Agreement and Reliability of Patient-Measured Post-Void Residual Bladder Volumes

Back to 2023 Abstracts

Agreement and Reliability of Patient-Measured Post-Void Residual Bladder Volumes
Michael Jalfon, MD¹, Mursal Gardezi, MD¹, Dylan Heckscher, MD¹, Devin Shaheen, BS², Kaitlin R. Maciejewski, MS³, Fangyong Li, MS, MPH³, Leslie Rickey, MD, MPH¹, Harris Foster, MD¹, Jaimie A. Cavallo, MD, MPHS¹.
¹Yale School of Medicine, New Haven, CT, USA, ²Yale School of Nursing, West Haven, CT, USA, ³Yale Center for Analytical Sciences, New Haven, CT, USA.

BACKGROUND: Post-void residual bladder volume (PVR) measurement in the non-healthcare setting would be a valuable opportunity for remote monitoring of patients with voiding dysfunction. We hypothesized that patient self-measurement of PVR using a smart device-integrated portable ultrasound probe equipped with artificial intelligence (AI)-based automated bladder volume measurement software would demonstrate high reliability, strong agreement with provider measurement, and be preferred by subjects over provider measurements in the healthcare setting.
METHODS: Eligible adult subjects were enrolled during outpatient Urology visits. Subject PVRs were measured in triplicate by providers using each of the following FDA-cleared devices: standard bladder scanners (either the Verathon^TM BVI 9400 System or the Verathon^TM BladderScan Prime System), the portable ultrasound probe with AI-based automated bladder volume measurement software using ultrasound images (Butterfly^TM US image mode), and the portable ultrasound probe with AI-based automated bladder volume measurement software using abstract images (Butterfly^TM abstract mode). Subjects blinded to the provider measurements then self-measured PVR in triplicate using both the US and abstract modes. Subjects were randomized to use either Butterfly^TM US mode or Butterfly^TM abstract mode first following a brief tutorial. Subject-reported preferences and evaluations were assessed via electronic survey. Agreement between PVR measurements for each combination of devices was assessed via Bland-Altman analyses using a clinically acceptable difference threshold of 50 mL. Reliability was assessed via intraclass correlation (ICC). Two-sided statistical significance was set at p<0.05.
RESULTS: 50 subjects were enrolled. ICCs ranged between 0.95-0.98 for each device. Results of the Bland-Altman analyses are displayed in Table 1. Mean difference in PVR between provider standard bladder scanner measurement and subject self-measurement were 7.5 mL (95% limits of agreement: -71.73 mL, 86.73 mL) using Butterfly^TM US mode and 9.34 mL (95% limits of agreement: -93.84 mL, 112.52 mL) using Butterfly^TM abstract mode, respectively. No statistically significant differences in agreement were observed for any combination of devices when stratified by age, sex, BMI, or subject-reported dexterity. Statistically significant differences in agreement were observed between subject US mode and both provider standard bladder scanner (p=0.02) and provider abstract modes (p=0.03) when stratified by subject-reported technological comfort. Most subjects preferred self-measurement over provider measurement (74% vs 26%, respectively).
CONCLUSIONS: Subject self-measurement of PVR using smartdevice-integrated portable ultrasound probes is feasible, reliable, and preferred by subjects. Limits of agreement between subject self-measurement and provider standard bladder scanner measurement of PVR exceeded our clinically acceptable difference threshold, though the published inherent error of each ultrasound-based bladder volume measurement device should be considered. Additional studies are warranted to determine if longitudinal subject self-measurements of PVR in real-world settings facilitate remote monitoring of patients with voiding dysfunction.

Table 1. Bland-Altman Bias and Limits of Agreement for All Method Combinations
Method 1	Method 2	Bias	95% Limits of Agreement
Standard Bladder Scanner PVR Mean	Provider Butterfly^TM US Image Mode PVR Mean	13.17	-59.65, 85.89
Standard Bladder Scanner PVR Mean	Provider Butterfly^TM Abstract Mode PVR Mean	12.51	-55.30, 80.33
Standard Bladder Scanner PVR Mean	Patient Butterfly^TM US Image Mode PVR Mean	7.50	-71.73, 86.73
Standard Bladder Scanner PVR Mean	Patient Butterfly^TM Abstract Mode PVR Mean	9.34	-93.84, 112.52
Provider Butterfly^TM US Image Mode PVR Mean	Provider Butterfly^TM Abstract Mode PVR Mean	-0.65	-71.28, 69.97
Provider Butterfly^TM US Image Mode PVR Mean	Patient Butterfly^TM US Image Mode PVR Mean	-5.67	-91.19, 79.86
Provider Butterfly^TM US Image Mode PVR Mean	Patient Butterfly^TM Abstract Mode PVR Mean	-3.83	-105.05, 97.40
Provider Butterfly^TM Abstract Mode PVR Mean	Patient Butterfly^TM US Image Mode PVR Mean	-5.01	-85.48, 75.45
Provider Butterfly^TM Abstract Mode PVR Mean	Patient Butterfly^TM Abstract Mode PVR Mean	-3.17	-71.56, 65.21
Patient Butterfly^TM US Image Mode PVR Mean	Patient Butterfly^TM Abstract Mode PVR Mean	1.84	-87.99, 91.67
Note:Bias = mean difference of the measurements.Limits of agreement = 1.96 standard deviations from the mean difference.All values are expressed in mL.

Back to 2023 Abstracts