Epidemiology, Trajectories and Outcomes of Acute Kidney Injury Among Hospitalized Patients: A Retrospective Multicenter Large Cohort Study
Authors:
Esra Adiyeke,
Yuanfang Ren,
Shmuel Fogel,
Parisa Rashidi,
Mark Segal,
Elizabeth A. Shenkman,
Azra Bihorac,
Tezcan Ozrazgat-Baslanti
Abstract:
Background: Acute kidney injury (AKI) is a clinical syndrome affecting almost one fifth of hospitalized patients, as well as more than half of the patients who are admitted to the intensive care unit (ICU). Stratifying AKI patients into groups based on severity and duration would facilitate more targeted efforts for treating AKI. Methods: In a retrospective, multicenter and longitudinal cohort stu…
▽ More
Background: Acute kidney injury (AKI) is a clinical syndrome affecting almost one fifth of hospitalized patients, as well as more than half of the patients who are admitted to the intensive care unit (ICU). Stratifying AKI patients into groups based on severity and duration would facilitate more targeted efforts for treating AKI. Methods: In a retrospective, multicenter and longitudinal cohort study of 935,679 patients who were admitted between 2012 and 2020 to health centers included in OneFlorida+ Network, we analyzed the impact of AKI trajectories which are rapidly reversed AKI, persistent AKI with renal recovery, and persistent AKI without renal recovery on patients' clinical outcomes, including hospital, 30-day, 1-year, and 3-year mortality, kidney replacement therapy, new chronic kidney disease (CKD) within 90 days or 1-year of discharge, CKD progression within 1-year of discharge, resource utilization, hospital disposition, and major complications during hospitalization. As analytical approaches, Kaplan-Meier estimators and survival curves, Cox proportional-hazards regression model, logistic regression model, Kruskal-Wallis test, analysis of variance, chi-square, Fisher's exact test were used. Results: Among 2,187,254 encounters, 14% had AKI, of which 63%, 21%, and 16% had Stage 1, 2, and 3, respectively, as the worst AKI stage. Fraction of patients with persistent AKI was 31%. Patients with AKI had worse clinical outcomes and increased resource utilization compared to patients without the condition. One-year mortality was 5 times greater for patients with persistent AKI compared to those without AKI. Conclusions: Persistent AKI was associated with prolonged hospitalization, increased ICU admission and mortality compared to the other groups. This may emphasize the critical need for devising strategies targeting effective management of AKI and prevention of persisting AKI.
△ Less
Submitted 12 March, 2024;
originally announced March 2024.
The impact of electronic health records (EHR) data continuity on prediction model fairness and racial-ethnic disparities
Authors:
Yu Huang,
**gchuan Guo,
Zhaoyi Chen,
Jie Xu,
William T Donahoo,
Olveen Carasquillo,
Hrushyang Adloori,
Jiang Bian,
Elizabeth A Shenkman
Abstract:
Electronic health records (EHR) data have considerable variability in data completeness across sites and patients. Lack of "EHR data-continuity" or "EHR data-discontinuity", defined as "having medical information recorded outside the reach of an EHR system" can lead to a substantial amount of information bias. The objective of this study was to comprehensively evaluate (1) how EHR data-discontinui…
▽ More
Electronic health records (EHR) data have considerable variability in data completeness across sites and patients. Lack of "EHR data-continuity" or "EHR data-discontinuity", defined as "having medical information recorded outside the reach of an EHR system" can lead to a substantial amount of information bias. The objective of this study was to comprehensively evaluate (1) how EHR data-discontinuity introduces data bias, (2) case finding algorithms affect downstream prediction models, and (3) how algorithmic fairness is associated with racial-ethnic disparities. We leveraged our EHRs linked with Medicaid and Medicare claims data in the OneFlorida+ network and used a validated measure (i.e., Mean Proportions of Encounters Captured [MPEC]) to estimate patients' EHR data continuity. We developed a machine learning model for predicting type 2 diabetes (T2D) diagnosis as the use case for this work. We found that using cohorts selected by different levels of EHR data-continuity affects utilities in disease prediction tasks. The prediction models trained on high continuity data will have a worse fit on low continuity data. We also found variations in racial and ethnic disparities in model performances and model fairness in models developed using different degrees of data continuity. Our results suggest that careful evaluation of data continuity is critical to improving the validity of real-world evidence generated by EHR data and health equity.
△ Less
Submitted 4 September, 2023;
originally announced September 2023.
