ABSTRACT

Objectives: Medicare Advantage (MA) plans have strong incentives to control costs, including postacute spending; however, to our knowledge, no research has examined the methods that MA plans use to control or reduce postacute costs. This study aimed to understand such MA plan efforts and the possible unintended consequences.

Study Design: A multiple case study method was used.

Methods: We conducted 154 interviews with administrative and clinical staff working in 10 MA plans, 16 hospitals, and 25 skilled nursing facilities (SNFs) in 8 geographically diverse markets across the United States.

Results: Participants discussed how MA plans attempted to reduce postacute care spending by controlling the SNF to which patients are discharged and SNF length of stay (LOS). Plans typically influenced SNF selection by providing patients with a list of facilities in which their care would be covered. To influence LOS, MA plans most commonly authorized patient stays in SNFs for a certain number of days and required that SNFs adhere to this limitation, but they did not provide guidance or assistance in ensuring that the LOS goals were met. Hospital and SNF responses to the largely authorization-based system were frequently negative, and participants expressed concerns about potential unintended consequences.

Conclusions: In their interactions with hospitals and SNFs, MA plans attempted to influence the choice of SNF and LOS to control postacute spending. However, exerting too much influence over hospitals and SNFs, as these results seem to indicate, may have the negative consequences of delayed hospital discharge and SNFs’ avoidance of burdensome plans.

Am J Manag Care. 2018;24(12):e386-e392Takeaway Points

We examined the methods that Medicare Advantage (MA) plans use to control or reduce postacute spending, as well as their possible unintended consequences. Plans attempted to reduce spending by controlling the skilled nursing facility (SNF) to which patients are discharged and the SNF length of stay:

• Some plans used a more hands-on approach by engaging with SNFs and patients.
• Most plans used authorization processes.
• SNFs and hospitals reported negative consequences of these authorization processes, including longer hospital lengths of stay and SNF avoidance of some plans.
• These negative consequences could result in MA patients being sent to lower-quality SNFs.

Medicare Advantage (MA) offers Medicare beneficiaries the option of receiving healthcare benefits through private insurance plans rather than through traditional fee-for-service Medicare. In 2017, MA beneficiaries made up 33% of the Medicare population, and Medicare’s capitated payments to MA plans comprised 30% of total Medicare spending.¹ Following hospitalizations, one-fifth of Medicare beneficiaries, both fee-for-service and MA, are discharged to skilled nursing facilities (SNFs) for postacute care (PAC).² The growth in SNF utilization and spending³ has placed pressure on payers to identify effective strategies to reduce postacute spending.⁴

Because MA plans receive capitated payments in exchange for bearing the risk of providing Medicare-covered services, plans may attempt to control PAC spending by requiring prior authorization for each PAC stay and for specific length of stay (LOS) durations or by limiting the network of SNFs in which patients’ care is covered.² Health plans employ prior authorization to determine whether they will pay for a SNF stay and, if so, the duration of the stay that will be covered. Restricting the network of SNFs may reduce administrative costs associated with working with a larger number of SNFs and preferentially direct patients to SNFs with more efficient practice patterns. Additionally, because MA patients can go directly to a SNF without an acute stay, plans may be further motivated to narrow their SNF networks. These cost-containment strategies are unavailable in traditional Medicare because the program does not selectively contract with providers and does not use prior authorization for SNF care beyond requiring a 3-day qualifying hospital stay.

Prior research comparing PAC in traditional Medicare versus MA has reported that MA patients use less PAC, have a shorter LOS, experience fewer readmissions, and are more likely to be discharged to the community.⁵ However, MA patients are more likely to receive care in low-quality SNFs and have high rates of switching to traditional Medicare following a SNF stay.^6,7 Of note, there is little empirical study of the specific strategies that MA plans use to reduce PAC spending and manage/coordinate care for their enrollees. This lack of prior research raises a critical gap for 3 reasons. First, there is a pressing need to identify effective approaches, including those used by MA, to improve the value of PAC.⁸ Second, federal policy has stimulated enrollment in MA under the theory that capitated payments to private plans will improve quality and lower costs, but the specific strategies that these plans use to achieve these objectives are unknown. Third, MA strategies to reduce the use of PAC may have adverse unintended consequences, and PAC providers that contract with MA plans and serve MA patients may be well positioned to identify and report these consequences.

The present study aimed to explore these mechanisms and potential consequences and sought to describe perceptions from plans, hospitals, and SNFs. Whereas quantitative analysis of secondary data may shed light on patient outcomes, LOS, or other utilization trends, these data are often out-of-date and do not give information on actual mechanisms at work. This study required a qualitative approach, as such data can help uncover and examine these dynamics in detail and lead to the development of further hypotheses that can be tested in future quantitative work.

METHODS

Design and Sample

We conducted a qualitative study of 154 participants from 10 MA plans, 16 hospitals, and 25 SNFs in 8 markets across the country. We selected markets that varied based on region of the country, county size, MA penetration rates, and the absence or presence of accountable care organizations. For further information on market selection, see McHugh et al.⁹

Procedures

We first recruited the 1 or 2 largest MA plans in each of the 8 markets, then recruited from each of those markets 1 hospital with a low readmission rate and 1 with a higher rate. Using Medicare claims data, we then selected at least 3 SNFs to which the 2 hospitals discharged patients. During in-person facility visits, we conducted 154 interviews, representing 10 MA plans, 16 hospitals, and 25 SNFs. We interviewed the chief medical officer (CMO) and a care manager for each MA plan; the vice president of strategy, the CMO, a discharge planner, and a hospitalist in each hospital; and the administrator, director of nursing, and admissions coordinator, among other staff, for each SNF. These interviews were designed, in part, to understand relationships among MA plans, hospitals, and SNFs. Participants were asked about hospital discharge planning, SNF placement and LOS, and the role that different actors, including MA plans, play in placement and LOS decisions. Sample questions from different interview participant roles are included in Table 1. These interviews took place in participants’ offices or, in the case of many of the MA plan interviews, on the phone, and they lasted approximately 40 minutes each. All interviews were audio recorded and transcribed for analysis.

Analysis

Interviews were qualitatively analyzed to identify overarching themes and patterns of responses.^10-13 First, we developed a preliminary coding scheme based on the questions included in our interview protocols. We then adjusted the scheme in an iterative fashion to add codes and refine code definitions; additional codes were added depending on the material that emerged from the interviews. The scheme was then applied to each transcript and analyzed by members of the research team. For detailed information regarding data analysis, see Tyler et al.¹⁴

During analysis, an audit trail was kept to record ongoing team decisions, including selection and definitions of codes and discussion of emerging themes and competing interpretations.^11,15-18 Coded data were entered into the qualitative software package NVivo (QSR International Pty Ltd; Melbourne, Australia) for data management. The research protocol and associated materials were approved by Brown University’s institutional review board, and informed consent was obtained from all participants.