Concurrent validity of inertial sensors for postural balance analysis in patients with chronic low back pain and asymptomatic individuals

Abstract

Introduction: Human movement analysis is an aspect of great relevance to the health field. Routinely, in research and clinical settings, assessments are conducted to investigated effects caused by clinical conditions -as chronic low back pain (CLBP)- and screening rehabilitation process. Most of the studies about movement analysis include postural balance assessment. Several methods are used, but the most accurate tools have high cost and are not feasible for use in the clinical routine. With technology advancement, low-cost equipment, such as inertial sensors, is currently available. These devices are portable and therefore, more accessible. Therefore, there is no consensus about the validity of the inertial sensors. Objective: To investigate the concurrent validity of inertial sensors as an instrument of measuring body sway during the upright standing in patients with nonspecific low back pain and asymptomatic patients. Methods: A cross-sectional study was conducted according to the requirements of the Standards for Diagnostic Accuracy Reports , in 39 patients with chronic low back pain (CLBP) (mean [SD]: age = 33.38 [11.60] years, pain intensity at the moment of the assessment = 47 [1,3]) and 39 asymptomatic subjects (age = 32.50 [10.26 years]). The balance analysis was performed through the upright posture with two inertial sensors positioned on the participants, one on the lumbar region and another on the sternum. The force plate was the gold standard instrument used for the concurrent validation that was investigated by Spearman's correlation coefficient. The variables analysed were Root Mean Square (RMS), index of sway smoothness (JERK), total trajectory length (Path) and eliptic area (AREA). Results: The asymptomatic group showed statistically significant correlations for RMS variables (r = 0.38; p = 0.03) with sensor positioned on the lumbar and (r = 0.42; p = 0.02) with sensor positioned on the sternum. Patients in CLBP showed a statistically significant correlation with the sensors positioned in the lower back for RMS variables (r = 0.59; p <0.01), path (r = 0.42, p = 0.01) and area (r = 0.59; p <0.01). Conclusion: The body oscillation measurements obtained by the inertial sensors showed weak to moderate correlations concerning the force plate. The positioning of the inertial sensor in the lumbar region seems to be the most adequate for balance analysis since it obtained more relevant correlations than the inertial sensor positioned in the sternum.