In this first of its kind study, we introduced in a public hospital in a Brazil a system used in the private sector to analyze resources and cost management. The total operational costs at HMVSC in 2016 were US$ 48.7 million (within the initially proposed budget of approximately US$ 50.1 million). Overall, the factors that determined heightened hospital costs were age, disease severity and medical procedure complexity.
As stressed by Kaplan and Porter [14], understanding hospitalization profiles allows for a better allocation of increasingly finite resources.
Given the financial restraints experienced in any current health care system in the world, such understanding is critical for resource planning and allocation. In its first year of operation, the HMVSC admitted oncology and transplant patients, which are mostly in the high cost group, since in this study we calculated costs associated with hospitalization pre-, during and post-transplantation, and did not consider consultation and outpatient care costs.
A study conducted in Ontario, Canada, also labeled oncology patients as non-high cost group, reporting operational costs in the diagnostic investigation of outpatients with breast cancer.[15] Omachonu et al. (2004) examined how patient characteristics and clinical indicators affected the length of stay for key DRG-related groups of Medicare patients in a teaching hospital in the United States. Using mathematical models, they found that age is associated with an increase in length of stay (LOS) for psychiatric DRGs and that men had longer LOS for three DRG-related conditions: heart failure and shock, psychiatric disorders, and rehabilitation. In our study, longer LOS increased costs in the high cost group (median: 3.9 days). In terms of age, each additional year was associated with up to 6.6% higher operating costs, consistent with Dormont et al. (2007).[16] However, unlike this last study, women and men in our study were equally represented in the higher cost group.[17]
Ferraz et al [18] found that expenditures progressively increase over the four years prior to a patient’s death. The highest average cost in our study was for patients who died during 2016. Thus, in line with Koenig et. al. [18,19], severely ill patients in our study resulted in higher relative costs.
In 2015, Sacks et al. showed that excess alcohol consumption represents 76.7% of health care costs, with an average government cost of US$ 3.5 billion per US state [18]. In line with these findings, in our cohort, the highest average cost in the MDC “alcohol/drug use and alcohol drug induced organic mental disorders” group came from three patients who were awaiting an organ in the transplant unit.
Although here we only aimed to measure costs and to identify factors that increase them, we must recognize that the success of this model also depends on identifying the direct cost of labor per health care professional, an approach that employs the micro count methodology, which we intend to use in future studies.[14]
The average length of stay of the DRG tool can be used to calculate outcome costs. This cost model remains controversial until fully evaluated, but it allowed us to verify the profile of hospitalizations, which followed HMVSC’s intended criteria.
This work was the first to use the DRG system as a discharge diagnosis in a municipal public hospital, which would normally use the ICD (International Classification of Diseases; employed by the SUS for such purposes). The goal is to enable health policy makers to develop meaningful comparisons of the relative performance of hospitals in terms of efficiency and quality.[16]