Abstract Background and Aims Assessing and monitoring cognitive frailty (CF; coexistence of physical frailty and cognitive impairment) are critical for hemodialysis patients. For administering physical frailty and cognitive assessments, however, clinical settings need to address challenges (e.g., cost, human and technical resources, patient’s fatigue, etc.). This study aims to investigate the feasibility and accuracy of a novel intradialytic cognitive-demanding exercise program based on wearables, called, intradialytic-exergame for screening CF in hemodialysis patients. Method Individuals diagnosed with diabetes and end-stage renal disease requiring hemodialysis (n=28, age: 61.36 ± 6.85 years, 54% female) participated. All participants were assessed for physical frailty using the Fried frailty phenotype method and cognitive impairment using the Mini-Mental State Examination (MMSE). The Fried frailty phenotype method assesses physical frailty, which ranges between 0 and 5 based on five criteria (unintentional weight loss, weakness, slowness, exhaustion, and low physical activity). CF was determined with a frailty phenotype greater than or equal to 1 and a MMSE score less than or equal to 24. The intradialytic-exergame system consists of an inertial wearable sensor and an interactive software installed on a standard laptop. A clinician attaches one wearable sensor to each foot after the participant sits or lies down on a bed (Figure A). While undergoing hemodialysis treatment, the participant performs 15 non-weight-bearing cognitively-demanding dorsiflexion and plantarflexion exercises for each foot. For each exercise, the laptop’s screen displays a target (solid yellow circle) and a cursor (solid red square), which the participant points down for a dorsiflexion motion or up for a plantarflexion motion (Figure B). When the participant successfully puts the cursor in the target, the target disappears, and a new target appears in a different location on the screen (dashed yellow circle shown in Figure B). If the participant moves the cursor to a target in less than 2s, the custom software provides both visual (the target explodes) and auditory (positive sound) feedback as a reward (success). The three outcome measures (Exergame-slowness, frailty, and cognitive performance) were analyzed. Exergame-slowness was computed as an average of the ranges of ankle angular velocity measured by wearable sensors. Linear regression analysis was conducted for the three outcome measures to examine correlations between each outcome measure. Binary logistic regression model was conducted, and its area-under-curve (AUC) was calculated to determine the ability of Exergame-slowness to identify CF. An independent t-test was conducted to compare the differences of Exergame-slowness for CF and non-CF conditions. Significance was defined at the 2-sided p < 0.05 level. Results Five out of 28 participants were identified with CF. Significant correlations were observed between Exergame-slowness and frailty (p = 0.004, R = -0.531), Exergame-slowness and cognitive performance (p = 0.023, R = 0.437), and frailty and cognitive performance (p = 0.015, R = -0.463). The model was significantly reliable (p = 0.012) and its AUC was 0.90. Results indicated that Exergame-slowness was significantly higher (lower velocity) for participants with CF than for those with non-CF (CF: 27.41 ± 3.58 degree/s, non-CF: 34.25 ± 5.24 degree/s, p = 0.010). Conclusion To our knowledge, this is the first study to investigate the feasibility and accuracy of intradialytic-exergame with wearable sensors, as a practical tool for routine screening CF assessment in hemodialysis patients. The results of this study indicate that speed of ankle rotation, measurable using a wearable sensor during a simple intradialytic cognitive-demanding exercise, can be used as a practical digital biomarker for screening CF in hemodialysis patients.