Patterns of unplanned readmissions after heart failure hospitalization: novel longitudinal perspectives from Australia and New Zealand

This article refers to ‘Frequency, trends and institutional variation in 30-day all-cause mortality and unplanned readmissions following hospitalisation for heart failure in Australia and New Zealand’ by C. Labrosciano et al., published in this issue on pages 31–40. Heart failure (HF) is a condition associated with frequent hospitalizations needed to decongest patients with intravenous diuretics, whilst closely monitoring biochemistry and thoroughly evaluating the patients' cardiac and wider health status. However, such admissions should be balanced against the inherent costs, not just for health care systems but also for the individual patient. Immobilization, associated muscle loss, infection risk and loss of self-empowerment are all reasons why length and number of hospitalizations should be kept to a minimum. It is of great interest therefore to study and monitor changes in rates of HF hospitalization, and unplanned readmissions over time. It is well known that an unplanned (re)admission to hospital is associated with an increased mortality risk, impaired functional status, and quality of life and consequently it has been introduced as a measure for monitoring treatment quality across health care institutions together with mortality1 (Figure 1). In the US, the Hospital Readmissions Reduction Program (HRRP) that was established in 2012 — as a part of the Affordable Care Acts — has taken this a step further by applying financial penalties to hospitals with 30-day readmission rates higher than expected.2 The intent was to reduce health care cost by penalizing early readmissions after hospitalization due to acute myocardial infarction, HF, pneumonia (and since 2015, chronic obstructive pulmonary disease, and elective knee and hip arthroplasty). While reduction of hospitalizations by drug or device therapy in a clinical randomized trial are likely to represent important steps forward in the treatment of patients, a deliberate reduction in the number of readmissions due to financial disincentives may not. The decision of whether or not to hospitalize a patient is complex, and often depends on many factors beyond clinical assessment of HF status, including comorbidity, patient and physician preferences, logistic and other practical issues. This decision should not be influenced by financial and structural incentives, which could lead to abstaining from hospitalization with consequent further deterioration of the patient's condition. Several US studies have shown that 30-day hospital readmission rates have declined since the implementation of HRRP,3-5 while the derived effect on 30-day mortality remains debated.4, 6, 7 However, data regarding trends in 30-day mortality and unplanned readmission rates following HF hospitalization during this period outside of the US are sparse, but needed to get a better barometer of secular trends in countries where health systems are not penalized by higher readmissions rates. In this issue of the Journal, Labrosciano et al.8 provide new findings that enrich our understanding of this area. The authors present data from 2010 through 2015 for Australia and New Zealand, two countries with highly developed health care systems that have not undergone a broad political intervention to bring down short-term readmission rates. Notably, Labrosciano et al.8 report a substantial decline in unadjusted 30-day mortality risk (12.5% to 8.1%) but a lesser decline in 30-day unplanned readmission rate (23.2% to 21.9%) for HF patients. Both trends remained significant after adjusting for temporal differences in patient characteristics. Further, when the authors examined the institutional risk-standardized mortality rate (RSMR) and readmission rate, the former ranged from 6.1% to 17.3% across institutions, and the latter from 17.7% to 27.1%. Importantly, only around one third of the unplanned readmissions were due to HF, while more than half were due to non-cardiovascular conditions. These hospitalization patterns have been described previously,9, 10 but the fact that non-cardiovascular admissions dominate the early phase after a HF hospitalization emphasizes the multi-morbid characteristics of patients with HF who are at greatest risk of death and readmissions. This is further illustrated by the higher prevalence of renal disease, chronic lung disease and diabetes along with more prior cancers, among those readmitted. Comparing the trends seen in the present article with those reported from the US, what stands out is the substantial decline in mortality seen in Australia/New Zealand which is not mirrored in the US. This decline in mortality is observed despite very modest reductions in unplanned readmissions, indicating that the readmission changes seen in the US cannot guarantee reductions in mortality. The decline in mortality after HF admission over time is likely explained by factors such as implementation of novel therapies, earlier diagnosis and intervention, and better treatment of comorbid conditions as well as increased focus on patient education. Improving survival may likely result in a more fragile population thus more susceptible to readmission, which is underscored by Labrosciano and colleagues, reporting that less than half of the early readmissions were due to cardiovascular conditions. Unplanned hospitalizations are a marker of deterioration but can also be an opportunity to evaluate and optimize therapy and eventually acknowledge terminal disease and initiate palliative actions. Therefore, hospitalization rates are multi-faceted, and their interpretation is complex. Although the present study is impressively large, some limitations have to be acknowledged. The study lacked information on left ventricular ejection fraction, natriuretic peptides, renal function, etc., limiting its value in comparing outcomes across institutions. The almost three-fold difference in 30-day mortality between the institutions with the lowest and highest RSMR may, as the authors suggest, reflect differences in clinical processes and transitional care practices and patient demographics. With more than a fourth of the population living in rural or remote areas, it is also possible that variations in access to health care and threshold for seeking medical help are relevant to RSMR variations.11 In conclusion, this important work by Labrosciano and colleagues adds a more global perspective on early outcomes after a hospitalization with HF. It shows early post-hospital mortality rates have declined substantially in Australia and New Zealand between 2010 and 2015, seemingly independent of readmission rates. This finding, in turn, underscores that the road to improved patient care and survival in HF is not necessarily paved with avoidance of hospitalizations. Conflict of interest: none declared.