Search

Background: Simultaneous urine testing for albumin (UAlb) and serum creatinine (SCr), that is, 'dual testing,' is an accepted quality measure in the management of diabetes. As chronic kidney disease (CKD) is defined by both UAlb and SCr testing, this approach could be more widely adopted in kidney care., Objective: We assessed time trends and facility-level variation in the performance of outpatient dual testing in the integrated Veterans Health Administration (VHA) system., Design, Subjects and Main Measures: This retrospective cohort study included patients with any inpatient or outpatient visit to the VHA system during the period 2009-2018. Dual testing was defined as UAlb and SCr testing in the outpatient setting within a calendar year. We assessed time trends in dual testing by demographics, comorbidities, high-risk (eg, diabetes) specialty care and facilities. A generalised linear mixed-effects model was applied to explore individual and facility-level predictors of receiving dual testing., Key Results: We analysed data from approximately 6.9 million veterans per year. Dual testing increased, on average, from 17.4% to 21.2%, but varied substantially among VHA centres (0.3%-43.7% in 2018). Dual testing was strongly associated with diabetes (OR 10.4, 95% CI 10.3 to 10.5, p<0.0001) and not associated with VHA centre complexity level. However, among patients with high-risk conditions including diabetes, <50% received dual testing in any given year. As compared with white veterans, black veterans were less likely to be tested after adjusting for other individual and facility characteristics (OR 0.93, 95% CI 0.92 to 0.93, p<0.0001)., Conclusions: Dual testing for CKD in high-risk specialties is increasing but remains low. This appears primarily due to low rates of testing for albuminuria. Promoting dual testing in high-risk patients will help to improve disease management and patient outcomes., Competing Interests: Competing interests: None declared., (© Author(s) (or their employer(s)) 2024. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

Delays in the identification of acute kidney injury (AKI) in hospitalized patients are a major barrier to the development of effective interventions to treat AKI. A recent study by Tomasev and colleagues at DeepMind described a model that achieved a state-of-the-art performance in predicting AKI up to 48 hours in advance. 1 Because this model was trained in a population of US Veterans that was 94% male, questions have arisen about its reproducibility and generalizability. In this study, we aimed to reproduce key aspects of this model, trained and evaluated it in a similar population of US Veterans, and evaluated its generalizability in a large academic hospital setting. We found that the model performed worse in predicting AKI in females in both populations, with miscalibration in lower stages of AKI and worse discrimination (a lower area under the curve) in higher stages of AKI. We demonstrate that while this discrepancy in performance can be largely corrected in non-Veterans by updating the original model using data from a sex-balanced academic hospital cohort, the worse model performance persists in Veterans. Our study sheds light on the importance of reproducing artificial intelligence studies, and on the complexity of discrepancies in model performance in subgroups that cannot be explained simply on the basis of sample size.

Background and Objectives: AKI, a frequent complication among hospitalized patients, confers excess short- and long-term mortality. We sought to determine trends in in-hospital and 1-year mortality associated with AKI as defined by Kidney Disease Improving Global Outcomes consensus criteria., Design, Setting, Participants, & Measurements: This retrospective cohort study used data from the national Veterans Health Administration on all patients hospitalized from October 1, 2008 to September 31, 2017. AKI was defined by Kidney Disease Improving Global Outcomes serum creatinine criteria. In-hospital and 1-year mortality trends were analyzed in patients with and without AKI using Cox regression with year as a continuous variable., Results: We identified 1,688,457 patients and 2,689,093 hospitalizations across the study period. Among patients with AKI, 6% died in hospital, and 28% died within 1 year. In contrast, in-hospital and 1-year mortality rates were 0.8% and 14%, respectively, among non-AKI hospitalizations. During the study period, there was a slight decline in crude in-hospital AKI-associated mortality (hazard ratio, 0.98 per year; 95% confidence interval, 0.98 to 0.99) that was attenuated after accounting for patient demographics, comorbid conditions, and acute hospitalization characteristics (adjusted hazard ratio, 0.99 per year; 95% confidence interval, 0.99 to 1.00). This stable temporal trend in mortality persisted at 1 year (adjusted hazard ratio, 1.00 per year; 95% confidence interval, 0.99 to 1.00)., Conclusions: AKI associated mortality remains high, as greater than one in four patients with AKI died within 1 year of hospitalization. Over the past decade, there seems to have been no significant progress toward improving in-hospital or long-term AKI survivorship., (Copyright © 2022 by the American Society of Nephrology.)