Our findings show that in 2014, in Catalonia, the results for readmission, mortality, and mean length of stay for the two HaH modalities, in spite of their heterogeneous development, are similar to those observed in previous studies throughout the world [4–6, 16]. In addition, irrespective of the modality, it was observed that comorbidity and type of hospital are closely related to readmission and mean length of stay whilst the patient’s age is strongly linked to mortality. Such information could help define more precisely the most suitable helthcare circuits, and the type of patient who could most benefit from the different HaH modalities.
Previous studies comparing HaH and CH [4, 11, 17–19] have reported that home care can have similar results to conventional hospitalization and would save both human and economic resources [20–23]. In this sense, both the results obtained and the available evidence suggest that, provided the patient’s indication permit it, HaH treatment could be a suitable, effective, and possibly efficient alternative to CH [8, 16, 21].
With respect to the volume of HaH activity in Catalonia , in spite of a progressively increasing trend of HaH utilization, its implementation rate is low as it represents approximately only 1.5% of total hospital activity . In this regard, encouraging the use of HaH could decrease the medical burden within hospitals and reduce the consumption of resources linked to patient care, thus providing benefits to all levels in the healthcare system .
Regarding the HaH modalities, it was observed that whilst admission avoidance was more frequent in less complex hospitals, early discharge was more common in more complex ones. More complex or more serious cases could be addressed to a greater extent in high-technology and high-resolution hospitals. In addition, these hospitals generally have a wide range of technology and services which results in a greater demand for healthcare. Thus, adopting an early discharge program in hospitals of greater complexity, providing the patient’s indication allows it, could result in the optimization of services that are only available in these institutions. Further studies centered on the early discharge modality in high-technology and high-resolution hospitals could help assess the optimization of such specialized services.
With respect to the factors related to the considered outcomes, in both modalities it was observed that whilst readmission and median length of stay were linked to the CCI and type of hospital, mortality was related to the patient’s age. In addition, in the admission avoidance modality, mortality was only related to comorbidity in those patients with the highest CCI scores. Nevertheless, in spite of being non-significant, the direction of the relationship in patients with a lower CCI was expected. In this manner, as reported by previous studies [5, 26], our findings suggest that comorbidity could be a particularly relevant factor when choosing this modality. Further longitudinal studies with a greater sample size could be valuable to confirm this hypothesis and thus aid better indication and improve results.
As limitations of the study, it is worth highlighting the reductionism in which the classification of HaH modalities falls. Nevertheless, this classification has been previously employed on numerous occasions when evaluating HaH [5, 6, 18, 19]. In addition, given the heterogeneity observed amongst the HaH units in Catalonia, it permits a general evaluation of the results at a populational level. Another limitation is related to the cross-sectional design of the study. It does not allow the direction of the relationship amongst the variables to be established. Nevertheless, as the analyzed indicators are results from the healthcare utilization, it would be reasonable to argue that the described direction of the relationship between variables is accurate. Finally, it should be mentioned the limitation regarding the variables included. The inclusion of others not considered related to the patient or hospital, as social support or specific treatments, could help to define more accurately what factors are associated with the indicators that we analyzed. However, we consider that the included variables allow for the adjustment of more parsimonious and easily interpretable models. In addition, these variables cover to a large extent basic aspects of both the patients and the care process, being also the analysis of these factors a basic previous step for the conducting studies in greater depth.