INTRODUCTION

Deep sternal wound infections (DSWI) as a result of open-heart surgery are so serious and costly that hospital services continue to strive to control and prevent these outcomes^[1,2]. The prevalence of DSWI in coronary artery bypass grafting (CABG) patients varies between 1% and 4% worldwide^[3,4]. And multiple risk factors are associated with infections in cardiac surgery, such as female sex, age, diabetes, obesity, renal failure, smoking, steroid use, and chronic obstructive pulmonary disease (COPD)^[1,4,5].

In Brazil, a study reported the total cost of CABG per patient of US$7,992.55^[6]. Prior estimates of the cost of hospitalizations after surgical infections vary widely across hospitals, states, and regions, and range from US$24,000 to US$58,000^[7-9].

Quality improvement practices were first implemented by Ernest Codman, who migrated from the technology industry to clinical practice and collaborated to improve outcomes, even with high-complexity procedures involved^[10]. Throughout the study of the Registro Paulista de Cirurgia Cardiovascular (REPLICCAR), our team participates in the Cardiac Surgery Quality Program to identify opportunities for quality improvement in cases in which high-cost and resource-intensive frequently preventable outcomes might occur. Quality interventions do not necessarily imply increased hospital costs, as it focuses on the optimization of an existing organizational process. Some examples are medical care focuses on patients, protocols based on the best available evidence, decisions made by a multidisciplinary team, real-time data to show quality improvement protocols, benefits of interventions and their impact on patients, and education leading to positive changes^[10,11].

There is a need to establish appropriate priorities between patients’ groups with an effective selection for treatment within particular characteristics, based on the risk of complications and chance of survival, rehabilitation, and acceptable quality of life. Risk scores have become an important tool in patient assessment, including factors such as age, the severity of heart disease, and co-morbidity in the type of cardiac procedure. However, most scoring systems are used to predict mortality, and further refinement to specific morbidity risk scores is necessary to predict both outcome and hospital costs^[11].

Microcosting is the more accurate method to describe economic evaluation in healthcare. It can provide the most precise approach of deriving interventional costs because it involves direct enumeration and costing of each interventional input, such as nurse or pharmacist time for the procedure, and capital inputs, such as facilities space. The process includes three stages: (1) identification of all resources involved in the provision of care (e.g., human resources, consumables/materials); (2) accurate measurement of each resource (e.g., time and motion studies); (3) valuation of the resources used^[12]. Only a few studies reported this method in cardiac surgery^[11,13-16], none of them used the microcosting approach to estimate hospital costs for DSWI as a severe and costly complication in postoperative patients of cardiac surgery.

In that view, DSWI and mediastinitis represent a preventable outcome with a resource-intensive environment. The purpose of our study was to estimate the cost of DSWI and mediastinitis in a sample of isolated CABG patients from a referenced cardiac hospital in São Paulo, Brazil.

METHODS

This is a retrospective observational cohort study using a single center for the microcosting analysis of patients with DSWI and without complications after isolated CABG as the first cardiac procedure. Data were obtained from the REPLICCAR II database, which was a multicentric cohort study performed by voluntary participant hospitals between August 2017 and June 2019. The variables included in REPLICCAR II were defined using the Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD) collection tool (version 2.9 - 2017). Approximately 760 variables were collected preoperatively, intraoperatively, and postoperatively, and included risk factors, clinical and laboratory characteristics, and complications of surgery. The data were collected using a secure web application for building and managing online surveys and databases, the REDCap platform (Research Electronic Data Capture, https://www.project-redcap.org/).

The Comissão de Ética para Análise de Projetos de Pesquisa (ethical committee board) approved the study under the protocol number CAPPesq: 2.507.078 and received funding from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) (CNPq PPSUS FAPESP 2016/15163-0); patient consent was not required.

RESULTS

A total of 1,120 CABGs were performed in the hospital between 2017 and 2019. The DSWI prevalence during the period was 4.7% (n=53), and prevalence of mediastinitis/osteomyelitis was 1.4% (n=16). The STS risk score for mortality was on average 1.14% (standard deviation [SD]=0.9) in the DSWI group and 1.07% (SD=0.8) for patients without infections; and the STS model for DSWI (including mediastinitis) was 0.19% in average (SD=0.07) on the infected and 0.12% (SD=0.8) on non-infected patients. Seven patients died within 30 days (10.7%) - one in the same hospitalization (in-hospital mortality) and another six were re-admitted until 30 days after CABG (five with mediastinitis and one with deep infection). After PSM, 66 patients were allocated between groups. Sample characteristics after matching groups are described in Table 1.

Table 1 - Sample characteristics after propensity score matching between control and infected patients (DSWI and mediastinitis) after CABG (Instituto do Coração, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, 2020).

Figure 1 describes the hospitalization total costs between groups. As expected, deep infection increased the costs by 152%, and mediastinitis increased it by 188%. The mean cost for the control group was $6,863 (SD=4,615); for deep infections, it was $17,329 (SD=8,471); and for mediastinitis it was $19,805 (SD=13,383).

Fig. 1 - Financial impact on coronary artery bypass grafting total costs related to deep infections and mediastinitis (Instituto do Coração, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, 2020).

Download

Fig. 1 - Financial impact on coronary artery bypass grafting total costs related to deep infections and mediastinitis (Instituto do Coração, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, 2020).

Figure 2 describes the microcosting analysis with the average financial impact of infections on costs for each hospital department involved with the cardiac surgery process. Hospital stays, special materials, laboratory and image exams, medications and nutrition, and multidisciplinary labor were the costliest services for infected patients and were statistically different compared to the control group. However, the cost driver identified by the multiple linear regression was the hospital stay (b=0.0001; adj R2=0.48; 95% CI .00006-.0001; P=0.000).

Fig. 2 - Average cost of deep sternal wound infections after coronary artery bypass grafting divided by departments using microcosting approach. Bars indicate the average cost (95% confidence intervals [CI] are shown on the right) (Instituto do Coração, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, 2020).

Download

Fig. 2 - Average cost of deep sternal wound infections after coronary artery bypass grafting divided by departments using microcosting approach. Bars indicate the average cost (95% confidence intervals [CI] are shown on the right) (Instituto do Coração, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, 2020).

Considering time in intensive care unit (ICU), 19.2% and 40% of patients with deep infections and mediastinitis were readmitted for intensive care. The mean LOS in ICU after readmission was 13 days in the deep infections group and 21 days in the mediastinitis group. The groups were statistically different in total postoperative duration, where the control group total LOS was 9.5 days, the DSWI was 37 days, and the mediastinitis was 59 days (P=0.000).

DISCUSSION

Little information is available on the final impact of infections after major cardiac surgery. DSWI results in significant patient morbidity and consumes considerable resources and hospital LOS^[18,19]. The main purpose of this study was to provide an accurate description of how much infections (DSWI and mediastinitis) increase costs related to open surgical treatment in CABG patients. It should come as no surprise that there is a strong positive correlation between LOS and hospital costs. What is of interest is the relative importance of each complication to the cost structure of isolated CABG patients and the hospital’s ability to create policymakers to predict and manage these circumstances^[9].

Still, previous studies confirm our findings and reinforce the need for preventive methods^[18]. Graf et al. report that the main proportion of costs related to DSWI concerns to ward care, additional interventions, and prolonged ICU care^[20]. Besides the prolonged hospital stay, mortality and morbidity rates during treatment are high. Gib et al. report that using preconditioning wounds, such as vacuum therapy, could reduce mortality by 22% compared to time-only procedures^[21].

On the other hand, Brandt et al.^[22] published a structured literature review with data from 14 countries, including Brazil. The paper explores the burden of surgical infections after CABG. However, infection rates varied considerably between settings, with infections occurring in 2.8% (the United Kingdom) to 10.4% (the Netherlands) of CABG procedures, while the costs per surgical infection varied between $8,172 (Brazil) to $54,180 (Japan). Important limitations of this analysis include uncertainty about the surveillance methods, criteria and definitions, and superficial infections.

Economic evaluation studies in surgery frequently use “top-down” or “gross-costing” approaches and, usually, are based on healthcare resource groups, which can be used to estimate the average cost per inpatient episode for groups of surgical procedures^[11]. The diagnosis-related groups (DRG) were a framework created for monitoring hospital activity and efficiency and thus to control the increasing hospital costs better. The unique DRG to which any procedure is assigned is based on disease, comorbidities, and complications, as recorded by the International Classification of Diseases, Tenth Revision (ICD-10). Treatment codes are given by the NOMESCO Classification of Surgical Procedures (NCSP); the national DRG code is automatically calculated by a computer algorithm introducing the ICD-10 codes and NCSP codes. The algorithm is constructed to allocate the most complicated patients to the lowest DRG code number^[6]. However, there are several limitations related to these methods, such as the need to compare two different surgical procedures within the same group or evaluate a modification/actualization to an existing process^[12].

Brazilian economic and health systems databases are challenging to manage costs or quality because the ICD-10 is biased and does not represent clinical characteristics. Data collection is very scarce, and most observations are summaries per region, so individual observations aren’t possible using the National Database. In this study, we opted to use the REPLICCAR II database due to all limitations in Brazilian sources.

Prediction models designed for specific outcomes consistent with real population parameters may provide more accurate information about patients and hospital resources. Developers of performance measures will also be expected to promote their measures to health systems and payers. Another tough work would be to best educate consumers about why the measures are important to increase quality. Fortunately, such measure-promotion efforts will be synergistic with registry-promotion activities. Through advocacy for wider measure adoption, models can simultaneously promote the use of their performance measures and the registries that report those measures, thereby furthering the goals of patient-centered care^[22].

The STS models, for example, are widely used by Cardiac Surgery Quality Initiatives (CSQI) programs, such as the Virginia CSQI and the American Association for Cardiac Surgery^[23]. The Virginia CSQI evaluates adherence to clinical and process metrics derived from performance measures from the STS ACSD. This voluntary consortium of 17 hospitals and 13 cardiac surgical practices in Virginia (United States of America) identified quality improvement opportunities. It tracked patient outcomes but also found options for cost containment, such as improved patient outcomes and decreased resource utilization^[24,25,26].

In 2018, the Centers for Medicare & Medicaid Services announced that the bundled payments for care improvement advanced. The bundled payments, also described as episode payment models, are designed to move toward value-based care by incentivizing providers to go above the target price for an episode, including those that arise from complications and hospital readmissions. The idea is to support quality programs that invest in practice innovation and care redesign to better coordinate and reduce expenditures while improving the quality of care^[27,28].

In addition, Brescia et al.^[27] (2020) reported that assessing tradeoffs between spending and quality is essential for success in bundled reimbursement models. Although the authors didn’t evaluate the tradeoffs, they made a retrospective observation of 33 nonfederal hospitals in Michigan (United States of America) and identified determinants of variability between hospitals.

Implementing quality programs may represent the key to success for continuous improvement results.

CONCLUSION

In summary, our results demonstrate the incremental costs of a detailed microcosting evaluation of infections on CABG patients in São Paulo, Brazil. Hospital stay was an important cost driver identified, demonstrating the importance of evaluating patients’ characteristics and managing risks for a faster, safer, and more effective discharge.