Accountable Care Organizations Are Transforming the Management of Post-Acute Care through Preferred SNF Networks

Abstract

Due to high magnitude and variation in spending on post-acute care, accountable care organizations (ACOs) are focusing on transforming management of hospital discharge through relationships with preferred skilled nursing facilities (SNFs). Using a mixed-methods design, we examined survey data from 366 respondents to the National Survey of ACOs along with 16 semi-structured interviews with ACOs who performed well on cost and quality measures. Survey data revealed that over half of ACOs had no formal relationship with SNFs; however, the majority of ACO interviewees had formed preferred SNF networks. Common elements of networks included a comprehensive focus on care transitions beginning at hospital admission, embedded ACO staff across settings, solutions to support information sharing, and jointly established care protocols. Misaligned incentives, unclear regulations, and a lack of integrated health records remained challenges, yet preferred networks are beginning to transform the ACO post-acute care landscape.

INTRODUCTION

As the transition to value-based payment continues, health care organizations are transforming delivery models to coordinate services across a range of providers and settings. Toward this aim, accountable care organizations (ACOs) have been widely adopted in public and commercial markets and have demonstrated modest improvements in spending (Colla, Lewis, Kao, et al., 2016; McWilliams, Chernew, Zaslavsky, & Landon, 2013; Muhlestein & McClellan, 2016). Concurrently, post-acute care services including rehabilitation facilities, long term care hospitals, skilled nursing facilities, and home health care services have been highlighted as a driver of spending and variation in spending, making them a focus of efforts to improve efficiency (Medicare Payment Advisory Commission, June 2015). A report from the National Academy of Medicine found that post-acute care accounted for 73% of variation in total Medicare spending (Institute of Medicine, 2013). From 2001 to 2013, Medicare post-acute care expenditures for fee-for-service patients increased from $29.3 billion to $59.4 billion, with the greatest share from skilled nursing facilities (SNFs). Approximately 1.7 million beneficiaries used SNF facilities at least once in 2013, spending $298 per day for 37 days on average (Medicare Payment Advisory Commission, June 2015, March 2015). As the ACO model evolves and organizations strive to improve quality and optimize spending, innovation in post-acute care is accelerating (Ackerly & Grabowski, 2014; Mechanic, 2014). Participation in Medicare Shared Savings Program (MSSP) ACOs has been associated with reductions in post-acute spending without reducing quality, primarily through clinician efforts within hospitals and SNFs to influence care for ACO patients (McWilliams et al., 2017), while reduced SNF spending has been shown to increase ACO savings rates (Muhlestein et al., 2018).

Enhanced coordination of care between hospital and post-acute care settings presents a potentially rich opportunity for ACOs to improve cost performance. Effective communication between providers allows for more accurate care planning based on medical, functional, and cognitive status of patients (Ackerly & Grabowski, 2014), yet barriers exist including challenges with information sharing due to low interoperability of technology, and infrequent communication between acute and post-acute settings (Huckfeldt, Mehrotra, & Hussey, 2016; Kripalani et al., 2007). Additionally, relationships between ACOs and post-acute care providers are still in early stages with few ACOs including post-acute providers (Colla, Lewis, Bergquist, & Shortell, 2016), and limited case studies highlighting approaches used by ACOs to coordinate care with these providers (Lage, Rusinak, Carr, Grabowski, & Ackerly, 2015). The Center for Medicare and Medicaid Services (CMS) plans to implement the Skilled Nursing Facility Value-Based Purchasing Program (SNFVBP) in 2019 and has facilitated ACO-SNF innovation through introduction of the 3-day SNF waiver allowing Pioneer, MSSP Track 3 and Track 1+, and Next Generation ACOs to bypass the traditional 3-day hospital stay required to transition patients to SNFs. Further, the recently introduced ‘30-day hospital readmissions from SNFs’ ACO quality measure will contribute to determining the amount of shared savings an ACO is eligible to receive (Smith et al., 2015). These payment and delivery system reforms reflect a growing expectation that post-acute care providers and ACOs must take greater accountability for the movement of their patients across settings (Centers for Medicare & Medicaid Services, 2015; Mor, Intrator, Feng, & Grabowski, 2010; Pitts, 2014).

NEW CONTRIBUTION

While the magnitude and variation in post-acute spending has been highlighted as an area for potential improvement, little is known about the strategies ACOs are implementing to improve transitions and reduce spending, and factors that mediate these strategies have not been examined in detail. In this paper, we use a mixed-methods research design relying on survey data and qualitative interviews to study the approaches high-performing ACOs use to improve post-acute care, focusing on skilled nursing facilities due to the relative progress in this area. The contribution of this paper is to describe the ACO-SNF landscape, to delineate the developmental process and strategies that innovative ACOs have used to increase coordination with SNFs, and to suggest emerging trends in the post-acute care services environment.

CONCEPTUAL FRAMEWORK

The National Academy of Medicine report showed wide variation in post-acute spending, indicating that there may be opportunities for cost savings through ACOs. Research to date has identified small savings in post-acute care associated with ACO implementation, largely focused in reductions in average length of stay in skilled nursing facilities. However, little work has shown how ACOs derived these savings. Opportunities for post-acute care improvements in spending and quality are rooted in choosing clinically-appropriate discharge settings, emphasis on the coordination of care, the formation of meaningful partnerships and networks, and access to timely, relevant data (Mechanic, 2014). ACOs have the potential to affect post-acute care through: 1) proactive management of the hospital discharge process to prevent complications and readmissions, 2) increased referral to high-quality post-acute care providers, and 3) care coordination across clinicians and settings of care. ACOs are driven to make these changes in post-acute care management due to financial incentives to utilize cost-effective services and ACO quality measures related to hospitals readmissions from SNF, care coordination, and patient experience. We hypothesize that ACOs identified as high-performing have developed innovative strategies and tactics to manage post-acute care. We additionally hypothesize that they have done this by leveraging relationships with hospitals (to affect the discharge process) and skilled nursing facilities, where studies have shown early savings and prior research has shown great variation in quality and spending. Describing specific strategies, and analyzing the performance of ACOs with varying relationships with SNFs, is important to providers to diffuse innovative ideas, and to policymakers in order to spread ideas that improve quality or produce cost savings. It is also important for policymakers to understand barriers and facilitators to coordination from a regulatory perspective.

STUDY DATA AND METHODS

We examined ACO-SNF integration from the National Survey of ACOs (NSACO) and supplemented this analysis with in-depth qualitative interviews with ACOs leaders to describe the considerations, tactics, and experiences of ACOs in improving the management of post-acute care services. Previous ACO research demonstrates that service redesign requires time to yield performance (McWilliams, Hatfield, Chernew, Landon, & Schwartz, 2016; National Association of ACOs, 2016). Pairing quantitative and qualitative data elements allowed us to build a comprehensive picture of the current state of ACO post-acute care efforts. Quantitative data describe current ACO-SNF relationships while qualitative results elucidate strategies being used to improve ACO-SNF coordination.

Quantitative Data Sources and Sample

We conducted a cross-sectional analysis to describe the extent to which ACOs are formally incorporating SNFs with three waves of data from the NSACO. Each wave surveyed ACOs that were newly formed in the time period preceding the survey (Wave 1: October 2012 to May 2013; 2: October 2013 to March 2014; and 3: August 2014 to June 2015). We define an ACO as a group of providers held responsible for total cost of care and quality of performance for a defined patient population. We contacted 780 potentially eligible organizations, and of these, 397 ACOs were eligible and completed the survey. The combined response rate across the three waves of the survey was 64 percent (Frankel, 1983).

Quantitative Analyses

NSACO respondents were asked to select the highest level of engagement with each post-acute service: inclusion within the ACO, contracted outside, or no formal relationship. We evaluated differences in ACO leadership structure, types of services provided, and number of facilities and providers within the ACO by ACO level of engagement with SNFs.

Qualitative Data Sources and Sample

We conducted semi-structured telephone interviews between October 2015 and January 2016 and completed a site visit with a single ACO. To identify high-performing ACOs for qualitative interviewing, we examined several data sources including CMS and NSACO data, as well as website information and press releases to identify ACO composition, quality, and initiatives. Key characteristics such as extent of care transition systems, gross savings per beneficiary, and a composite mean quality measure (see appendix) were used to select potential interviewees. We narrowed our list of ACOs to 35 by selecting those reporting medium to high levels of care system transitions (NSACO), “yes” responses to reduction of waste by managing hospital admissions (NSACO), positive savings per beneficiary (CMS), and quality composite scores above the national mean for Medicare ACOs in their second performance year (CMS).

We finalized our list of 22 potential interviewees by cross referencing our data-driven list with ACO information collected from websites and press releases. We chose to represent a heterogeneous sample of location, size, types of existing innovations, and savings trajectories in order to analyze trends that are adaptable across patient markets, regions, and size of ACO. In order to retain flexibility to perform complementary claims analyses, the ACO candidate pool for qualitative interviews was restricted to Medicare ACOs. A de-identified table of interviewed ACOs’ characteristics, selection criteria, and interviewees are included as an appendix.

We recruited interviewees by email. Of the 22 we contacted, 16 agreed to participate. Those that declined cited a lack of bandwidth among organizational leaders, or a preference to participate in case studies where the ACO would be identified and have the ability to influence the narrative, as rationale for non-participation. The interviews were conducted with ACO chief executive officers, medical directors, population health directors, and chief strategy officers; one to four interviewees participated in each interview. Interviews lasted approximately one hour. Prior to conducting interviews, we provided our contact with an overview of interview topics and asked for assistance to ensure that the appropriate leader(s) would be identified within the organization to speak to the topics. Our questions were centered on three topics: (1) organizational characteristics, (2) strategies used to manage emergency, inpatient, and post-acute care services, and (3) challenges and barriers faced by high-performing ACOs, and possible solutions to overcome them. In March 2016, we conducted a two-day follow-up site visit with one interviewed ACO that exhibited a range of different stages of post-acute care transformation across its geographical regions, speaking with twelve ACO clinical and executive leaders across six interviews, and interviewing directors of four preferred SNFs.

Qualitative Analyses

Interview data were recorded under informed consent and transcribed by a transcription service. We developed a code list focused on the three topics outlined below and added additional codes from interviews to capture topics emerging from the data (sample codes included in parentheses): (1) ACO and environmental characteristics (e.g. Market power status, ACO PAC priority), (2) post-acute care management strategies (e.g. SNF selection, ACO staff embedded in SNF), and 3) exploring the challenges and regulatory barriers and possible ways the ACOs attempted to overcome these obstacles (e.g. Regulations, Aligning incentives). Three researchers independently coded the data using NVivo software (GK, SK, JM). The coding structure was revised after every three transcripts, reconciling each code across all transcripts. Analyses included 16 telephone interviews of high-performing ACOs, one site visit interview with the ACO post-acute care services division, and four site visit interviews with preferred SNFs. Additional interviews from the site visit were excluded as their focus was beyond the scope of the paper. Practical examples and quotes are presented below to illustrate themes.

STUDY RESULTS

ACO Characteristics by Relationship with Skilled Nursing Facilities

One-fifth of ACOs include SNFs formally within the ACO. Over half of ACOs have no formal relationship with SNFs (56 percent) and the remainder (24 percent) have contractual relationships with SNFs (Table 1). ACOs that include SNFs within their ACO are significantly more likely to include a comprehensive array of services including emergency care, hospital inpatient care, specialty care, and pediatric health. ACOs that formally include SNFs within their ACO are also most likely to include other post-acute care services such as outpatient rehabilitation, inpatient rehabilitation, and home health services. Additionally, these ACOs are more likely to include less commonly integrated services such as palliative and hospice care, behavioral health, and outpatient pharmacy services. ACOs that formally include SNFs tend to be larger, with higher counts of hospitals (mean 4.8; p< .001), medical groups (mean 63.2; p< .01), specialist groups (mean 44.4; p< .01), higher counts of primary care clinicians (mean 211.2; p < .01), and higher counts of specialist clinicians (mean 389.8; p < .001) when compared to ACOs that contract outside or have no formal relationship with SNFs.

Table 1:

Organizational Characteristics by Relationship with Skilled Nursing Facilities

	Level of Engagement with SNF
	Total (N= 366)	Within ACO (N=74)	Contracted Outside (N=88)	No Formal Relationship (N=204)
ACO Leadership
Physician Led*	54%	39%	56%	58%
Hospital Led	6%	8%	3%	6%
Jointly Led*	33%	46%	31%	28%
Coalition Led	4%	4%	2%	4%
Other‡*	5%	3%	11%	4%
Regular Care Services
Pediatric health***	57%	81%	63%	47%
Routine Specialty Care†***	64%	84%	67%	55%
Specialized care†***	18%	32%	11%	16%
Hospital inpatient care***	58%	93%	52%	48%
Emergency care***	56%	95%	51%	44%
Post-acute Care Services
Inpatient rehabilitation***	41%	80%	31%	32%
Outpatient rehabilitation***	44%	86%	34%	33%
Home health***	36%	79%	29%	24%
Other Services
Palliative/hospice***	42%	77%	38%	31%
Behavioral health***	43%	73%	32%	37%
Outpatient pharmacy***	28%	49%	28%	21%
Participant Providers
Community health center	30%	38%	35%	25%
Integrated delivery system***	53%	75%	52%	45%
No. of Facilities/Providers	Mean (SD;IQR)	Mean (SD;IQR)	Mean (SD;IQR)	Mean (SD;IQR)
Hospitals***	3.1 (6;4)	4.8 (5;4)	4.5 (9;6)	1.8 (3;3)
Nursing facilities***	1.4 (4;1)	2.7 (4;3)	3.2 (7;3)	0.21 (1;0)
Medical groups**	27.8 (105;24)	63.2 (223;34)	25.8 (44;28)	16.0 (25;20)
Specialist groups**	18.8 (75;10)	44.4 (157;19)	21.9 (42;30)	8.41 (23;3)
Community health centers	1.3 (5;1)	2.3 (10;1)	0.95 (2;1)	1.1 (4;1)
Primary care clinicians**	166.9 (196;150)	211.2 (216;205)	203.6 (203;202)	135.8 (180;130)
Specialist clinicians***	259.9 (367;320)	389.8  (443;520)	284.7 (368;365)	194.2 (316;215)

Data source: National Survey of ACOs

^*p < .05

^**p < .01

^***p < .001.

SD = standard deviation

IQR = interquartile range

^‡Other types of leadership include state, region or county led ACOs, FQHC-led ACOs, and other arrangements

^†Routine Specialty care encompasses all non-primary care visits to specialty medicine such as orthopedics; Specialized care involves non-routine procedures such as transplants.

Qualitative Results

While quantitative data provide information across a large, representative sample of ACOs, our qualitative results illuminate some of the strategies successful ACOs are using to change patient care. Across interviews, ACO leaders’ responses to questions regarding strategies used to manage post-acute care services gravitated to preferred SNF networks. Shaped by themes emerging from these responses, qualitative findings are arranged according to three categories of insights shared by ACO leaders: 1) considerations mediating the formation of preferred SNF networks; 2) strategies used by ACOs in establishing and optimizing preferred SNF networks; and 3) challenges to the success of preferred SNF networks. The second section details strategies employed by ACOs along the simplified continuum of a patient moving from hospital, to SNF, to home. Due to the infancy of preferred SNF networks and continued learning in progress, all sections present a breadth of key insights and practical tactics used by high-performing ACOs toward improved management of post-acute care services, offering other ACO administrators and policy makers a blueprint for ACO-post-acute care transformation.

Considerations Mediating the Formation of Preferred SNF Networks

Timing of and approaches to post-acute care improvement:

Leaders of high-performing ACOs interviewed recognized the value of efficient post-acute care service delivery early in their accountable care participation, yet addressed redesign of these services as an improvement priority at variable times in their evolution as an ACO. The majority of ACOs initially focused on expanding primary care services, building robust care management programs, improving analytics and integration of patient records, and reducing emergency department visits and hospitalizations. As these services and supports improved, ACOs shifted focus to post-acute care. This shift in focus included an effort to ensure patients were directed to the most appropriate setting of care based on clinical and social support needs. According to one ACO leader,

“We have a saying…that historically when patients would discharge from the hospital we just threw them over the wall, and whatever post-acute provider would catch them, and we never really looked to see what was happening on the other side of the wall…we just have to make it our business now to know how the post-acute providers are carrying out the triple aim.”

All interviewed ACO leaders reported that in 2015 they had implemented or were in the process of implementing strategies to manage the utilization of post-acute care. To improve quality, reduce readmissions, and reduce SNF length of stay (SNFs are paid on a per-diem basis), twelve ACOs had chosen to focus improvement work in SNFs on the establishment of a preferred SNF network. The four ACOs not in the process of establishing a preferred SNF network were placing a continued focus on initial priorities outlined above, yet indicated exploration of preferred networks as a subsequent strategic priority. “Preferred network” was the term used by interviewees to describe a group of high-performing SNFs selected by the ACO to receive preferential referrals and to form improvement-focused partnerships. We observed no sharing of risk and minimal sharing of reward with preferred SNFs. Simply by controlling the referral pattern for discharged patients, ACOs were able to engage SNFs in improvement processes. There was widespread recognition that improvement was contingent upon effective transitions across the entire continuum of care, prompting ACOs to simultaneously integrate efforts in SNFs with those to improve discharge planning in the hospital, and to connect efforts in both settings with the underpinnings of case management and primary care.

Market power:

ACO leaders indicated that a key factor in determining their ability to partner effectively with SNFs was market power; the real and potential referral volume that an ACO provided for the SNF. ACOs who indicated they had high market power were able to engage and influence SNFs, whereas ACOs stating they had lower market power indicated greater difficulty in forming mutually beneficial relationships. However, misaligned incentives could act as a strong barrier to partnership, even for ACOs with high market power. ACO competition, number of SNFs, quality of SNFs, and geographic spread of SNFs in a region impacted ACOs’ ability to build effective partnerships. For example, a high quality SNF in a market with high ACO competition and low SNF concentration had little incentive to engage in a preferred partnership as they would receive adequate referral volumes and revenues without modifying services to meet ACOs’ value-based partnership requirements.

Strategies Used by ACOs in Establishing and Optimizing Preferred SNF Networks

SNF selection and engagement:

The twelve ACOs that had developed preferred SNF networks utilized established criteria for inclusion. Some ACOs publicly released criteria using a request for proposals process, while others assessed SNFs against criteria internally and directly engaged target SNFs. Common criteria included but were not limited to: CMS five-star quality rating system scores, state-based ratings, SNF length of stay, hospital readmission rates, the capacity to manage high acuity patients, a 24/7 admission referral line, the ability to admit patients within a short time window, 24/7 ability to administer IVs, and a Registered Nurse on staff.

ACO leaders also considered the geographical coverage needs of patients and existing referral volumes to SNFs. ACOs covering expansive geographic regions, particularly ACOs serving rural areas, explained the challenge of establishing network coverage for patients seeking services close to home, a primary driver of post-acute setting choice (Buntin et al., 2005). In regions where SNF concentration and quality was low, a subset of these ACOs invested staff time and educational resources to work with key SNFs to improve quality. To offset challenges in providing coverage for complex groups such as behavioral health and dementia patients, a small number of ACOs targeted inclusion of specialized SNFs in their networks. Less tangibly, many ACOs placed an emphasis on shared value-based missions and patient care values.

ACO and SNF leaders unanimously stated that sustained or increased referral volumes were the foremost driver of SNF participation in preferred networks. The four SNFs interviewed during the site visit indicated additional benefits including: the ability to capitalize on ACO services such as telehealth, education and clinical support of SNF staff by ACO providers, ACO purchase of SNF medical equipment (e.g. Prothrombin Time/International Normalized Ratio (PT/INR) machines), and the ability to market themselves as a natural extension of ACO primary care and hospital services.

In-Hospital Strategies

Transition planning:

ACOs initiated the transition process from hospital to post-acute care early in the hospital stay, with half of ACOs in the early stages of developing guidelines to identify the appropriate level of post-acute care required by patients. This was an iterative process based on emerging data and many cited the lack of evidence as a barrier to widespread testing and adoption of formalized tools. Comprehensive approaches to care planning as patients transitioned into and out of SNFs were moving beyond traditional clinical considerations, incorporating factors such as behavioral health, housing and the home environment, nutrition, health literacy, and family and social supports. Table 2 provides a summary of strategies used by ACOs to improve care as patients move from hospitals, to preferred SNFs, home.

Table 2:

Strategies Supporting Care of ACO Patients in Hospitals and Preferred SNFs

Objective	Strategies Supporting Care of ACO Patients in Hospitals
Identify ACO patients, notify ACO/primary Care provider	Flags in integrated EHRs with daily reports to ACO +/− patient wristbands Admission, discharge, transfer (ADT) feeds Health information exchange (HIE) notifications Health management dashboards
Influence patient care	Evidence-based guidelines and treatment protocols Embedded ACO staff (non-ACO hospitals): hospitalists, RNs, case managers
Patient Information transfer	ACO EHR login privileges for hospital staff Hospital EHR login privileges for ACO staff Hospitalists double enter patient information into hospital and ACO EHRs
Patient steering	Patient and provider education re: benefits of preferred SNFs Information brochures highlighting preferred SNFs Embedding preferred SNFs in hospital EHR Patient choice policies
Seamless Transitions to preferred SNFs	Centralized transition teams Interdisciplinary patient rounds (ACO and SNF staff) Guidelines for identifying appropriate post-acute setting Medication reconciliation Standardized discharge summaries Warm handoffs with preferred SNFs
Objective	Strategies Supporting Care of ACO Patients in Preferred SNFs
Influence patient care	Embedded ACO staff: SNFists, APNs, NPs, RNs, case managers SNF care teams (trained SNF staff in SNFs with low ACO referral volume) ACO call center Required treatment protocols and performance measures: targeted length of stay, targeted readmission rates, frequency of ambulation, frequency of bathing, ability to perform on-site lab draws, pressure ulcer prevention, pain management, and condition-specific protocols (e.g. diabetes, heart failure), communication with ACO clinical staff SNF staff education
Clinical Support (technology)	PatientPing Provider-to-provider mobile applications Telehealth stations in SNFs (connect to ACO providers)
Performance management	Monthly SNF reporting requirements Training sessions led by ACO staff Review meetings with ACO-SNF liaisons
Information transfer	ACO EHR login privileges for SNF staff SNF EHR login privileges for embedded ACO staff SNFists double enter patient information into ACO and SNF EHRs
Discharge	Discharge summaries to primary care providers Nightly tuck-ins for high-risk patients at home (phone call) Triggers for ACO PCP or case manager follow-up (call/visit within specified time window) Home care/home support programs

Data source: Qualitative interviews and site visits with high-performing ACOs.

EHR: Electronic health record; RN: registered nurse; SNFists: skilled nursing physicians; APN: advanced practice nurse; NP: nurse practitioner; PCP: primary care physician

PatientPing: A commercial platform providing real-time clinical event notifications (e.g. admissions, discharges).

Embedded ACO staff:

In non-ACO hospitals, some ACOs embedded staff such as hospitalists, registered nurses, or case managers to monitor and direct the care and transitions of ACO patients. Depending on ACO patient volumes, these staff were placed full-time or circulated between multiple facilities. Alternatively, some ACOs developed centralized care transition teams that operated outside of the hospital to bridge gaps in care by connecting ACO inpatients in non-ACO hospitals to appropriate services, inside or outside the hospital.

Patient ‘steering’:

ACOs revealed hesitation around use of the word ‘steering’ and cited the inability to directly steer (preferentially refer) patients to specific facilities for post-acute care as a major challenge.

“We are also using that information to try to help steer prior to discharge – and maybe steer is the wrong word – but to help identify the higher quality facilities in terms of, you know, readmission rates and the facilities [SNFs] that will collaborate with us. We’re trying to help the facilities [hospitals] also make that choice – why SNF and which SNF.”

Recognizing the need to preserve patient choice, multiple ACOs provided patients with promotional materials highlighting preferred facilities, including the facilities’ services, ratings, and contact information, and listing non-preferred facilities on a back page. ACOs supplemented these materials with provider education to enable conversations presenting patients with the advantages of moving to a preferred SNF (such as better coordination with their care team), and then allowed patients to make the choice of any facility. In their own hospitals, one ACO embedded preferred SNFs in the EHR referral system. Within preferred networks, a small number of ACOs attempted to steer orthopedic, cardiac, or other niche patient groups to specialized SNFs. Some ACOs had adopted patient choice policies approved by legal departments and/or CMS, with one ACO providing extensive staff education on the policy, conducting annual reviews of the process, and periodically auditing its use.

ACO leaders shared that physicians would often take the historical default pathway of sending patients to SNFs, and cited a need to leverage case management supports to assure that shifting patient discharges to settings such as home care would not increase physician workload. Provider education was focused on changing perceptions that certain patient groups always required SNF care, and on discharging patients to the appropriate setting of care, including home (with or without home care services) or alternative facility-based options.

Information transfer:

None of the ACOs interviewed had integrated their physician group and/or hospital EHRs with SNF systems. In most cases preferred SNFs were granted privileges to review and download patient information, often through embedded ACO staff or terminals put in place by the ACO. In a small number of cases information was transferred to preferred SNFs via fax or jointly developed forms. One ACO had developed a provider-to-provider phone app to connect ACO hospitalists and primary care providers to share patient information, and were in the process of integrating ACO skilled nursing physicians (SNFists) to the platform.

Transition from hospital:

For patients transitioning to SNF care, half of ACOs had established processes for warm handoffs with SNF staff. When SNF care was indicated, transition processes involved hospital and SNF providers working in collaboration. A small number of ACOs were using interdisciplinary patient rounds, in some instances including both ACO and SNF providers, as an opportunity to ensure comprehensive transition planning. Prior to discharge from hospital, an ACO SNFist or a staff member from a preferred SNF would visit the patient in-hospital. Once this visit had occurred, hospitalists were notified and proceeded to reconcile medications, ensure labs were complete, and complete the discharge summary, after which time patients were transported to the receiving SNF. Medication reconciliation was a key component of all strategies and was conducted at one or more time points including the inpatient setting, upon discharge from hospital, and following admission to a SNF, using a variety of physician and pharmacy-led approaches.

“Then the process starts with the SNFists too. We try to – from our medical management team, we alert them [SNFist] that there’s a member that’s going to be admitted to a skilled nursing facility…If they’re able to – if it’s one of our [preferred] facilities where they can do a warm handoff, they’ll come and see the patient [in hospital]. If not, then they’ll see them once they get to a skilled nursing facility. So then after all that’s done our hospitalists are made aware. They’ll do all the typical functions with the med reconciliation and the discharge summary, make sure all the labs are complete, and the patient will be transported over to a skilled nursing facility.”

In-SNF Strategies

Embedded ACO staff:

The primary mechanism ACOs used to influence care in preferred SNFs was embedded staff (ACO-funded staff embedded in non-ACO SNFs). Embedded staff included combinations of skilled nursing physicians (SNFists), advanced practice nurses (APNs), nurse practitioners (NPs), registered nurses (RNs), and case managers. In most cases, embedded ACO staff circulated between multiple preferred SNFs, holding regular hours at each and participating in rounds for ACO patients. ACO staff was consulted on care planning and discharge; however, decisions rested with SNF clinicians. ACO staff also provided on-call clinical support to minimize preventable emergency department visits and readmissions. One ACO used a centralized call center to refer SNF clinical inquiries to on-call APNs or the patient’s designated SNFist. Alternatively, in SNFs with low ACO patient volumes, ACOs identified on-site teams to care for their patients. These teams were trained in the ACO SNF protocols, processes, and aims.

Treatment protocols and performance measures:

To ensure quality care in preferred SNFs, most ACOs established a range of required performance measures and clinical protocols including but not limited to: targeted length of stay, targeted readmission rates, frequency of ambulation, frequency of bathing, ability to perform on-site lab draws, pressure ulcer prevention, pain management, and condition-specific protocols (e.g. diabetes, congestive heart failure), and communication with ACO clinical staff. As mentioned above, several ACOs employed a SNF liaison to regularly visit with SNF administrators to review performance targets, and address barriers to meeting them. Depending on the market power of the ACO, elements of protocols were imposed as a condition of participation as a preferred SNF (high market power), or negotiated with the SNF (low-medium market power).

Clinical support technology:

Many ACOs used a common commercial platform to notify ACO primary care providers or case managers when ACO patients were admitted to hospitals or SNFs (PatientPing). This was noted as especially important to monitor care for patients admitted to non-ACO SNFs and/or discharged from non-ACO hospitals. Interoperability between ACO and SNF medical records was minimal. None of the ACOs interviewed had achieved integration of ACO EHRs with SNF systems, largely due to differences in the commercial EHR brands used across settings such as Epic for hospitals and PointClickCare for SNFs. A small number of technologically advanced ACOs were working with preferred SNFs to improve communication and reported potential in emerging modules from commercially available EHR platforms. A commonly stated solution to the lack of interoperability was to provide SNFs with read-only access to ACO primary care and/or hospital EHRs. To create one comprehensive medical record for patients across systems, some ACOs required that SNFists doubly enter information into the SNF and ACO patient record. Additional technologies were emerging to support patient care in SNFs. One ACO utilized a telehealth station placed in SNFs with a high-definition camera and a monitor. SNF staff could press a button for immediate access to an ACO physician to provide clinical support for ACO patients.

Discharge from SNFs:

Discharge from preferred SNFs triggered automatic follow-up from ACO primary care staff or case managers. ACO follow-up strategies included a mixture of patient phone calls within a set time window, scheduling of follow-up visits with primary care or specialists, initiation of ‘nightly tuck-ins’ for high-risk patients, or initiation of home care services.

Challenges to the Success of Preferred SNF Networks

Building effective partnerships:

Interviewed leaders indicated a learning curve for ACOs in working effectively with SNFs due to less familiar staffing arrangements (higher staff to patient ratios in hospitals), treatment protocols, as well as regulators and reporting requirements. When individual ACO referral volumes were not critical to SNF survival, partnerships formed on the basis of mutually beneficial two-way communication and shared ownership of improvement solutions were cited as most effective. When individual ACO patient volumes were critical to SNF survival, ACOs took a more dictatorial approach to defining SNF requirements for network participation. To transmit priorities to SNF administrators, several ACOs conducted monthly meetings between ACO and SNF leaders. One ACO hired a former SNF administrator to act as a liaison, consulting on regulation and strategy. Multiple ACO leaders found the use of liaisons familiar with both SNF and ACO operations integral to facilitation of these evolving partnerships. Regarding oversight, one ACO’s leaders emphasized advantages in nesting administration of the preferred network in a portfolio with ACO-owned home care services due to the centrally managed, regionally deployed relationships essential to both services.

Alignment of incentives:

A challenge unanimously reported by ACO leaders was the lack of alignment between ACO and SNF incentives due to conflicting payment models. Apart from one ACO participating in a demonstration using a basic rate withhold for SNFs that could be earned back through performance, there were no mechanisms for ACOs to financially reward SNFs. To compound this challenge, ACO leaders voiced complaints that Medicare pays SNFs on a per diem basis, incentivizing longer stays. One ACO leader explained these challenges,

“Currently we’re working on mechanisms to decrease the length of stay in the sub-acute facilities [SNFs] as well as optimizing the visiting nurse services [home health]. Once again, this is a difficult process because the incentives are not aligned. These are private companies. They’re for-profit. And the more services they can provide, the more money they make. The more durable medical equipment they provide, the more money they make. It’s an incestuous process that really requires governmental reform. But prior to that, it requires some major motivation in order to initiate change.”

Additionally, one ACO leader identified that Medicare reimburses swing beds - hospital beds that can be used to provide acute or SNF services - in qualifying hospitals at a much higher rate than alternative facilities with skilled nursing beds available. This created conflicting incentives for hospitals operating under both risk-based and fee-for-service models in trying to move patients to lower levels of care, while also preserving revenue. ACOs experienced greatest success in working with hospitals toward reduction of readmissions due to hospital incentives to avoid readmission penalties.

The presence of multiple competing alternative payment models such as bundled payments, Medicare Advantage, and ACOs created additional complexity for organizations in prioritizing and aligning strategies. Some ACOs were responding by creating a population health strategy for all patients, while others continued to segment efforts by patient-payer group. One ACO leader explained these frustrations,

“I think one of the things that’s challenging for us [is] … there’s so many different programs. And they’re trying to make different opportunities available depending on if you’re an independent physician, if you’re a large system, if you’re a specialty provider…it feels like a lot of spaghetti being thrown on the wall. And trying to identify which of those programs are going to be most impactful for this transition is challenging. And where do we, as a system, spend our energy?”

Regulation and waivers:

Approved Pioneer ACOs were capitalizing on the 3-day SNF waiver. This waiver allowed ACOs to bypass the traditional 3-day hospital stay requirement for SNF reimbursement, and use direct-to-SNF admissions from emergency departments or primary care. Both ACOs and SNFs viewed the waiver as mutually advantageous to efficiently move patients to appropriate lower cost settings, to assist with maintaining patient volumes at preferred SNFs as length of stays decreased, and as a cost-effective mechanism to bypass unnecessary and costly utilization of the emergency department for patients with escalating care needs. ACO leaders also expressed optimism about the 3-day SNF waiver, and new telehealth expansion and post-discharge home visit waivers under the Next Generation ACO model.

ACO leaders emphasized the need for strong, timely regulatory clarity regarding Stark and antitrust regulations. For some ACOs, the establishment of preferred networks created opposition from state and national skilled nursing associations, as explained by one interviewee,

“I think the one that often comes to the top of the list is the challenges in and around the various regulations, whether that’s Stark or anti-trust or anti-competitive legislation. On one hand, we are called and I think rightly so, to develop a more clinically integrated healthcare delivery model across the country… There’s another group of regulators who are looking to not prevent that from happening, but are concerned about the potential anti-trust outcomes.”

Post-acute care data:

As described above, interviewed all ACO leaders placed strong emphasis on analytics to support ACO strategies, yet voiced challenges in obtaining timely, targeted data to improve post-acute care services. Specifically, the time lag in receiving claims data from Medicare, the lack of data to track post-acute care utilization, and a lack of data on the causes of readmissions. One ACO had begun incorporating patient-reported data from readmissions into risk prediction models. According to ACO leaders, the lack of data was compounded by the constantly shifting attribution of ACO patients. Finally, there was scepticism amongst many ACO leaders about the accuracy of the CMS SNF star ratings. In response, some ACOs were collecting in-house data to evaluate SNF performance in their region.

Sustainability of SNF operating models:

Interviewees indicated an ACO trend towards moving patients away from higher cost facility-based care, particularly by reducing length of stay in SNFs. Many ACOs were increasing focus on moving patients to home care as an alternative to SNFs, and some were narrowing home care networks in a similar manner to SNFs. SNFs shared concerns that as ACO models evolved they would be pressured to use ACO-owned or affiliated providers for post-SNF services such as home care or physical therapy, eroding partnerships built over time. All ACO leaders shared the view that telemonitoring, the 3-day SNF waiver, and emerging Next Generation ACO waivers would facilitate moving patients away from facility-based care.

Under the ACO model, patients in interviewed ACOs were being discharged from the hospital earlier resulting in increased acuity of patients admitted to SNFs. As a result, some SNFs were experiencing challenges in training and retention of staff, and also voiced an increased need to educate patients and families about differences in the level of care across SNF and hospital environments, and to change patient expectations regarding long lengths of stay. Similar to hospitals, all SNFs were emphasizing an earlier start to discharge planning beginning at admission to the SNF.

LIMITATIONS

Our study has a number of limitations. First, the National Survey of ACOs is a cross-sectional survey. The survey was administered to newly formed ACOs in three waves from 2012–2015, and therefore has no ability to measure changes in relationships with SNFs over time. Second, the NSACO relied on self-report, leaving room for interpretation. For example, respondents could categorize their level of engagement with SNF providers as ‘contracted out’ or ‘no formal relationship’. To enrich our understanding of these evolving relationships, we supplemented NSACO data with in-depth examination of ACO-SNF relationships through qualitative interviews with ACO and SNFs. A third limitation is related to our qualitative interviews. Research interviews included sixteen ACOs; however, only four SNFs. Therefore, the perspective of SNFs is under-represented in our qualitative data. A final limitation relates to the sample of ACOs interviewed. ACOs were selected across a range of characteristics related to cost, quality, and capabilities that led to their classification as high-performers, yet the ACOs were not selected based on whether or not they employed preferred SNF networks, nor their post-acute care performance.

DISCUSSION

As ACOs mature and begin to shift their focus to improving the management of post-acute care services, our interviews suggest that some high-performing ACOs have made substantial progress in building partnerships with SNFs through preferred networks. However, the National Survey of ACOs indicates that over half of ACOs have no formal relationship with SNFs signalling that widespread implementation of post-acute care improvement efforts across the country remain in their infancy.

Consistent with early evidence that concentrated hospital-to-SNF referral patterns are associated with lower readmission rates (Lage et al., 2015; Rahman, Foster, Grabowski, Zinn, & Mor, 2013) and reduced rehospitalizations (Winblad, Mor, McHugh, & Rahman, 2017), high-performing ACOs are establishing preferred SNF networks to an effort to improve coordination of care, reduce hospital readmissions, and shorten lengths of stay. Our research found common elements of preferred SNF networks include a comprehensive focus on care transitions in hospitals and SNFs, embedded ACO staff to support care across settings, workarounds and emerging technologies to support sharing of patient information, and jointly established patient care protocols designed to connect patient health and social needs with appropriate services from hospital to home.

As preferred SNF networks evolve, there remain a number of challenges and opportunities for ACOs to translate these relationships into sustained improvements in cost and quality performance. The foremost challenge for ACOs in improving coordination of care with SNFs is a current lack of payment frameworks and aligned incentives to support these efforts. In advance of the 2019 Skilled Nursing Facility Value-Based Purchasing Program (SNFVBP), ongoing pay-for-performance experiments between ACOs and SNFs such as basic rate withhold models may be a first step in solving this lack of alignment, yet further evaluation of such programs is needed. Clinical decision-making regarding the need for provision of post-acute care in SNFs also lacks a clear framework, often leading to the use of SNFs as a “safety net” for older adults (Burke et al., 2017). Additional challenges for ACOs in advancing preferred SNF networks include unclear regulations around anti-trust and patient choice, a lack of integrated health records across settings of care, and a lack of actionable post-acute care data to drive performance improvement.

Given these multi-faceted challenges, evolving strategies to manage post-acute care, including preferred networks, may take time to yield improvements. Despite advantages in coordination of care that might be achieved by financially integrating SNFs within ACOs, less than a quarter of ACOs have taken this step. Rationale for not including SNFs in ACOs may include recognition of the substantially different administrative, clinical, and regulatory competencies required to manage these facilities, and the need to establish a distributed network of multiple facilities to serve patients close to home. Additionally, the benefits of gaining patients attributed through skilled nursing physicians are unclear (McWilliams et al., 2013), as is the stability of the SNF market while referral patterns shift under payment reform. Beginning in 2017, CMS is no longer treating certain HCPCS codes in listed in physician claims with SNFs as the place of service (POS=31) as primary care services for attribution purposes (Centers for Medicare & Medicaid Services, 2017), which could further complicate ACO decisions to affiliate with SNFs.

As the post-acute care sector transforms under value-based healthcare, ACOs are implementing strategies to move some patients to lower cost settings when possible, away from facility-based care. For patients outside of the skilled nursing facility, new 3-day SNF, telehealth expansion, and post-discharge home visit waivers will combine with emerging programs and technologies to augment transformation, forcing ACOs and their post-acute care partners to find mutually beneficial, patient-centered approaches to redesign care pathways. This evolving process is supported by evidence that shorter inpatient stays are not necessarily correlated with higher readmission rates (Kaboli et al., 2012), and that patients have comparable outcomes when receiving care in different types of PAC facilities (Medicare Payment Advisory Commission, March 2015).

ACOs and their affiliated systems will need to find ways to achieve synergies by applying these new strategies across patient-payer groups. For SNFs, rising acuity, shorter lengths of stay, and a need for collaboration define a new normal. As the ACO model expands, SNFs who have elected not to participate in preferred networks may face consequences (such as decreases in volume) for their lack of involvement. SNFs participating in preferred networks may enhance clinical capabilities to care for higher acuity patients, improve cost and quality performance across patient populations, and solidify referral volumes from partner ACOs. As both ACOs and SNFs adjust to this evolving environment, there will be an increased need to align strategies with other post-acute care providers such as home care. The adaptability of preferred networks throughout this period of system transformation may be their biggest strength. Further research into the cost and quality performance of preferred SNF networks, and evaluation of programs to align ACO and SNF incentives will be paramount to optimizing post-acute care services under the ACO model.

CONCLUSION

The capacity of ACOs to improve post-acute care cost and quality performance will impact the success of the ACO model, the ability of ACOs to achieve shared savings, and ultimately patient outcomes. While enhanced management of post-acute services presents a profound opportunity for improved coordination of care, efforts including preferred SNF networks are in their infancy, and a current lack of payment mechanisms to mediate ACO-SNF partnerships is a major barrier. As increasing numbers of ACOs shift their focus to post-acute care, ACO-SNF pay-for-performance programs may elucidate near-term solutions to current challenges, and emerging waivers continue to create opportunities for innovation. ACO successes in integrating post-acute care services to create a seamless continuum of care will support the national transition to value-based healthcare, and serve as a model for engagement for other traditionally fragmented health services.