Inter-rater reliability and cross-validation of lumbar stability test

Abstract

Background: Clinical lumbar stability test can be used to assess individuals’ ability to control their lumbar spine during limb movements. However, inter-rater reliability and validity has not previously been investigated. Therefore, this study aimed to (1) establish inter-rater reliability of lumbar stability test, and (2) cross-validate the lumbar stability test using lab-based equipment as a reference standard.

Methods: Twenty-nine healthy university students were recruited. Two fourth year physical therapy students performed lumbar stability test. Lumbar stability levels were used to determine inter-rater reliability using kappa statistics. Then, the participants were asked to perform two trials of 30-s seated balance tests on an unstable chair in eye-open and eye-closed conditions. These trunk neuromuscular control data, including 95% confidence ellipse area (CEA), root-mean-square error along X and Y axes (RMSEx and RMSEy, respectively), were used to cross-validate the lumbar stability test using Spearman’s correlation.

Results: Kappa statistics revealed poor to almost perfect agreement (0–1.00). When adjusting for prevalence and bias, inter-rater reliability were fair to almost perfect (0.31–0.59). CEA, RMSEx and RMSEy were greater in eye-closed condition (8.81, 0.30 and 0.48, respectively) comparing with eye-open condition (2.79, 0.16 and 0.26, respectively). Correlation between lumbar stability level and trunk neuromuscular control were fair (r = –0.31–0.34) in eye-open condition and weak (r = –0.14–0.18) in eye-closed condition.

Conclusion: Lumbar stability test should be refined to improve inter-rater reliability. Lumbar stability level does not completely represent trunk postural control. Clinicians should use this test with caution.

Keywords:

• Trunk postural control
• lumbar stability
• lumbar spine
• unstable chair
• centre of pressure
• confidence ellipse area

Acknowledgements

The authors would like to thank Motor Control and Neural Plasticity Laboratory, Mahidol University for providing data collection space and equipment. The authors also would like to thank Ms. Tanatta Chichakan for helping in data collection, and all subjects who participated in this study.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was funded in part by Thailand Research Fund (TRG5880133, 2015).