Using CUSUM in real time to signal clinically relevant decreases in estimated glomerular filtration rate

Abstract

Background The electronic health record (EHR), utilized to apply statistical methodology, assists provider decision-making, including during the care of chronic kidney disease (CKD) patients. When estimated glomerular filtration (eGFR) decreases, the rate of that change adds meaning to a patient’s single eGFR and may represent severity of renal injury. Since the cumulative sum chart technique (CUSUM), often used in quality control and surveillance, continuously checks for change in a series of measurements, we selected this statistical tool to detect clinically relevant eGFR decreases and developed CUSUMGFR. Methods In a retrospective analysis we applied an age adjusted CUSUMGFR, to signal identification of eventual ESKD patients prior to diagnosis date. When the patient signaled by reaching a specified threshold value, days from CUSUM signal date to ESKD diagnosis date (earliness days) were measured, along with the corresponding eGFR measurement at the signal. Results Signaling occurred by CUSUMGFR on average 791 days (se = 12 days) prior to ESKD diagnosis date with sensitivity = 0.897, specificity = 0.877, and accuracy = .878. Mean days prior to ESKD diagnosis were significantly greater in Black patients (905 days) and patients with hypertension (852 days), diabetes (940 days), cardiovascular disease (1027 days), and hypercholesterolemia (971 days). Sensitivity and specificity did not vary by sociodemographic and clinical risk factors. Conclusions CUSUMGFR correctly identified 30.6% of CKD patients destined for ESKD when eGFR was > 60 ml/min/1.73 m2 and signaled 12.3% of patients that did not go on to ESKD (though almost all went on to later-stage CKD). If utilized in an EHR, signaling patients could focus providers’ efforts to slow or prevent progression to later stage CKD and ESKD.

Cost-effectiveness of screening for chronic kidney disease using a cumulative eGFR-based statistic

“…The goal of CUSUM GFR -based screening is to detect a clinically meaningful decline in eGFR as an early indicator for the risk of progression to severe stages of CKD and ESKD. A previous study showed the validity of using a statistical measure, CUSUM GFR , based on patients’ longitudinal eGFR values—a measure that only requires a highly accessible blood test, the basic metabolic panel (BMP)—for early detection of severe CKD [ 19 ]. For an individual patient, the CUSUM GFR statistic was updated after each measurement and captured the historical information accumulated from all prior eGFR observations.…”

“…For an individual patient, the CUSUM GFR statistic was updated after each measurement and captured the historical information accumulated from all prior eGFR observations. Specifically, CUSUM GFR is defined as follows [ 19 ]: where eGFR i indicates the eGFR measured at the i th observation, indicates the standard deviation of eGFR in a matched group of patients without chronic or acute kidney disease (CKD or AKD) diagnosis that was determined using ICD9/10 codes [ 19 ], and is the age-adjusted mean eGFR which decreases by 0.81 mL/min/1.73 m 2 annually. The variable w is the tuning parameter for the CUSUM GFR statistic.…”

“…At any time, if the CUSUM GFR statistic fell below a certain threshold T , the patient was identified as being at-risk for ESKD. The optimal values for the tuning parameter w and the threshold T had been identified to achieve the best predictive performance from the prior study [ 19 ].…”

“… * Value defined by 85.07 − 0.81Δ t , where Δ t is the difference in years between the age of the patient at time i and point of their first eGFR measurement [ 19 ]. ** Value estimated by comparing the slope of eGFR decline in ACE/ARBs treatment vs combined SGLT2 and ACE/ARBs treatment groups [ 32 ].…”

“…This is mainly due to the significant variation in eGFR measures and resulting temporal patterns. Zafarnejad et al [ 19 ] developed a statistical method based on the cumulative sum statistic (CUSUM GFR ), which could identify a statistically significant decrease in eGFR levels as the signal to indicate patients being at risk of ESKD and was validated using a large electronic health record (EHR) dataset. Given its desirable performance in effectively identifying at-risk patients in an earlier stage of CKD, the CUSUM GFR statistic offers a promising opportunity to be integrated into the screening strategy for CKD: Being easily interpretable by primary care providers, CUSUM GFR -based screening method can leverage the existing EHR data and require limited testing resources.…”