This article describes how one accountable care organization (ACO) created a risk-adjusted algorithm to evaluate current and potential candidates for skilled nursing facility partnerships.
Accountable care organizations (ACOs) continually strive to achieve the triple aim: an enhanced patient experience and improved population health at a decreased cost. One area of opportunity identified by a Philadelphia-area ACO (PAACO) to achieve this goal was to decrease skilled nursing facility (SNF) utilization and cost. The PAACO’s 2017 attributed population had utilized more than 130 facilities despite a group of preferred SNFs having been previously determined. There is currently no standard risk-adjusted algorithm for rating SNFs. The PAACO thus created a risk-adjusted algorithm that would rank the facilities based on a weighting system of key indicators of quality. Points were awarded or detracted based on expected amount thresholds; the sum of the points was multiplied by the case mix index, which served as a normalizing factor. The results revealed that the group of preferred SNFs was performing worse than all other facilities included in the analysis. Areas for opportunity were identified for the preferred group, and as a true community partner, the PAACO will work with these facilities to improve their ranking. This approach could not only decrease costs for the PAACO, but also increase the CMS star rating for these facilities.
The American Journal of Accountable Care. 2019;7(1):24-27An accountable care organization (ACO) continually strives to achieve the triple aim of enhanced patient experience and improved population health at a decreased cost. Although many ACOs face similar challenges to achieve this goal, each one has a unique set of obstacles to overcome. For one Philadelphia-area ACO (PAACO), a medical cost analysis directed attention to skilled nursing facility (SNF) utilization as a foremost area of opportunity. Postacute care is an important aspect in the continuum of care that is essential to achieving the triple aim.1 The literature indicates that 20% of Medicare patients require skilled nursing post hospital discharge, and in 2013, postacute care services accounted for 12% of Medicare annual expenditures ($60 billion).1,2
The New England Journal of Medicine published an article in 2014 that focused on postacute Medicare expenditures and discussed the importance of ACO—SNF relationships.3 A postacute care strategy is a critical part of value-based care and population health.1 A SNF admission can easily accumulate a burdensome cost, and with the implementation of the Bundled Payments for Care Improvement (BPCI) initiative, this can consume a sizable portion of the resources allocated for the 90 days of care.3,4 Due to the programmatic financial incentives, health systems are encouraged to send patients to postacute care facilities where efficient and high-quality recovery will take place.3 For an ACO, understanding postacute care partners’ performance is important, as many hospitals and providers only have a basic understanding of postacute care services and do not look beyond the facility’s length of stay (LOS); this lack of understanding can lead to the neglect of other quality metrics that play a role in patient care.1
Once partnerships are formed, goals may be aligned to benefit both the ACO and the SNF. Examples of this include quicker response times to urgent requests, provision of complete clinical information resulting in appropriate treatment options, development of evidence-based care pathways, and alignment of patient-specific care goals and discharge planning processes.3,4 A systematic review conducted in 2011 revealed that poorly coordinated care transitions contributed to more than $12 billion in healthcare expenditures.5 One of the top priorities of healthcare reform is to dissolve the silos that exist across the healthcare system and create a continuous line of communication among all facilities.5 Herein lies the vital role that an ACO can play.
One study in the Journal of the American Geriatrics Society utilized both publicly reported and self-reported metrics to rate the quality of a SNF beyond the CMS star rating.2 The authors’ findings revealed that although a facility may have a lower readmission rate with higher-acuity patients, other metrics may cause the facility to look less favorable compared with its counterparts in the absence of risk adjustment. Another study, published in Health Services Research, concluded that a statistically significant difference (P ≤.001) occurred between preferred and nonpreferred facilities in LOS (3.5 days shorter in preferred), Medicare spending ($687 lower in preferred), and readmission rates (1.4% lower in preferred) when risk adjusted.6 These results strengthen the notion that the incorporation of risk-adjustment factors, such as case mix index (CMI), will lead to better-informed decisions made by an ACO regarding which SNFs would be good candidates for partnership.2
In a densely populated market, the ability to identify facilities that can accommodate and provide high-quality care to patients with complex conditions is vital. Quality goes beyond the physical care rendered to patients; it also encompasses costs not captured in the SNF claim.7 Readmissions to hospitals not only increase the ACO’s overall costs and detract from the BPCI budget, but they may affect a hospital’s reimbursement as well. The national average readmission rate is 17%.1 Almost half of the SNFs utilized by the PAACO in 2017 exceeded this average, despite the fact that most ACOs are incentivized to bring down readmission rates. A readmission may imply that the care provided at the SNF, or hospital, was insufficient.7 Patients should be evaluated to see if home health is a more appropriate option, as it is highly cost-effective compared with a SNF stay. Some home health agencies are ramping up efforts to further the variety of nursing and therapy services in the home, more than the standard twice-a-week protocol.1 This allows the patient to receive services similar to those provided at a SNF at a much lower cost.
The PAACO’s attributed population was admitted to 134 distinct SNFs in 2017, for which claims data were examined. For comparison purposes, it should be noted that prior to the implementation of ACOs and the BPCI program, hospitals were referring patients to 58 postacute care facilities, on average.6 The PAACO’s referral patterns far exceed this amount and pinpoint the first area of opportunity for cost reduction. The intent of the PAACO’s analysis was to provide quantitative guidance for deciding with which SNFs a strategic partnership should be formed. Originally, quality key indicators used to evaluate and measure a SNF’s performance were limited to the facility’s LOS, readmissions, and costs. Although the most comprehensive analyses of SNF quality will include clinical components of care not captured in a claim, data on LOS and readmissions are critical to include in partnership discussions because both will contribute to the total medical cost of the ACO.1
The PAACO designed a risk-adjusted algorithm whereby a formulated point system for key indicators of quality created weight, which yielded a composite numerical rank for each facility. The PAACO learned that some facilities specialize in specific conditions, whereas others are highly selective in admissions criteria; therefore, the consideration and evaluation of diagnosis-related groups (DRGs) are important to counteract these differences.6 Regardless of which of these groups a facility fell into, several SNFs claimed that they admitted numerous high-acuity patients. With the intent to either validate or refute such assertions, the PAACO would apply its internally developed severity index to the analysis as a normalizing factor.
To begin this work, the PAACO needed to determine which measurement components would yield the most comprehensive model possible. First, they established a severity index at the DRG level, which incorporated the respective weights made available by CMS.8 The admitting DRG severity scores were formulated, which led to the facility’s individual CMI. This calculation displayed the acuity of the caseloads and allowed for ease of comparison across facilities; a higher CMI equated to a higher caseload acuity. From there, key indictors at the facility level were layered in: estimated versus actual LOS of caseload, estimated versus actual cost of caseload, percent of readmissions, and CMS star rating. With these outlined, facility-level claims data were assessed and calculations applied where appropriate.
The final step was to revise the ranking methodology (or weighting scale) that included the outcome of each key indicator. Points were awarded or detracted based on performance above or below the estimated LOS and cost thresholds, measured independently. Both the CMS star rating and readmission rate also yielded the potential to earn positive or negative points based on performance. Thresholds for the 4 metrics were derived from the literature.1,9 Completing the model was the multiplication of the sum of points by the CMI. This last step permitted the facility’s rank to be risk adjusted, which was embodied in a single numerical value. Risk adjusting quantifies the facility’s ability to cater to higher-acuity patients or inability to care for a lower-acuity caseload. For a hypothetical example, both SNF A and SNF B have a sum of points equal to 2. SNF A’s CMI is 1.20 and SNF B’s CMI is 0.98. Risk adjusting with the CMI leads to final scores of 2.40 and 1.96, respectively. This demonstrates that although SNF A is admitting higher-acuity patients, it is better able to control costs and LOS, and possibly delivering better patient care, thereby leading to fewer readmissions. Thus, SNF A would be considered for partnership over SNF B.
One method of delivery of these data to members of the PAACO team will be in the form of a SNF scorecard. It will call out the top-performing SNFs, highlighting those in the preferred network; additionally, it will present cost data, admissions data, and readmission rates. The data will be trended year over year and quarter over quarter to account for potential seasonality. Trending will also allow for the tracking of improvements within the aforementioned metrics. The data will be readily available for the PAACO’s Postacute Network Liaison—a registered nurse position within the ACO whose job is to manage the relationships with postacute partners—which will be used in partnership discussions along with improvement strategies.
The Figure illustrates the alarming discrepancy in estimated versus actual cost per SNF day, when comparing the PAACO’s preferred network with all other SNFs used by its attributed patients. For 2017, despite having a lower CMI than “all others,” the preferred network of SNFs exhibited a higher actual cost per day than what was expected. In a 2018 report to Congress, the standard cost per day for SNFs performing in the top margin quartile is $266; those in the bottom have a cost of $387.9 In knowing that across-the-board SNF expenditure is higher than expected, the PAACO has a definitive point from which to begin making improvements and data to aid in the establishment of clearly defined targets. The next step for the PAACO is to review these data in greater detail at the facility level and prospectively trend the scores of the SNFs over time. Trending these data quarterly will show seasonal change, and accumulative trending is expected to show improvements in agreed-upon areas of opportunity including, but not limited to, readmission rates and LOS. Additionally, as a true community partner, the PAACO will assist the preferred SNFs in improving these respective areas.
In the work by Abrams et al, top-performing SNFs were found to have an average LOS of fewer than 24 days, whereas lower-performing SNFs have an average LOS greater than 36 days for Medicare patients. This difference in days equates to about a $4000 cost difference per admission.1 The results from the analysis of the PAACO’s 2017 data were consistent with this statistic. The top 10% of SNFs had an average LOS of fewer than 22 days, whereas a majority of the SNFs in the bottom 10% had an average LOS of more than 24 days, half of which had an average LOS of more than 36 days. The difference between these 2 groups for cost per admission exceeded $10,000.
One limitation to be considered going forward is the constant change in CMS star rating. These ratings would require updating for succeeding iterations to ensure that proper points were applied or retracted. Of note, the PAACO intends to expand its DRG-specific severity scores as additional distinct cases are encountered. Not all DRGs were represented at the time of initial analysis.
Incorporating a risk-adjusted approach to ranking, then selecting preferred SNF partners is paramount for the future of ACOs. Although LOS and costs are important indicators for quality of care, they do not illustrate the full picture. Facilities with higher-acuity patients are expected to have higher costs and longer stays compared with those that have more selective criteria for accepting patients based on diagnosis/condition. Regardless of acuity level, the facility needs to be able to provide appropriate care at an appropriate cost so that patients can ideally return to the same quality of life they experienced prior to their acute care admission. Partnerships between ACOs and SNFs ultimately benefit both parties, because each will reap the resulting favorable outcomes of better care coordination and lower costs.
The authors would like to recognize and thank their colleagues from the Trinity Health System Office analytics team for their contributions, specifically Stephen Oliver and Anthony Agrusa, whose tireless efforts helped to write the backend coding. They would also like to extend a sincere thanks to those who showed their support and continued understanding in the value such work offers.Author Affiliations: Mercy Accountable Care, Mercy Health System (MEL, AMD), Conshohocken, PA.
Source of Funding: None.
Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (MEL, AMD); analysis and interpretation of data (MEL, AMD); drafting of the manuscript (MEL, AMD); critical revision of the manuscript for important intellectual content (MEL, AMD); administrative, technical, or logistic support (MEL); and supervision (MEL).
Send Correspondence to: Mark E. Lewis, MPH, Mercy Accountable Care, Mercy Health System, 1 W Elm St, Conshohocken, PA 19428. Email: Mark.Lewis@mercyhealth.org.REFERENCES
1. Abrams K, O’Rourke O, Gerhardt W. Viewing post-acute care in a new light: strategies to drive value. Deloitte website. www2.deloitte.com/content/dam/Deloitte/us/Documents/life-sciences-health-care/us-lshc-post-acute-care-innovation-report.pdf. Published 2017. Accessed October 8, 2018.
2. Lage DE, Rusinak D, Carr D, Grabowski DC, Ackerly DC. Creating a network of high-quality skilled nursing facilities: preliminary data on the postacute care quality improvement experiences of an accountable care organization. J Am Geriatr Soc. 2015;63(4):804-808. doi: 10.1111/jgs.13351.
3. Mechanic R. Post-acute care—the next frontier for controlling Medicare spending. N Engl J Med. 2014;370(8):692-694. doi: 10.1056/nejmp1315607.
4. Behery OA, Kouk S, Chen KK, et al. Skilled nursing facility partnerships may decrease 90-day costs in a total joint arthroplasty episode under the Bundled Payments for Care Improvement Initiative. J Arthroplasty. 2018;33(3):639-642. doi: 10.1016/j.arth.2017.10.013.
5. Clarke JL, Bourn S, Skoufalos A, Beck EH, Castillo DJ. An innovative approach to health care delivery for patients with chronic conditions. Popul Health Manag. 2017;20(1):23-30. doi: 10.1089/pop.2016.0076.
6. Huckfeldt PJ, Weissblum L, Escarce JJ, Karaca-Mandic P, Sood N. Do skilled nursing facilities selected to participate in preferred provider networks have higher quality and lower costs? Health Serv Res. 2018;53(6):4886-4905. doi: 10.1111/1475-6773.13027.
7. Larson A, Herbold JS. Performance of skilled nursing facilities for the Medicare population. Milliman website. milliman.com/uploadedFiles/insight/2016/2352HDP_Performance_20161212.pdf. Published December 2016. Accessed September 11, 2018.
8. Files for FY 2007 Final Rule and Correction Notice. CMS website. cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Acute-Inpatient-Files-for-Download-Items/CMS1255476.html. Published 2007. Accessed February 4, 2019.
9. Medicare Payment Advisory Commission. Skilled nursing facility services. In: Medicare Payment Advisory Commission. Report to the Congress: Medicare Payment Policy. Washington, DC: MedPAC; 2018:205-237. medpac.gov/docs/default-source/reports/mar18_medpac_ch8_sec.pdf. Accessed October 16, 2018.