Electronic health record-based predictive models for acute kidney injury screening in pediatric inpatients

Abstract

Background Acute kidney injury (AKI) is common in pediatric inpatients and associated with increased morbidity, mortality, and length of stay. Early identification can reduce severity. Methods To create and validate an electronic health record (EHR)-based AKI screening tool, we generated temporally distinct development and validation cohorts using retrospective data from our tertiary care children’s hospital, including children 28 days through 21 years old with sufficient serum creatinine measurements to determine AKI status. AKI was defined as 1.5-fold or 0.3 mg/dL increase in serum creatinine. Age, medication exposures, platelet count, red blood cell distribution width, serum phosphorus, serum transaminases, hypotension (ICU only), and pH (ICU only) were included in AKI risk prediction models. Results For ICU patients, 791/1332 (59%) of the development cohort and 470/866 (54%) of the validation cohort had AKI. In external validation, the ICU prediction model had C-statistic=0.74 (95% confidence interval 0.71–0.77). For non-ICU patients, 722/2337 (31%) and 469/1474 (32%) had AKI, and the prediction model had C-statistic=0.69 (0.66–0.72). Conclusions AKI screening can be performed using EHR data. The AKI screening tool can be incorporated into EHR systems to identify high risk patients without serum creatinine data, enabling targeted laboratory testing, early AKI identification, and modification of care.

Epidemiology and outcomes of community-acquired and hospital-acquired acute kidney injury in children and adolescents

“…AKI is often preventable and reversible. Computerized provider order entry intervention (25,26), electronic alerts (27,28), risk assessment toolkits (29), and care bundles (30) are being increasingly implemented to facilitate early detection, improve care processes, prevent worsening of patient outcomes, and reduce health-care resource utilization in different settings. For example, identified AKI risk factors can be used to generate predictive models to screen individual patients for AKI in specific settings, such as intensive care unit (ICU) and non-ICU settings (29).…”

“…Computerized provider order entry intervention (25,26), electronic alerts (27,28), risk assessment toolkits (29), and care bundles (30) are being increasingly implemented to facilitate early detection, improve care processes, prevent worsening of patient outcomes, and reduce health-care resource utilization in different settings. For example, identified AKI risk factors can be used to generate predictive models to screen individual patients for AKI in specific settings, such as intensive care unit (ICU) and non-ICU settings (29). Considering the increase in the incidence of CA-AKI, further studies could be useful in determining the application of risk factors identified in our analysis (patient-related underlying diseases, acute illnesses, and nephrotoxic medicines) in the development of e-alert systems and risk assessment tools for pediatric patients who are at risk for AKI in outpatient settings.…”

Big Data and Pediatric Acute Kidney Injury: The Promise of Electronic Health Record Systems

“…Regardless, once the AKI has resolved, the ability to identify these patients accurately, allows them to be "tagged" and followed whether they developed ESRD, CKD, or experienced full recovery. potential predictors were chosen a priori based upon their association with AKI in prior studies (40)(41)(42)(43)(44)(45). While certainly statistically sound, these approaches do not take full advantage of big data informatics methods.…”

Acute kidney injury risk-based screening in pediatric inpatients: a pragmatic randomized trial

“…The EHR-based risk prediction models for AKI have been described. (28) In brief, using EHR data from pediatric ICU patients, we developed and validated a statistical model to predict AKI risk during ICU hospitalization based on 10 characteristics (age, high-risk nephrotoxins, moderate risk nephrotoxins, total number of medications, platelet count, red blood cell distribution width, phosphorus, transaminases, pH, and hypotension). A second model was generated for pediatric ward patients using 8 characteristics (as above, without pH or hypotension).…”