Data-driving methods: More than merely trendy buzzwords?

Machine learning methods to improve bedside fluid responsiveness prediction in severe sepsis or septic shock: an observational study

BACKGROUND

Passive leg raising (PLR) predicts fluid responsiveness in critical illness, although restrictions in mobilising patients often preclude this haemodynamic challenge being used. We investigated whether machine learning applied on transthoracic echocardiography (TTE) data might be used as a tool for predicting fluid responsiveness in critically ill patients.

METHODS

We studied, 100 critically ill patients (mean age: 62 yr [standard deviation: 14]) with severe sepsis or septic shock prospectively over 24 months. Transthoracic echocardiography measurements were performed at baseline, after PLR, and before and after a standardised fluid challenge in learning and test populations (n=50 patients each). A 15% increase in stroke volume defined fluid responsiveness. The machine learning methods used were classification and regression tree (CART), partial least-squares regression (PLS), neural network (NNET), and linear discriminant analysis (LDA). Each method was applied offline to determine whether fluid responsiveness may be predicted from left and right cardiac ventricular physiological changes detected by cardiac ultrasound. Predictive values for fluid responsiveness were compared by receiver operating characteristics (area under the curve [AUC]; mean [95% confidence intervals]).

RESULTS

In the learning sample, the AUC values were PLR 0.76 (0.62-0.89), CART 0.83 (0.73-0.94), PLS 0.97 (0.93-1), NNET 0.93 (0.85-1), and LDA 0.90 (0.81-0.98). In the test sample, the AUC values were PLR 0.77 (0.64-0.91), CART 0.68 (0.54-0.81), PLS 0.83 (0.71-0.96), NNET 0.83 (0.71-0.94), and LDA 0.85 (0.74-0.96) respectively. The PLS model identified inferior vena cava collapsibility, velocity-time integral, S-wave, E/Ea ratio, and E-wave as key echocardiographic parameters.

CONCLUSIONS

Machine learning generated several models for predicting fluid responsiveness that were comparable with the haemodynamic response to PLR.

Predicting the microbial cause of community-acquired pneumonia: can physicians or a data-driven method differentiate viral from bacterial pneumonia at patient presentation?

BACKGROUND

Community-acquired pneumonia (CAP) requires urgent and specific antimicrobial therapy. However, the causal pathogen is typically unknown at the point when anti-infective therapeutics must be initiated. Physicians synthesize information from diverse data streams to make appropriate decisions. Artificial intelligence (AI) excels at finding complex relationships in large volumes of data. We aimed to evaluate the abilities of experienced physicians and AI to answer this question at patient admission: is it a viral or a bacterial pneumonia?

METHODS

We included patients hospitalized for CAP and recorded all data available in the first 3-h period of care (clinical, biological and radiological information). For this proof-of-concept investigation, we decided to study only CAP caused by a singular and identified pathogen. We built a machine learning model prediction using all collected data. Finally, an independent validation set of samples was used to test the pathogen prediction performance of: (i) a panel of three experts and (ii) the AI algorithm. Both were blinded regarding the final microbial diagnosis. Positive likelihood ratio (LR) values > 10 and negative LR values < 0.1 were considered clinically relevant.

RESULTS

We included 153 patients with CAP (70.6% men; 62 [51-73] years old; mean SAPSII, 37 [27-47]), 37% had viral pneumonia, 24% had bacterial pneumonia, 20% had a co-infection and 19% had no identified respiratory pathogen. We performed the analysis on 93 patients as co-pathogen and no-pathogen cases were excluded. The discriminant abilities of the AI approach were low to moderate (LR+ = 2.12 for viral and 6.29 for bacterial pneumonia), and the discriminant abilities of the experts were very low to low (LR+ = 3.81 for viral and 1.89 for bacterial pneumonia).

CONCLUSION

Neither experts nor an AI algorithm can predict the microbial etiology of CAP within the first hours of hospitalization when there is an urgent need to define the anti-infective therapeutic strategy.