Predictive analytics in the pediatric intensive care unit for early identification of sepsis: capturing the context of age

Abstract

BACKGROUND: Early recognition of patients at risk for sepsis is paramount to improve clinical outcomes. We hypothesized that subtle signatures of illness are present in physiological and biochemical time series of pediatric-intensive care unit (PICU) patients in the early stages of sepsis. METHODS: We developed multivariate models in a retrospective observational cohort to predict the clinical diagnosis of sepsis in children. We focused on age as a predictor and asked whether random forest models, with their potential for multiple cut points, had better performance than logistic regression. RESULTS: One thousand seven hundred and eleven admissions for 1425 patients admitted to a mixed cardiac and medical/surgical PICU were included. We identified, through individual chart review, 187 sepsis diagnoses that were not within 14 days of a prior sepsis diagnosis. Multivariate models predicted sepsis in the next 24 h: cross-validated C-statistic for logistic regression and random forest were 0.74 (95% confidence interval (CI): 0.71-0.77) and 0.76 (95% CI: 0.73-0.79), respectively. CONCLUSIONS: Statistical models based on physiological and biochemical data already available in the PICU identify high-risk patients up to 24 h prior to the clinical diagnosis of sepsis. The random forest model was superior to logistic regression in capturing the context of age.

Principles for Real-World Implementation of Bedside Predictive Analytics Monitoring

“…10,11,13,15 To accept new tools like predictive analytics, particularly those that derive information from the physiologic waveforms, clinicians need to understand what is happening inside. 11,13,16,17 We presented the published evidence base 6,[18][19][20][21][22] to provide transparency into the algorithms' underpinnings and to emphasize the strengths of the scientific foundation. 16 We focused other sessions on current patients to give clinicians a more detailed examination of how data elements interacted within the algorithm to produce a risk score.…”

Dynamic Transitions of Pediatric Sepsis: A Markov Chain Analysis

“…Continuous electrocardiogram data from bedside monitors, vital signs, laboratory values, and clinical assessment findings in the electronic health record can be analyzed to identify patients at rising risk of sepsis, prior to overt clinical signs. Predictive models exist that were developed as time series measures of changing risk based on clinical variables that detect physiological changes with illness (5)(6)(7)(8)(9). The utility of such continuous predictive analytics is intuitive: novel monitoring to alert busy clinicians to a change in the patient so diagnosis and treatment can occur early.…”

Accuracy and Monitoring of Pediatric Early Warning Score (PEWS) Scores Prior to Emergent Pediatric Intensive Care Unit (ICU) Transfer: Retrospective Analysis

“…Continuous monitoring in children can be challenging to implement because it can be difficult to keep continuous monitoring leads and probes on mobile children, and previous estimates have demonstrated as few as 1% of alarms in children are clinically meaningful [29]. We speculate that there can be clinical benefit to shifting the clinical monitoring paradigm away from its use only as a means of responding to critical physiological alarms and towards a means for early detection of clinical deterioration using continuous predictive analytics monitoring so clinicians can initiate proactive clinical actions [28,30,31]. To avoid medical overuse and further contribution to false alarms, there is a defined need to determine the populations that could benefit the most from continuous cardiorespiratory monitoring in the acute care pediatric setting while also determining the correct "dose" of continuous cardiorespiratory monitoring for those at risk of clinical deterioration.…”

“…Further, there is substantial heterogeneity in ages represented in any pediatric sample. When Spaeder and colleagues [28] developed a machine learning model to predict early onset of pediatric sepsis, they found that parameters performed differently in the model given the age of the pediatric patient, again indicating that no one model likely performs equally well in all age ranges represented in pediatric care (neonate, infant, child, adolescent).…”