Applying the OMERACT Truth filter to a New Electronic Spinal Mobility Index for Axial Spondyloarthritis Based on Inertial Measurement Unit (IMU) Sensors

Background: Loss of spinal mobility is one of the most characteristic problems for people with axial spondyloarthritis (axSpA) and is predictive of loss of function. Traditional measures such as the BASMI fail to capture many elements of spinal mobility and lack responsiveness to change. Inertial Motion Unit (IMU) sensors can be used to accurately measure spinal movement without requiring significant operator expertise.
Objectives: The primary objective of this study was to test the reliability of these new tools in patients with axSpA and to develop a new composite spinal mobility index. Our secondary objective was to apply the OMERACT ‘Truth’ filter to evaluate the new index for bias, for clinical relevance, and for convergent validity with existing measures 
Methods: Patients with axSpA fulfilling ASAS classification criteria were recruited. ViMove system was used to obtain ROM by attaching two IMU sensors at the cervical (Occiput-T3) and lumbar spine (L1-Sacrum). Intra-rater, inter-rater and test-retest reliability of IMU tests were assessed by intraclass correlation coefficients (ICC). The maximum range of movement for anterior flexion/extension (AFE), lateral flexion (Left+Right) and rotation (Left+Right) were obtained for the lumbar and cervical region. These six values were used in a composite score (IMU-ASMI) which referenced equivalent ROM values from normal subjects in an earlier criterion validity study. Pearson correlation coefficients with BASFI were calculated for each component as well as the overall score.
Results: The study included 40 patients (12 females, 28 males) with a mean age of 48 (27-41). Subjects had a wide range of severity of axSpA. The mean BASMI was 4.8 (range 0.7 to 8.2, SD 1.9). The mean IMU-ASMI was 4.0 (range 0.1-9.2, SD 2.1). The sensor based measurements had good to excellent reliability ( Table 1 ) and correlated closely with BASMI (r=0.79). The mean BASFI was 4.6 and the IMU-ASMI correlated closely with BASFI (r=0.71).Face Validity: Each IMU test presents spinal movement in angles and can also be represented as a normalized severity index analogous to BASMI. The mean cervical and lumbar IMU-ASMIs were 3.5 and 4.4 units, respectively.
Construct Validity: Do IMU movements correlate with their corresponding traditional measurements? As expected, the closest correlations were between IMU and goniometer cervical rotation (r=0.85) and between IMU and tape measure lumbar side flexion (r=0.84). Correlations between Schober’s test and IMU lumbar AFE and between tragus to wall and IMU cervical FE were moderate (r=0.62, 0.65).Do IMU movements correlate with BASFI? Correlation coefficients were as follows: lumbar AFE -0.57; rotation -0.59; side flexion -0.45; cervical F/E -0.55; rotation -0.61; side flexion -0.39. BASFI correlations with BASMI were comparable.
Content Validity/Comprehensiveness: No major ceiling or floor bias issues were found in the composite indices. Intermalleolar distance (BASMI) represents hip rather than spinal mobility, but it correlates with BASFI and is not in the IMU-ASMI. IMU-ASMI includes lumbar rotation which accounts for 27% (0-53%) of the lumbar mobility score.Table 1Results of ICC for different IMU spinal mobility testsCervicalLumbarMovementAFELatRotAFELatRotIMU- ASMIIntra-rater0.950.870.970.890.840.760.96Inter-rater0.960.960.980.940.940.770.98Test-retest0.880.960.850.910.940.810.97
Conclusion: IMU sensors can be used by non-experts to accurately and reliably measure spinal mobility in patients with axSpA. Lumbar rotation is an important new outcome measure.
REFERENCES:[1] Tugwell P, et al. J Rheum 2014 41(5):1000-1004
Acknowledgement: This study was funded by FOREUM (www.foreum.org)Disclosure of Interests: Philip Gardiner Grant/research support from: Educational support to attend rheumatology conferences from AbbVie, Abbott, MSD, Novartis, Amgen, Pfizer, Roche, UCB., Speakers bureau: Menarini, Dawn Small Employee of: Medical representative - Farmitalia Carlo Erba, Pharmacia, Pharmacia & Upjohn, Yamanouchi/Astellas., Paid instructor for: Medical representative - Farmitalia Carlo Erba, Pharmacia, Pharmacia & Upjohn, Yamanouchi/Astellas., Pedro Machado Consultant for: Abbvie, BMS, Celgene, Janssen, MSD, Novartis, Pfizer, Roche and UCB, Speakers bureau: Abbvie, BMS, Celgene, Janssen, MSD, Novartis, Pfizer, Roche and UCB, Karla Munoz Esquivel: None declared, Joan Condell: None declared, Antonio Cuesta-Vargas: None declared, Juan L. Garrido-Castro: None declared