Data from: Early, individualized recommendations for hospitalized patients with acute kidney injury: A randomized clinical trial
Data files
Apr 28, 2026 version files 326.54 KB
-
KAT-AKI_limited_dataset.04-23-26.csv
323.24 KB
-
README.md
3.30 KB
Abstract
Importance: Acute kidney injury (AKI) is a common complication during hospitalization associated with poor outcomes.
Objective: To evaluate whether diagnostic and therapeutic recommendations sent from a Kidney Action Team through the electronic health record improve outcomes among patients hospitalized with AKI, compared to usual care.
Design, Setting, participants: Randomized, clinical trial conducted at two US health systems including seven hospitals. Hospitalized patients with AKI were randomized between October 29th, 2021, to February 8th, 2024. Final follow-up occurred in February 2024.
Intervention: An alert was sent to the Kidney Action Team, consisting of a physician and pharmacist (never directly to the clinical providers) who provided personalized recommendations for patients with AKI in five major categories (general diagnostics, volume, potassium, acid-base, and medications) within 1 hour via a structured note in the electronic health record. The note is immediately visible to anyone who has access to the chart. Randomization to the intervention or usual care occurred after the note was generated, but the note was only delivered to clinicians of patients randomized to the intervention group.
Main Outcomes and Measures: The primary outcome was a composite outcome consisting of AKI progression to a higher stage of AKI, dialysis, or mortality occurring while the patient remained hospitalized but within 14 days from randomization.
Results: 4,003 patients (median age 72 (61-81) years; 1874 (47%) female; 931 (23%) Black were randomized. The KAT made 14,539 recommendations, with a median (IQR) of 3 (2-5) per patient. The primary outcome occurred in 19.8% of the intervention group and in 18.4% in the usual care group (difference: 1.4%, 95% CI: -1.1, 3.8, P=0.28). Of six secondary outcomes, only rates of recommendation implementation significantly differed between the two groups: 2459/7270 (33.8%) of recommendations were implemented in the intervention group, and 1766/7269 (24.3%) of the (undelivered) recommendations were implemented in the usual care group within 24 hours (difference: 9.5%, 95% CI 8.1%, 11.0%, p<0.001).
Conclusions and Relevance: Among patients hospitalized with AKI, recommendations from a Kidney Action Team did not significantly reduce the composite outcome of worsening AKI stage, dialysis, or mortality, despite a higher rate of recommendation implementation in the intervention group, compared to usual care.
Trial Registration: NCT04040296 (https://clinicaltrials.gov/study/NCT04040296)
https://doi.org/10.5061/dryad.sqv9s4ncs
Description of the data and file structure
The data was collected during the course of the Kidney Action Team clinical trial, conducted at Yale and Johns Hopkins from 2022 to 2024.
Files and variables
File: KAT-AKI_limited_dataset.04-23-26.csv
Description: Analytic dataset for the KAT AKI trial.
Variables
- age: Age of the patient (rounded to nearest year. Age above 90 is listed as 90 with a binary flag - see below)
- femalesex: 1 = female
- hospital: in order to account for stratified randomization
- ckd_pmhx: past medical history of chronic kidney disease
- chf_pmhx: past medical history of congestive heart failure
- dm_pmhx: past medical history of diabetes
- htn_pmhx: past medical history of hypertension
- cirrhosis_pmhx: past medical history of cirrhossi
- baseline_creat: baseline creatinine level (mg/dL)
- start2enrollment_days: days between the start of the trial and the patient enrollment
- randomization: 1 = intervention, 0 = usual care
- primary_outcome: 1= experienced primary outcome (composite of worsening stage of AKI, dialysis, or death within 14 days of randomization)
- aki_progression14: 1 = experienced the AKI progression outcome
- dialysis14: 1 = experienced the dialysis outcome
- death14: 1 = experienced the death outcome
- discharge_to_hospice: 1 = discharged to a hospice program
- consult14: 1 = received a renal consult within 14 days of randomization
- general: 1 = any general recs made
- general_diag_rec_count: the number of general recs made
- volume: 1 = any volume recs made
- volume_rec_count: the number of volume recs made
- potas: 1 = any potassium recs made
- potas_rec_count: the number of potassium recs made
- acidosis: 1 = any acidosis recs made
- acidosis_rec_count: the number of acidosis recs made
- meds: 1 = any medication recs made
- meds_rec_count: the number of medication recs made
- renal_consult_rec: 1 = KAT recommended a formal renal consult
- general_diag_done_count24: count of general recs completed
- volume_done_count24: count of volume recs completed
- potas_done_count24: count of potassium recs completed
- acidosis_done_count24: count of acidosis recs completed
- meds_done_count24: count of med recs completed
- renal_consulted24: 1 = renal consult was ordered within 24 hours of randomization
- total_recs: count of total recs made
- total_recs_done24: count of total recs completed
- age_over_90: 1 = age is over 90
Access information
Other publicly accessible locations of the data:
- N/A
Data was derived from the following sources:
- Electronic health record review
Human subjects data
This research was granted a waiver of informed consent by the Yale IRB, under 45 CFR § 46.116. To ensure participant privacy, the dataset was de-identified by removing all 18 standard categories of direct personal identifiers. Temporal data (such as dates of enrollment or clinical visits) were converted to elapsed days from a baseline index date (Day 0) rather than absolute dates. Participant ages >=90 years were recoded into a single top-coded category.