative results showed two major themes. First, ACOs with CHCs typically represent new relationships or formal partnerships between CHCs and other local health care providers. Second, CHCs are considered valued partners brought into ACOs to expand primary care capacity and expertise.

CONCLUSIONS

A substantial number of ACOs include CHCs. These results suggest that rather than reinforcing segmentation of safety net providers from the broader delivery system, the ACO model may lead to the integration of safety net primary care providers.

INTRODUCTION

As Medicaid reimbursements shrink and charitable and government funding for uncompensated care dwindles, care for disadvantaged Americans is increasingly concentrated among a subset of health care providers often termed the health care safety net. Community health centers are a critical component of the safety net, providing primary care for medically underserved communities and vulnerable populations. Over 1,100 community health centers with over 8,500 locations provide care for about 22 million Americans, including one in three of those under the poverty line, one in seven Medicaid beneficiaries, and 13 million racial and ethnic minorities.¹ Community health centers that qualify for federal designation receive grant funding and enhanced reimbursement from Medicare and Medicaid, as well as other benefits. Despite serving a population that is often sicker and limited in financial resources, evidence suggests community health centers perform equally or better than private practice primary care physicians on ambulatory care measures.^2,3 Studies have found community health centers have a strong positive impact on access to primary care and quality of care for vulnerable populations.^4–7 However, community health centers still remain highly segmented from other health care providers; primary care providers at health centers have repeatedly been shown to have difficulty obtaining specialty appointments, procedures, or non-emergency hospital care for their patients.^8–11

The move toward accountable care organizations provides one venue through which safety net primary care providers such as community health centers may integrate in new ways. Accountable care organizations (ACOs) are groups of providers that assume responsibility for the total cost and quality of care for their patient population.^12,13 The ACO model incentivizes physicians to coordinate care across providers, settings, and organizations, since uncoordinated care may result in higher costs and worse quality.^12,14–16 Some have expressed concern that ACOs may further exacerbate or entrench existing segregation in the health care system.^17,18 For example, health care providers serving disproportionately high numbers of disadvantaged or low-income patients may be further isolated as mainstream providers form networks that exclude safety net providers. In contrast, it is possible that community health centers and other safety net providers may have opportunities to work in new, meaningful ways with other health care providers in their local markets, particularly if primary care providers are in high local demand by ACOs¹⁹. If this is happening, it is important to understand the rationale and strategies behind inclusion of safety net providers. Importantly, new evidence has shown that health centers that are part of integrated systems with hospitals are better able to access specialty referrals and services for their patients.²⁰ Thus, if ACOs spur new integration of health centers with other providers, this integration may have important benefits for patients.

In this paper, we explore whether the ACO model is providing a venue through which safety net primary care providers may integrate with other health care providers to provide better-coordinated care for patients. In particular, we examine to what extent ACOs include community health centers (defined as federally qualified health centers), a major component of the health care safety net in the United States. We use data from the National Survey of ACOs to quantify how many ACOs include community health centers and how these ACOs compare to ACOs without community health centers. We supplement these data with interviews of 18 ACOs that include community health centers to understand the ways in which ACOs involve community health centers, why community health centers are or are not valued as partners, and how organizational partnerships may improve clinical integration.

DATA AND METHODS

We employ a mixed methods design to learn about the involvement of community health centers in ACOs, making use of national survey data paired with in-depth qualitative interviews. Throughout this paper, we use the term “community health center” to refer specifically to federally qualified health centers. To qualify for this federal designation and receive associated benefits, a community health center must: (1) serve a medically underserved area or population; (2) provide primary care services without regard to patients’ insurance status; and (3) offer a sliding-scale payment tied to the income of uninsured patients.

Survey Data

Our survey data come from the National Survey of ACOs (NSACO), a survey of all ACOs in the United States as of August 2012, including Medicare Pioneer ACOs, Medicare Shared Savings Program 2012 participants, Medicaid ACOs, and commercial payer ACOs. Potential commercial payer ACOs were identified through a variety of sources, including public notices (e.g., press releases); participation in ACO learning collaboratives (e.g., run by Premier, the American Medical Group Association, and Brookings-Dartmouth); and responses to national surveys identifying ACOs (i.e., the National Survey of Physician Organizations and the Health Research and Education Trust Care Coordination Survey). In total, 292 potentially eligible organizations were identified as our survey population.

The survey was fielded November 2012 through May 2013. A set of screening questions determined if organizations were eligible; organizations needed to have at least one contract (Medicare, Medicaid, or commercial payer) under which they were financially responsible for both quality performance and total cost of care for a patient population to be eligible. In total, 173 eligible organizations completed the survey, resulting in a response rate of 70.0%. Analysis of non-response based on provider composition indicated our sample was representative of the population of ACOs nationally. We targeted an individual respondent at each ACO who was most knowledgeable about the survey domains, typically a member of senior or executive leadership, such as chief executive officers, executive directors, or medical directors.

For each respondent ACO, we identified provider organizations participating in the ACO through a question that asks how many of the following are part of the ACO: hospitals, medical groups, specialist groups, community health centers (federally qualified health centers or rural health clinics), nursing facilities, or other organizations. The data include 156 valid responses for this question, and our analysis is based on this sample of ACOs. We compared ACOs with and without health centers on a variety of characteristics, including number of physicians, types of contracts, payment reform experience, care management, quality improvement, and health information technology capabilities. For all results, statistical significance is based on two-tailed t-tests.

Interview Data

From June through September 2013, we conducted semi-structured, in-depth telephone interviews at 18 ACOs that include community health centers, speaking with an executive-level individual at the ACO. These ACOs were sampled from a total of 44 ACOs in the NSACO data that included at least one community health center. We also interviewed leadership at eight community health centers affiliated with those ACOs, for a total of 22 interviews. Our sample includes an oversample ACOs that are coalitions of community health centers (N = 3), and our sample also includes two ACOs led by safety net hospitals. All other interviewed ACOs included non-safety net health care providers, such as hospitals or physician practices serving largely commercial or Medicare populations. Further background research suggests that the ACOs with community health centers that we interviewed are representative of the larger NSACO sample of ACOs with community health centers.

Interviews lasted approximately one hour in length. Topics such as governance structure, communication, and clinical integration processes were covered in-depth. A portion of the interview focused specifically on the role of community health centers in the ACO. Interviews were recorded, transcribed, coded, and analyzed for themes.

SURVEY RESULTS

Organizational Composition

Community health centers (CHCs) are becoming key players in the move towards accountable care, participating in more than one-quarter of ACOs (28%). Almost all ACOs with a community health center include at least one other type of provider organization, such as a hospital or other medical group.

ACOs with community health centers are similar in structure to those without, with a few notable differences (Table 1). Compared to other ACOs, those with community health centers are equally likely to include hospitals and medical groups and slightly more likely to include nursing facilities and specialty groups. They have similar numbers of primary care clinicians (180 vs. 170) and specialty clinicians (223 vs. 268). About half of ACOs identified as integrated delivery systems, with fewer ACOs with health centers identifying compared to ACOs without health centers (46% vs. 55%), although this difference is not statistically significant.

Table 1.

Organizational Composition of ACOs, by CHC Inclusion

	Includes CHC (N = 44, 28 %)	Does not include CHC (N = 112, 72 %)	P Value on Difference*
ACO participants
Only CHCs (%)	7 %	N/A
ACO contains hospital (%)	68 %	61 %	0.388
ACO contains nursing facility (%)	34 %	18 %	0.029
ACO contains medical group (%)	89 %	90 %	0.777
ACO contains specialty group (%)	73 %	50 %	0.010
Integrated delivery system (%)	46 %	55 %	0.361
Full time equivalent (FTE) physicians (mean, [sd])
Primary care	180 (180)	170 (170)	0.755
Specialists	223 (274)	239 (322)	0.766

*p values are from two tailed t-tests for means or proportions (depending on measurement of the variable)

Are these ACOs with health centers exclusively composed of safety net providers? It appears that a very small number include only safety net providers. For example, of the 18 ACOs we interviewed with participating health centers, five consisted of predominantly safety net providers, e.g., led by safety net hospitals or coalitions of health centers; the other ACOs that involved community health centers consisted otherwise of health care providers largely operating outside of the safety net serving large proportions of commercially insured and Medicare patients.

Contracts and Payment Experience

Compared to other ACOs, ACOs with community health centers are equally likely to have Medicare ACO contracts (Table 2), more likely to have a Medicaid ACO contract (50% vs. 16%, p < 0.001), and possibly less likely to have a commercial payer ACO contract (43% vs. 55%, p = 0.173). More than half (54%) of ACOs with community health centers are currently participating in at least one risk-based contract (where the ACO is held responsible for losses in addition to savings), a larger proportion than among ACOs without health centers (38%, p = 0.092). This pattern mirrors ACOs’ experience with payment reforms: a greater proportion of ACOs with CHCs report experience with public reporting (97% vs. 82%, p = 0.019), patient-centered medical homes (95% vs. 82%, p = 0.057), and other risk-bearing contracts (88% vs. 56%, p < 0.001).

Table 2.

ACO Contracts and Payment Experience, by CHC Inclusion

	Includes CHC (N = 44, 28 %)	Does not include CHC (N = 112, 72 %)	P Value on Difference*
ACO contracts (%)
Medicaid	50 %	16 %	< 0.001
Medicare	64 %	66 %	0.775
Commercial	43 %	55 %	0.173
Multi-payer	45 %	43 %	0.770
Any current contract with risk	54 %	38 %	0.092
Receiving upfront investment (%)
Advanced Payment participant	8 %	15 %	0.255
Care management pmt from commercial payer	53 %	56 %	0.827
Capital investment from commercial payer	17 %	12 %	0.596
Securing sufficient funds: very challenging (%)	40 %	24 %	0.043
Payment reform experience (%)
Bundled payment	42 %	34 %	0.415
Patient-centered medical home	95 %	82 %	0.057
Pay for performance	89 %	90 %	0.895
Public reporting	97 %	82 %	0.019
Other risk-bearing contracts	88 %	56 %	< 0.001

*p values are from two tailed t-tests for means or proportions (depending on measurement of the variable)

ACOs with health centers were significantly more likely to report that securing sufficient funds to launch the ACO was a challenge (40% vs. 24%, p = 0.044). They reported being no more likely to receive upfront investments (17% vs. 12%), prospective care management payments (53% vs. 56%) from a commercial payer, or to be a participant in Medicare’s Advanced Payment Program (8% vs. 15%).

Capabilities

Table 3 shows the proportion of ACOs that have attained advanced capabilities on health information technology, care management, and quality improvement. Less than one-third (32%) of ACOs with community health centers report advanced information technology capabilities (including integration of outpatient and inpatient data, predictive risk assessment, and clinical decision support) compared to half of ACOs without community health centers (p = 0.047). There is no difference in primary care physicians meeting meaningful use criteria.

Table 3.

ACOS Reporting Advanced Capabilities, by CHC Inclusion: Health Information Technology (HIT), Care Management, and Quality Improvement

	Includes CHC (N = 44, 28 %)	Does not include CHC (N = 112, 72 %)	P Value on Difference*
Health information technology (%)
Advanced HIT capabilities	32 %	50 %	0.047
Meaningful use by majority of PCPs	83 %	89 %	0.283
Care management (%)
Pre-visit planning and medication management	23 %	19 %	0.671
Chronic care management processes and programs	41 %	26 %	0.069
Systems for care transitions across practice setting	18 %	19 %	0.859
Behavioral health integration into primary care	23 %	8 %	0.016
Patient involvement in care decisions, self-management	32 %	18 %	0.062
Established end-of-life care processes and protocols	20 %	21 %	0.988
Quality improvement (%)
Assessment of preventable hospital readmissions	45 %	47 %	0.882
Reduction of hospital admissions for ambulatory care sensitive conditions	41 %	44 %	0.770
Assessment of inappropriate ED use	52 %	41 %	0.218
Use of disease monitoring data	66 %	60 %	0.487
Assessment of patient care satisfaction	53 %	60 %	0.455
Clinician training in continuous QI methods	36 %	34 %	0.751
Use of ACO-wide formulary	44 %	36 %	0.353

*p values are from two tailed t-tests for means or proportions (depending on measurement of the variable)

PCP primary care provider; ED emergency department; QI quality improvement

ACOs with community health centers generally report greater care management capabilities; differences are not all statistically significant, perhaps due in part to our small sample size. ACOs with community health centers are significantly more likely to have integrated behavioral health into primary care (23% vs. 8%, p = 0.016), to report having fully developed chronic care management programs (41% vs. 26%, p = 0.069), and to report patients have high involvement in self-management (32% vs. 18%, p = 0.062). There are no statistically significant differences in quality improvement capabilities between ACOs with and without health centers, although ACOs with health centers are possibly more likely to report that they routinely assess inappropriate emergency department use (52% vs. 41%, p = 0.218).

Summary of Survey Findings

Overall, findings from survey data on ACOs suggest that community health centers are relatively common participants in ACOs, and that ACOs with health centers are generally similar to other ACOs in terms of organization, size, structure, and capabilities, although there are differences of note. ACOs with community health centers were more likely to have ACO contracts with Medicaid and less likely to have ACO contracts with commercial payers; they were also more likely to have a contract that included downside risk. In addition, ACOs with health centers were more likely to report challenges in securing funding for their initiative. In contrast to challenges, ACOs with community health centers had more experience with potentially valuable reforms, including patient centered medical home, public reporting, and risk bearing. Finally, ACOs with community health centers report less advanced health information technology capabilities, but some aspects of more advanced care management, such as chronic care management programs and patient involvement in care.

STRATEGIES OF INCORPORATION

Our interviews with ACOs and participating community health centers provided richer, in-depth information on the nature of these ACOs. Most interviewees represented ACOs with providers largely serving commercially insured or Medicare patients, not typically defined as safety net providers. In this section, we discuss the two main types of partnerships between community health centers and mainstream providers; the reasons ACOs sought to include community health centers; and how ACOs planned to work toward clinical integration of health centers’ primary care with other ACO services.

The Nature of Partnerships

In nearly all cases, community health centers were not viewed as unique partners despite generally serving a more disadvantaged patient population than the rest of an ACO’s providers. The relationships between community health centers and other participants in ACOs fell into two groups: new formal partnerships (established to pursue an ACO contract) and existing partnerships revamped or restructured for an ACO contract. The first group includes community health centers working with other health care organizations under their first formal relationship for a joint payment contract. For example, one ACO included a community hospital, private practices, and community health centers; the health centers had not previously participated in a joint payment contract with the other ACO members. In these cases, although a joint payment contract was new, the organizations involved had existing informal partnerships. For example, in the ACO described above, the community health centers had referral relationships with the hospital in the ACO. In other cases, the community health centers had worked with others in the ACO on collaborative efforts such as local or regional quality initiatives.

Some ACOs with community health centers had pre-existing formal relationships. For example, one ACO consisted of a hospital and five local health centers that operated under the hospital’s license. Another ACO was repurposing a physician-hospital organization to become an ACO, including rewriting bylaws and participation agreements to include standards around quality performance; four community health centers participating in the ACO were part of the organization and participating in this process. In these cases, the ACO contract often was seen as opportunity to improve existing partnerships and work together more closely toward coordinated and improved care. As stated by one respondent, “[We] are working on a participating agreement…thinking about what does it really mean to be part of the ACO and part of the PHO [physician hospital organization], and what are the expectations that we have for you as a practice for quality, for cost, for patient experience?”

Primary Care Supply and Expertise

Across ACOs interviewed, community health centers were being brought into ACOs largely to strengthen the ACO’s base of primary care, both to increase the number of primary care providers and to enhance primary care expertise. Many ACOs deliberately reached out to community health centers to increase the number of primary care physicians to which patients could be assigned. For example, one ACO interviewed was led by a large, integrated hospital system. “[In] two of our markets we don’t have any primary care physicians that are employed really in any significant numbers by the hospitals…. So in each region we have worked with the leadership in the region…and said, who are your key physician partners and how do we get them to the table in a single effort to manage a population? We’ve offered to have them come in and join us in our [Medicare Shared Savings Program] and share savings with them.” In one markets, the independent practice identified was a community health center, which subsequently joined with the hospital system under a Medicare Shared Savings Program ACO arrangement.

In addition to capacity, ACOs sought primary care expertise from community health center physicians. At one ACO, staff members from two participating community health centers were brought onto the Primary Care Transformation Committee; one became a physician champion for the ACO’s patient-centered medical home, providing mentorship to other practices working toward certification. Another ACO reached out to the health centers to learn from their experience with high-cost, high-need patients—a major focus for nearly all ACOs. As the respondent described, “On the evidence-based medicine committee, one of the key representatives [is] from our largest federally qualified health center (FQHC)…. He is very valuable because the FQHCs have been charged with and have tried to work with evidence-based guidelines for a great many years. So expertise like that is really valuable.” This ACO’s care coordination team was also working closely with one of the ACO’s health centers to learn how the health center has decreased hospitalizations among high-risk patients. As part of this effort, the Medical Director of the ACO was meeting regular with the health center’s clinical leadership to replicate the care delivery model for high-risk patients across ACO practices.

Beginnings of Clinical Integration

In general, ACOs interviewed were at early stages of integrating and transforming care delivery. Many had begun efforts to better coordinate care that involved community health centers. One ACO was creating new care coordination teams that worked to better connect the ACO’s primary care practices (including four community health centers) with the ACO’s hospital. For example, this care coordination team connects hospital discharge planners with clinics’ existing care managers to help ensure appropriate follow up after hospitalization.

In other cases, clinical integration strategies were being tailored or targeted for specific disadvantaged and minority patient populations served by health centers. In one ACO, a participating community health center was located in a rural area that involved a significant drive to the city where most other ACO member organizations were located, such as a community hospital and private specialty practices. The health center served a significant population of Hispanic patients, and these patients were generally hesitant to make the drive to the nearby city (for example to see specialists) because a border patrol stop was located on the main route to the city. The health center and the ACO were working together to alleviate this problem, perhaps by asking specialists to travel to the health center periodically. Other ACOs described similarly narrow targeted strategies for homeless, refugee, and non-English speaking patients.

CONCLUSIONS

In this study, we paired survey analysis with qualitative interviews to understand the extent to which ACOs involve partnerships between community health centers and other health care providers. Our results show that ACOs are involving community health centers: 28% of ACOs nationally include a community health center among participating provider organizations. ACOs with health centers are generally similar to other ACOs in terms of organization, size, structure, and capabilities, although there are differences of note. Importantly, ACOs with community health centers had more experience with potentially valuable reforms, such as patient-centered medical home, public reporting, and risk bearing. Additionally, although ACOs with community health centers report less advanced health information technology capabilities, they are more advanced in some aspects of patient management. These differences may be related directly to community health center involvement; health centers may be bringing valuable experience in care management for high-need patients, for example. In contrast, these differences may reflect selection, as ACOs more developed in these areas pursue partnerships with community health centers.

Qualitative interviews provided insight into how and why non-safety net providers were partnering with community health centers under ACO contracts. Generally, ACOs were looking to community health centers to expand their primary care capacity and expertise. Health centers were valued partners, looked to for expertise in primary care transformation and caring for high need patients. These partnerships were often new formal partnerships or renewed, existing partnerships between non-safety net providers and community health centers, suggesting that ACOs may integrate safety net providers in new ways. Notably, ACOs that are new formal partnerships were built on existing relationships, such as referral relationships or prior collaboration on regional quality initiatives.

A moment’s pause on causality and the role of ACOs specifically in spurring integration is important. Our interviews probed heavily on the role of ACO initiatives, and the results we present on partnerships reflect ACO-specific initiatives around partnership formation that respondents often directly attributed to ACO initiatives. However, we are unable to measure the counterfactual: whether or not the partnerships between community health centers and other health care providers would have occurred in the absence of ACOs. Other reforms such as bundled payment initiatives might also encourage integration or partnerships. Additionally, broader market forces might encourage partnerships of hospitals with community health centers; for example, a hospital with many specialty services may benefit from partnering with a community health center to secure referrals of health center patients.

Our study has several strengths. We use new survey data to empirically examine the involvement of community health centers in ACOs, bringing data to bear on the question of whether ACOs are incorporating safety net providers into broader delivery systems. The use of qualitative interviews allowed us to better understand the nature of community health center involvement in ACOs, including how and why ACOs that are not largely safety net organizations are working with health centers. This study represents one of the first examining the structure and nature of ACOs, and specifically how ACOs are touching the safety net. This study provides only a first look at the involvement of community health centers in ACOs; future work is needed to examine how these partnerships develop over time, if participation in ACOs leads to meaningful clinical integration and care coordination for patients, and how these ACOs perform.

Our study also has limitations. First, our survey data represent only ACOs that began contracts prior to August 2012. It is possible that newer ACOs may be systematically different than those that began in 2012 or earlier. Second, this study covers only community health centers and specifically federally qualified health centers. Notably, our survey data does not distinguish between rural health centers and federally qualified health centers, although background research and interviews indicated that nearly all ACOs with health centers involve one or more federally qualified health centers. Other safety net providers, such as safety net hospitals, likely are facing a different set of challenges and opportunities with accountable care organizations. Our data and results are solely from the perspective of the population of ACOs. In short, although we can say a substantial proportion of ACOs include community health centers, we cannot speak to whether a substantial proportion of community health centers are participating in ACOs, and why some health centers are involved while others are not. Finally, our work thus far has only examined the extent of formal community health center involvement in ACOs, so we cannot speak at length to questions about how patterns of care for patients are actually changing.

Our findings have several implications. First, these results suggest that ACOs may help integrate safety net organizations with other health care providers. Careful planning of state Medicaid initiatives and on-going Medicare programs may facilitate continued inclusion of health centers in ACO programs, such as ensuring (as Medicare did) that community health centers are eligible to have patients attributed to them under ACO rules and contracts. The ACO model may provide a venue for clinical integration apart from corporate consolidation. That is, our interviews suggested that the inclusion of health centers in ACOs typically did not involve an ownership model or consolidation of health care organizations; rather, health centers were being included in ACOs through partnership models, whereby the health centers and other ACO members retained their independence but were working together under an ACO contract in new partnerships. This model may be of interest to policymakers and providers looking to facilitate integration through mechanisms other than consolidation of ownership.