Rasch analysis of an outcome tool using data from questionnaires completed by patients with lumbar spinal stenosis.
To evaluate the psychometric properties of the Swiss Spinal Stenosis questionnaire using the Rasch measurement model.
Lumbar spinal stenosis (LSS) is a common cause of low back pain and leg symptoms in older patients. The SSS questionnaire was devised as a condition-specific outcome measure for patients with LSS. It comprises two subscales; one to measure symptom severity and another to measure physical functioning. The SSS is commonly used in trials for patients with LSS but has not, to our knowledge, been subjected to rigorous modern pschometric analysis methods.
Data from a total of 190 SSS assessments in 98 people with LSS were fitted to the Rasch measurement model. This method was used to examine the validity of the item scoring functions, the presence of item bias or differential item functioning and the fit of data to model expectations. In addition, the analysis was used to assess whether or not the item set in each scale formed a unidimensional scale to provide a valid summed score.
The symptom severity scale and the physical function scale of the SSS were found to be reliable and well targeted to the study population. However, the symptom severity scale was found to be multidimensional; three items focused on pain and four items on neuroischemic symptoms. The physical function scale performed well as a unidimensional scale, but needed to be modified by excluding one item in order for it to fit the Rasch model.
Development of a more robust measurement tool for this patient group may be warranted. A new tool may need to include three scales to measure the separate domains of pain severity, neuroischemic symptom severity, and physical function.
A Rasch analysis of the physical function scale and symptom severity scale of the Swiss Spinal Stenosis questionnaire indicated that the physical function scale performed well with the removal of item 11. However, the symptom severity scale demonstrated multidimensionality, with two separate constructs of pain and of neuroischemic symptoms.
*Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, Faculty of Medicine and Health, University of Leeds, United Kingdom;
†NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom;
‡Department of Rehabilitation Medicine, Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, Faculty of Medicine and Health, University of Leeds, United Kingdom.
Address correspondence and reprint requests to Alan Tennant, PhD, Department of Rehabilitation Medicine, Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, Faculty of Medicine and Health, D Floor Martin Wing, The General Infirmary at Leeds, Great George St, Leeds LS1 3EX, United Kingdom; E-mail: firstname.lastname@example.org
Acknowledgment date: March 17, 2010. Revision date: September 2, 2010. Acceptance date: September 7, 2010.
The manuscript submitted does not contain information about medical device(s)/drug(s).
Arthritis Research UK funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript.