Photo Credit: The best photo for all

The following is a summary of “Machine learning for the prediction of mortality in patients with sepsis-associated acute kidney injury: a systematic review and meta-analysis,” published in the December 2024 issue of Infectious Disease by Lv et al. 

Machine learning is increasingly used to predict mortality in sepsis-related acute kidney injury, aiding early data-driven treatment decisions.  

Researchers conducted a retrospective study to assess the predictive value of machine learning for mortality in individuals with septic acute kidney injury.  

They searched the PubMed, Web of Science, Cochrane Library, and Embase databases up to July 20, 2024, with a manual search of references and review articles. Data analysis was performed by STATA 14.0 software. The risk of bias in the prediction model was evaluated using the Predictive Model Risk of Bias Assessment Tool.  

The results showed that 8 studies were included, utilizing 53 predictive models and 17 machine learning algorithms. Meta-analysis employing a random effects model revealed that the overall C-index in the training set was 0.81 (95% CI: 0.78–0.84), with a sensitivity of 0.39 (95% CI: 0.32–0.47) and specificity of 0.92 (95% CI: 0.89–0.95). In the validation set, the overall C-index was 0.73 (95% CI: 0.71–0.74), sensitivity 0.54 (95% CI: 0.48–0.60), and specificity 0.90 (95% CI: 0.88–0.91). These findings indicated that machine learning algorithms demonstrated good performance in predicting mortality in sepsis-related acute kidney injury.  

Investigators concluded the machine learning effectively predicted sepsis-associated acute kidney injury mortality, with potential for improved risk assessment and clinical decision-making, and future research should focus on larger, multi-center studies to validate these models across diverse patient populations. 

Source: bmcinfectdis.biomedcentral.com/articles/10.1186/s12879-024-10380-6