The Computerized Algorithm for Renal Assessment Improves Diagnostic Accuracy of Renal Impairment in Hospitalized Patients

Background: Acute kidney injury (AKI) is an important adverse event of hospitalized patients that associates high mortality and high medical cost. While accurate diagnosis and timely management of AKI is essential for improving the outcomes of in-hospital AKI, delayed diagnosis or misdiagnosis by clinicians hinders the advance in the care of AKI. To ameliorate this problem, several electronic AKI alert systems had been proposed, which had shown inconsistent effects on the outcomes of AKI. Before electronic systems could improve the outcomes of AKI, an important issue to be elucidated is to confirm their diagnostic accuracy. The purpose of the present study is to establish an easy-to-construct Computerized Algorithm for the diagnosis of renal impairment and to test its accuracy. Methods: The present study retrospectively included 1551 hospitalized patients with serum creatinine (SCr) of > 1.3 mg/dL in Wanfang Hospital. A Computerized Algorithm was constructed to identify AKI events and CKD cases among these patients. Previous SCr tests were reviewed to define the baseline SCr level. Increased SCr level of > 1.5 times from baseline was defined as AKI. Estimated glomerular filtration rate (eGFR) of < 60 mL/min/1.73m2 for > 90 days was defined as CKD. The standard diagnoses were made by experienced nephrologists according to the same criteria. The discharge diagnoses made by the attending physician was defined as clinician’s diagnoses. The diagnoses made by Computerized Algorithm and the clinicians were compared with the researcher’s diagnoses to define their accuracy. Results: Of the included patients, the mean age was 73.0 years old; the in-hospital mortality was 14.8%; AKI was present in 28.6% of the patients. Regarding the diagnostic accuracy for AKI, the Computerized Algorithm achieved a sensitivity of 85.6% and a specificity of 98.8%. The main cause of false negative (FN) AKI diagnosis was that AKI occurred earlier than the outpatient visiting before the indexed hospitalization. Regarding the diagnostic accuracy for CKD, the Computerized Algorithm achieved a sensitivity of 94.7% and a specificity of 100%. The cause of FN CKD diagnosis was exclusively lack of previous eGFR records. Compared to the clinician’s diagnoses, the Computerized Algorithm exhibited significantly superior accuracy than the clinician’s diagnoses in both AKI (95.0% versus 57.0%) and CKD (96.5% versus 73.6%). Conclusions: We developed a simple and easy-to-construct Computerized Algorithm for the diagnosis of renal impairment, which significantly improved the diagnostic accuracy of AKI and CKD than clinicians did. In future, an algorithmic approach for differential diagnosis of AKI and decision guide with be incorporated into this system.