Opportunities for Collaboration: Refining Post Operative Readmission Risk for Skilled Nursing Facility Patients

Post-acute and long-term care providers (both physicians and advanced practice providers) depend on well executed hospital discharge plans. This is due to concerns about patient safety as well as ensuring our ability to deliver prompt, high level medical care.^1,2 The nursing facility also relies on well communicated discharge plans to ensure necessary medications are available from the pharmacy in a timely manner, a patient’s necessary medical equipment has been delivered, and to ensure that staff education, skills and competency are vetted to meet the needs of this increasingly complex posthospital population.^3,4 As accepting facility providers, we are asked to manage high risk patients whose “acute care” episodes in the hospital are spilling into our “post acute” care environment, due to shorter inpatient stays.⁵

As providers, we may be comfortable managing post-hospital patients with a variety of needs, but, if our facility does not have the support of the pharmacy, the necessary medical equipment, nursing skills, transportation capacity and hospital staff access required to assume care of these patients, we are much less likely to be successful in our role as post-acute care providers. Even as we and our facility staff are challenged to elevate the level of care provided we are also challenged to reduce hospital readmissions of these increasingly complex patients.⁶

Hospital readmission penalties, estimated from claims data as a marker of a successful transition have been under increasing scrutiny lately.⁷ Overly zealous efforts to avoid readmissions may lead to poor outcomes, including mortality.^8,9 While readmissions are an incomplete measure of poor quality of care transitions because some hospital readmissions are necessary and unavoidable, efforts to reduce readmissions are one means for stimulating improvement in patient care.^6,10

Usual care will likely be insufficient for highly complex post-operative patients; so some method of identifying high-risk patients is important. There are many variables to consider in this population ranging from the provider’s perspective, facility’s perspective and the health system’s perspective. The goal of reducing avoidable readmissions necessitates input from all three perspectives in addition to consideration of patient level variables.^11–13

In this issue of JAMDA, Kim et al describe a straightforward readmission risk prediction score for post-operative patients discharged to skilled nursing facilities (SNFs).¹⁴ Although the study was from one academic medical system they followed patients discharged to an impressive number of nursing facilities (n=110) and readmitted to any of nine affiliated hospitals. They did appropriately exclude patients discharged to SNFs affiliated with other hospital systems and who, due to distance, were more likely to be readmitted to a different hospital system.

The score proposed by Kim et al is validated against the HOSPITAL score. The HOSPITAL score is comprised of seven elements, much of which are dependent upon access to inpatient information such as hemoglobin and sodium levels. Thus, despite being a relatively simple score, the HOSPITAL score cannot be calculated without access to the inpatient medical records.¹⁵ The HOSPITAL score is also designed to predict readmission of medical patients prior to hospital discharge regardless of discharge destination. The score from Kim et al offers an even simpler and more targeted mechanism for identifying risk of readmission.

There are a number of promising implications of Kim et al’s score which will make it useful for both inpatient surgery teams and potentially for post-acute care providers. The score can be applied to any post-operative patient, and they developed a three-tiered system to stratify which category a procedure may fall into. The score is comprised of just four components: number of hospitalizations in the previous year, hospital length of stay, non-elective surgery, and the three-tiered system of identifying the risk associated with the primary surgical service. The four components make it relatively easy for inpatient providers to calculate and implement. For the most part it agrees with previous studies of readmission risk. For example, acuity of the surgical admission (elective versus non-elective) has been demonstrated as a variable of interest for hospital readmissions in other studies.¹⁶

It is surprising that although comorbidities were considered they were not included in the final score because they were not significant in the multivariate model. Other studies have demonstrated the influence of comorbidities on readmissions from post acute care however, these studies have included both surgical and medical patients.^17–19 Surgical candidates with more comorbidities may have been removed from the surgical population going to the SNF because they did not have surgery in the first place.²⁰

The four components of Kim, et al.’s readmission risk score were derived from a number of potential variables that they tested, ranging from patient demographics to comorbidities to length of stay. Notably absent from the variables that were tested for inclusion in the final score are those which reflect the care received in the nursing facility. There are multiple facility-specific factors that may affect a patient’s outcomes from nurse staffing ratios to frequency of worsening pressure ulcers. Many of these facility-specific factors are included in the Centers for Medicare and Medicaid Services star rating system. Facilities with better star ratings may be less likely to have a patient readmitted to the hospital.²¹

Furthermore, nursing facility care is highly variable both within a region and when comparing different regions of the United States.²² For-profit facilities have been linked to poorer overall quality indicators which may also be linked to readmissions.¹⁷Longer duration in the SNF may be protective against readmissions¹⁸ and for those who are discharged home there is evidence that rapid establishment of home care may also be protective.²³ These are all post-hospital discharge variables that could affect readmission and are not included in Kim et al’s readmission risk score.

The authors note that the score has higher sensitivity (72.5% vs. 33.8%) and lower specificity (56.3% vs. 79.4%) than the HOSPITAL score.¹⁴ In other words, this score will yield more false positives, which can lead to more lower risk patients, including those who would not necessarily be readmitted to the hospital, to receive enhanced care because of this score. The question is what this enhanced care may look like, and the authors only briefly speculate on this. If the enhanced care is relatively low risk, one imagines little harm to the patient erroneously labeled as high risk due to the false positive rate. However, this may lead to dilution of resources and dilution of the enhanced care that can be provided to the “true positive” patients.

Although an individual facility provider may possess the clinical skills to care for complex post-operative patients, the nursing facility is an equal partner in executing prompt, high level medical care. Health systems also need to improve support of post-acute care partners by investing in facility communication and expanding access to specialty follow-up.²⁴ One simple but important action that the authors suggest is that inpatient providers may take is to improve communication. When a post-operative patient scores high enough, that may be used as a trigger for a “warm handover”. In addition to the standard nursing care report that is supposed to occur when patients are discharged from the hospital to a facility, the inpatient providers could communicate the post-operative course to the post-acute providers. Simply alerting the post-acute provider may trigger actions that can prevent the patient reentering the cycle of acute care however; improving communication alone will not stem readmissions.²⁵

The score developed by Kim et al can be calculated within a day of hospital admission. By calculating the readmission score while patients are still hospitalized, they can begin the process of planning for the patient to eventually return home, even with a brief SNF stay inserted prior to going home. It can be used as a criterion to engage the inpatient geriatrics consultant team which can help to reduce readmissions and care costs.²⁶ This may also aid in decision making about whether to discharge patients directly to home from the hospital or to a SNF since the criteria to determine ideal hospital discharge destination are debatable.^27–29

This study provides important clues for providers at receiving facilities to identify post operative patients with an elevated risk of acute care readmission. Evidence based care transitions are a major research need in post acute and long term care.³⁰ There have been numerous calls to action for increased hospital-facility collaborative care yet, there are few examples of how to move the needle forward on this front.^13,31 This study points the way for increased collaboration by introducing an easily calculable score that if communicated, can help facilities and providers better prepare for complex post-operative patient admissions.