RESUMEN
Patients in an Intensive Care Unit (ICU) are closely and continuously monitored, and many machine learning (ML) solutions have been proposed to predict specific outcomes like death, bleeding, or organ failure. Forecasting of vital parameters is a more general approach to ML-based patient monitoring, but the literature on its feasibility and robust benchmarks of achievable accuracy are scarce. We implemented five univariate statistical models (the naïve model, the Theta method, exponential smoothing, the autoregressive integrated moving average model, and an autoregressive single-layer neural network), two univariate neural networks (N-BEATS and N-HiTS), and two multivariate neural networks designed for sequential data (a recurrent neural network with gated recurrent unit, GRU, and a Transformer network) to produce forecasts for six vital parameters recorded at five-minute intervals during intensive care monitoring. Vital parameters were the diastolic, systolic, and mean arterial blood pressure, central venous pressure, peripheral oxygen saturation (measured by non-invasive pulse oximetry) and heart rate, and forecasts were made for 5 through 120 minutes into the future. Patients used in this study recovered from cardiothoracic surgery in an ICU. The patient cohort used for model development (n = 22,348) and internal testing (n = 2,483) originated from a heart center in Germany, while a patient sub-set from the eICU collaborative research database, an American multicenter ICU cohort, was used for external testing (n = 7,477). The GRU was the predominant method in this study. Uni- and multivariate neural network models proved to be superior to univariate statistical models across vital parameters and forecast horizons, and their advantage steadily became more pronounced for increasing forecast horizons. With this study, we established an extensive set of benchmarks for forecast performance in the ICU. Our findings suggest that supplying physicians with short-term forecasts of vital parameters in the ICU is feasible, and that multivariate neural networks are most suited for the task due to their ability to learn patterns across thousands of patients.
RESUMEN
BACKGROUND: The Society for Cardiovascular Angiography and Interventions (SCAI) shock classification has been shown to provide robust mortality risk stratification in a variety of cardiovascular patients. OBJECTIVES: This study sought to evaluate the SCAI shock classification in postoperative cardiac surgery intensive care unit (CSICU) patients. METHODS: This study retrospectively analyzed 26,792 postoperative CSICU admissions at a heart center between 2012 and 2022. Patients were classified into SCAI shock stages A to E using electronic health record data. Moreover, the impact of late deterioration (LD) as an additional risk modifier was investigated. RESULTS: The proportions of patients in SCAI shock stages A to E were 24.4%, 18.8%, 8.4%, 35.5%, and 12.9%, and crude hospital mortality rates were 0.4%, 0.6%, 3.3%, 4.9%, and 30.2%, respectively. Similarly, the prevalence of postoperative complications and organ dysfunction increased across SCAI shock stages. After multivariable adjustment, each higher SCAI shock stage was associated with increased hospital mortality (adjusted OR: 1.26-16.59) compared with SCAI shock stage A, as was LD (adjusted OR: 8.2). The SCAI shock classification demonstrated a strong diagnostic performance for hospital mortality (area under the receiver operating characteristic: 0.84), which noticeably increased when LD was incorporated into the model (area under the receiver operating characteristic: 0.90). CONCLUSIONS: The SCAI shock classification effectively risk-stratifies postoperative CSICU patients for mortality, postoperative complications, and organ dysfunction. Its application could, therefore, be extended to the field of cardiac surgery as a triage tool in postoperative care and as a selection criterion in research.