A simple non-physiological artifact filter for invasive arterial blood pressure monitoring: a study of 1852 trauma ICU patients

Effect of artifacts upon the pressure reactivity index

The pressure reactivity index (PRx) is a parameter for the assessment of cerebrovascular autoregulation, but its calculation is affected by artifacts in the source biosignals—intracranial pressure (ICP) and arterial blood pressure. We sought to describe the most common short-duration artifacts and their effect on the PRx. A retrospective analysis of 935 h of multimodal monitoring data was conducted, and five types of artifacts, characterized by their shape, duration, and amplitude, were identified: rectangular, fast impulse, isoline drift, saw tooth, and constant ICP value. Subsequently, all types of artifacts were mathematically modeled and inserted into undisturbed segments of biosignals. Fast impulse, the most common artifact, did not alter the PRx index significantly when inserted into one or both signals. Artifacts present in one signal exceeded the threshold PRx in less than 5% of samples, except for isoline drift. Compared to that, the shortest rectangular artifact inserted into both signals changed PRx to a value above the set threshold in 55.4% of cases. Our analysis shows that the effect of individual artifacts on the PRx index is variable, depending on their occurrence in one or both signals, duration, and shape. This different effect suggests that potentially not all artifacts need to be removed.

Real alerts and artifact classification in archived multi-signal vital sign monitoring data: implications for mining big data

Huge hospital information system databases can be mined for knowledge discovery and decision support, but artifact in stored non-invasive vital sign (VS) high-frequency data streams limits its use. We used machine-learning (ML) algorithms trained on expert-labeled VS data streams to automatically classify VS alerts as real or artifact, thereby "cleaning" such data for future modeling. 634 admissions to a step-down unit had recorded continuous noninvasive VS monitoring data [heart rate (HR), respiratory rate (RR), peripheral arterial oxygen saturation (SpO₂) at 1/20 Hz, and noninvasive oscillometric blood pressure (BP)]. Time data were across stability thresholds defined VS event epochs. Data were divided Block 1 as the ML training/cross-validation set and Block 2 the test set. Expert clinicians annotated Block 1 events as perceived real or artifact. After feature extraction, ML algorithms were trained to create and validate models automatically classifying events as real or artifact. The models were then tested on Block 2. Block 1 yielded 812 VS events, with 214 (26 %) judged by experts as artifact (RR 43 %, SpO₂ 40 %, BP 15 %, HR 2 %). ML algorithms applied to the Block 1 training/cross-validation set (tenfold cross-validation) gave area under the curve (AUC) scores of 0.97 RR, 0.91 BP and 0.76 SpO₂. Performance when applied to Block 2 test data was AUC 0.94 RR, 0.84 BP and 0.72 SpO₂. ML-defined algorithms applied to archived multi-signal continuous VS monitoring data allowed accurate automated classification of VS alerts as real or artifact, and could support data mining for future model building.

Multimodal monitoring in the pediatric intensive care unit: new modalities and informatics challenges

We review several newer modalities to monitor the brain in children with acute neurologic disease in the pediatric intensive care unit, such as partial brain tissue oxygen tension (PbtO2), jugular venous oxygen saturation (SjvO2), near infrared spectroscopy (NIRS), thermal diffusion measurement of cerebral blood flow, cerebral microdialysis, and EEG. We then discuss the informatics challenges to acquire, consolidate, analyze, and display the data. Acquisition includes multiple data types: discrete, waveform, and continuous. Consolidation requires device interoperability and time synchronization. Analysis could include pressure reactivity index and quantitative EEG. Displays should communicate the patient's current status, longitudinal and trend information, and critical alarms.