Evaluation of Risk-Adjusted Home Time After Hospitalization for Heart Failure as a Potential Hospital Performance Metric

“Although the implementation of HRRP [Hospital Readmission Reduction Program] has been associated with significant reductions in readmission rates for HF [heart failure], it remains unclear whether current health policies have contributed to improvement in patients’ overall experience or quality of life. Although there has been an increasing emphasis on use of patient-oriented outcomes in evaluation of therapeutic benefits of newer HF therapies in clinical trials, the role of patient-oriented outcomes in defining hospital-level care quality for patients with HF is limited.

[..] we assessed home time after hospitalization for HF through Medicare administrative claims data and its association with currently used CMS performance metrics of 30-day risk-standardized readmission rate (RSRR) and risk-standardized mortality rate (RSMR).

[..] The primary exposure of interest was risk-adjusted 30-day home time and was defined as time spent alive and out of an acute care or a subacute care nursing or intermediate/long-term care facility. Stays at facilities were assessed through MedPAR [CMS Medicare Provider Analysis and Review], which included claims for inpatient hospitalization, long-term care facilities, and SNFs. Hospital days during the index HF-related admission were not included in the home time calculation. [..] Patients who were discharged home and remained home were assigned a home time of 30 days. Any part of a day spent after discharge from the hospital that was spent in a facility was considered as a full day for the home time calculation. Home time was modeled using generalized linear mixed models with log link and Poisson distribution using maximum likelihood estimates.

[..] A significant inverse correlation was observed between continuous measures of hospital-level 30-day home time and 30-day RSRR (r = −0.23, P < .001), 30-day RSMR (r = −0.31, P < .001), and 1-year RSRR (r = −0.35, P < .001). In the categorical analysis, a statistically significant decrease in 30-day RSRR (worst-performing, 0.23; best-performing, 0.21; P < .001) and 30-day RSMR (worst-performing, 0.09; best-performing, 0.07; P < .001) was observed across increasing categories of 30-day home time. Among long-term outcomes, 1-year RSMR also decreased in a graded fashion across increasing 30-day home time categories from 27% in Q1 to 24% in Q4 .

[..] A total of 30.6% of hospitals had a meaningful reclassification in their performance status based on 30-day home time compared with 30-day RSRR with a similar proportion of hospitals up-classified (15.6%) and down-classified (15.0%).

[..] A 30-day home time–based hospital performance metric may have other important downstream effects on care patterns among patients with HF. We observed that stays at intermediate- or long-term care facilities and SNFs were the main contributors to loss of home time during 30-day follow-up. This finding is particularly relevant considering the increasing use of post–acute care services among patients with HF and substantial variability in care quality and patient outcomes across post–acute care facilities. However, the length of stay at SNFs or intermediate- or long-term care facilities may not be directly associated with care provided at the discharging hospital. Thus, the 30-day home time performance metric will hold discharging hospitals accountable, to some extent, for the care received at the post–acute care facilities, which is consistent with the current readmission-based performance model whereby patients readmitted from post–acute care facilities within 30 days of discharge contribute to the discharging hospital’s 30-day RSRR and associated financial penalties. These observations suggest the need for a more shared approach to value-based care and associated financial incentives such that both discharging hospitals and the post–acute care facilities are held accountable for the 30-day outcomes among patients. Furthermore, the home time metric may lead to greater scrutiny into use of post–acute care facilities by discharging hospitals and encourage preferential use of home health care and higher-performing post–acute care facilities.”

Full article, Pandey A, Keshvani N, Vaughan-Sarrazin MS et al. JAMA Cardiology, 2020.10.28

[Accompanying editorial]
“Home time is the number of days a patient spends alive and out of any health care institution. It is a potentially useful metric because it incorporates mortality and the length of readmissions, along with time spent in rehabilitation and skilled nursing facilities, an important consideration for patients. In the current analysis, the authors showed that this metric can be easily measured from Medicare claims and that substantial variation existed among hospitals. Using data from more than 3000 hospitals, the median hospital-level home time after an HF-related hospitalization was 22 days (interquartile range, 8-28). As a hospital quality metric, home time was inversely correlated with 30-day risk-standardized readmission rates and mortality rates, metrics currently used in the HRRP and Medicare’s Hospital Value-Based Purchasing program, respectively. Most interestingly, however, there was substantial reclassification (31%) when considering top-performing hospitals for home time after an HF-related hospitalization compared with top-performing hospitals for 30-day HF risk-standardized readmission rates.

[..] although Pandey et al reported that reclassification existed for home time vs risk-standardized readmission rates, additional data are needed to understand which is more reflective of true quality of care. There should be a link between hospital-level process measures (eg, use of guideline-directed medical therapy for HF with reduced ejection fraction) and hospital quality. Such a link is lacking for the readmissions measure; it should be explored with home time to provide additional reassurance that this measure is more reflective of high-quality care.

[..] additional knowledge is necessary to understand how a hospital or health care system can impact home time. A key limitation of the HRRP is that the initiative did not include evidence-based strategies to reduce 30-day readmissions; consequently, many hospitals may have used strategies, such as increased use of observation status, that were ultimately harmful to patients. Use of a hospital metric that is grounded in evidence on how to improve performance should minimize unintended practices, such as patient selection and coding modification.

[..] The home time measure explicitly disincentivizes the use of post–acute care, which for some patients may be crucial in ensuring their long-term health. Although a previous study of bundled payment programs suggested that post–acute care use can be diminished without evident harm after hospitalizations for medical conditions, broad evidence is lacking for HF.

[..] Prior studies found that frailty, disability, and social risk factors were associated with readmissions, but they are not included in current risk-adjustment models. Given evidence that people living in poverty, in addition to those with frailty and disability, are more likely to need post–acute care, failing to account for these characteristics could both harm hospitals that care for patients with complex conditions and disincentivize the care they need.”

Full editorial, DeVore AD and Joynt Maddox KE. JAMA Cardiology 2020.10.28