Assessing Walking Performance in Patients with Lumbar Spinal Stenosis

Background
Individuals with Lumbar Spinal Stenosis (LSS) face challenges with walking. Interventions often aim to improve their daily walking performance. Consequently, assessing this performance effectively is essential in order to evaluate interventions thoroughly. Current methods lack validation, but ankle-worn step counters appear to be the most accurate for counting daily steps. This study aims to develop and evaluate an accelerometer-based method for assessing walking performance in individuals with lumbar spinal stenosis using criterion and construct validation.

Methods
The study consists of three parts, with each part building on the previous one.

1. Criterion Validation: Steps counted by five accelerometers worn in different body locations by individuals with LSS while performing a walking test were compared to directly observed steps. The walking test was performed in a standardized hospital environment. Factors like wear-site, walking aids, interruptions while walking, and non-walking activities were assessed for their impact on accuracy.

2. Construct Validation: We determined for how long time individuals with LSS could walk in a hospital corridor until stopped by their LSS symptoms. In addition, they wore a thigh-worn accelerometer and an ankle-worn step counter for a week during everyday life. We tested two hypotheses: a) the longest continuous walking period detected by the accelerometer should match their corridor walking ability, and b) the ankle-worn step counter should align with the thigh-worn accelerometer’s step count.

3. Construct Validation in the form of known-group Validity: One week of accelerometer data from individuals with LSS, individuals with Low Back Pain (LBP), and a background population were analyzed. Three variables combining the length and number of continuous walking periods per day; short (4-9 seconds), moderate (10-89 seconds), and extended (≥90 seconds) were drawn from the data to investigate differences in continuous walking between groups. We compared steps per day between groups to assess known-group validity.  

Results
- Ankle-worn accelerometers were the most accurate during the walking test but registered false steps while biking.
- Thigh-worn accelerometers provided reasonable step count during continuous walking. Patients reported habitual biking in everyday life, and since a thigh-worn accelerometer can adjust for false steps while biking, the thigh wear-site was chosen for the continued construct validation.
- The longest continuous walking periods recorded by the accelerometer during everyday life matched hospital corridor walking times.
- The total step count from the thigh-worn accelerometer and ankleworn step counter did not align, but the step count aligned during continuous walking periods.
- LBP participants and those from the background population had significantly higher steps per day and significantly more extended continuous walking periods than LSS participants.

Conclusion
A thigh-worn accelerometer seems to be able to count steps during daily continuous walking in people with lumbar spinal stenosis. The results support the construct validity of steps per day and the number of extended continuous walking periods per day for assessing walking performance in individuals with lumbar spinal stenosis. This lays the foundation for a valid and comprehensive assessment of walking performance in daily living, which can help inform on the best approach to achieve active aging in patients with lumbar spinal stenosis.