Initiating a lymphedema screening program for breast cancer patients: Determining variability as a first step towards standardization.

Abstract

267 Background: Breast cancer-related lymphedema (BCRL) is a known morbidity from treatment, and yet published estimates of its incidence are highly variable. Typically, a single volume measurement is recorded and the degree of intra-rater variability undefined. A threshold 5% limb volume change over time is used to define “mild lymphedema.” Therefore, small variations in volume measurements could substantially lead to over or under-diagnosis of BRCL. Gage R&R is a measurement systems analysis tool that quantifies the amount of variability induced in measurements by the measurement system itself and compares this to the total variability observed. The objective of this study was to define the degree of variability in perometer arm volume measurements and factors contributing to this. Methods: A medical assistant (MA) was trained by a certified lymphedema therapist (CLT) to perform perometer-based lymphedema screening in the MD Anderson Nellie B. Connally Breast Center. Three measurements of each arm were taken and intra-rater variability against the mean calculated. Gage R&R testing was performed by the MA and CLT, to determine the degree of repeatability and reproducibility variation. As a part of this analysis, 30 patients had both their left and right arm volume measured 3 times by each measurer. Results: Over 6 months, 565 breast cancer patients had preoperative BCRL screening by the MA, with an overall intra-rater variability of 2.2%. Among the 30 patients in the Gage R&R study, median body mass index (BMI) was 27.7 (range 19.7-38.3), median height 1.61m (range 1.52-1.73m), intrarater variability of the MA was 2.07% and of the CLT was 1.21%. Gage R&R accounted for 4.77% of total variability, 1.26% repeatability and 3.51% reproducibility. Regression analysis demonstrated that patient BMI and height impacted variability in measurements (R2= 16.27). Conclusions: Intra-rater variability for perometer arm measurements diminishes with experience. Gage R&R variability is acceptable per industry standards, but there is room for improvement. Our findings suggest a role for targeted interventions to minimize variability in the diagnosis of BCRL based on patient characteristics.