Improving acute kidney injury diagnostics using predictive analytics

Rajit K. Basu*, Katja Gist, Derek S. Wheeler

*Corresponding author for this work

Research output: Contribution to journalReview articlepeer-review

9 Scopus citations


Purpose of review Acute kidney injury (AKI) is a multifactorial syndrome affecting an alarming proportion of hospitalized patients. Although early recognition may expedite management, the ability to identify patients at-risk and those suffering real-time injury is inconsistent. The review will summarize the recent reports describing advancements in the area of AKI epidemiology, specifically focusing on risk scoring and predictive analytics. Recent findings In the critical care population, the primary underlying factors limiting prediction models include an inability to properly account for patient heterogeneity and underperforming metrics used to assess kidney function. Severity of illness scores demonstrate limited AKI predictive performance. Recent evidence suggests traditional methods for detecting AKI may be leveraged and ultimately replaced by newer, more sophisticated analytical tools capable of prediction and identification: risk stratification, novel AKI biomarkers, and clinical information systems. Additionally, the utility of novel biomarkers may be optimized through targeting using patient context, and may provide more granular information about the injury phenotype. Finally, manipulation of the electronic health record allows for real-time recognition of injury. Summary Integrating a high-functioning clinical information system with risk stratification methodology and novel biomarker yields a predictive analytic model for AKI diagnostics.

Original languageEnglish (US)
Pages (from-to)473-478
Number of pages6
JournalCurrent opinion in critical care
Issue number6
StatePublished - 2015


  • Biomarkers
  • Clinical information systems
  • Renal angina
  • Risk stratification

ASJC Scopus subject areas

  • Critical Care and Intensive Care Medicine


Dive into the research topics of 'Improving acute kidney injury diagnostics using predictive analytics'. Together they form a unique fingerprint.

Cite this