Therese Ljungquist ^A1, Irene B. Jensen ^A2, Åke Nygren ^A2, Karin Harms-Ringdahl ^A3
A1 Neurotec Department, Division of Physiotherapy and the Department of Clinical Neurosciences, Section for Personal Injury Prevention, Karolinska Institutet, Stockholm, Sweden
^A2 Department of Clinical Neurosciences, Section for Personal Injury Prevention, Karolinska Institutet, Stockholm, Sweden
^A3 Neurotec Department, Division of Physiotherapy, Karolinska Institutet, Stockholm, Sweden

DOI: 10.1080/16501970306117

Objective: The aim of this study was to investigate the construct validity of 6 physical performance tests that had already been shown to have acceptable repeatability. Design: Data were collected in a randomized controlled multi-centre study. Subjects: 126 women and 105 men sick-listed for spinal pain carried out the tests and provided personal and background data at inclusion in the study. Methods: One test measured stepping up onto and down from a stool, 2 measured lifting ability and 3 walking speed. Construct validity was examined by analysing the influence of some variables on test performance. Results: High-rated pain behaviour and perceived high pain intensity during testing or during the previous 4 weeks were connected with low test performance. Exercise twice a week was connected with high test performance. The test with the highest ability to detect disability in the women with lumbar pain was a lumbar lifting test, while for the men, it was a cervical lifting test. The test with the highest ability to detect disability in the participants with neck pain was the cervical lifting test in addition to a gait test with burden for the women. Conclusion: Back pain hampered the test performance more than neck pain. Impairments and activity limitations expressed by the patient should guide the choice of test.