Presence and origin of variability of the pull test and push-and-release test in people with Parkinson's disease

The pull test and the push-and-release test evaluate postural instability in Parkinson's disease (PD). We systematically evaluated the impact of within- and between-assessor variability in test delivery by the clinician on the test outcome. We also evaluated whether using standardized treadmill-based mechanical perturbations may enhance the consistency of the patient's test outcomes.

Fifty persons with PD underwent a series of backward balance perturbations: three repetitions of both the pull test and the push-and-release test delivered by three different assessors (i.e., nine repetitions of each test), plus five standardized treadmill-induced perturbations at 1.5 m/s², in pseudo-random order.

We found substantial within-assessor variability on both manual tests. A difference in scores of 2 points or more was found in 30% of participants for the pull tests, and in 42% for the push-and-release tests. Similarly, large variability in scores was observed between assessors. Inconsistent test delivery was demonstrated by a wide range of sternum and center of mass displacements following the pull test and body inclination angles in the push-and-release test. Across five repeated treadmill-based perturbations at 1.5 m/s², ≥ 2 points difference in test outcomes was found in 18% of participants, with significantly greater consistency in sternum and center of mass displacements.

Variability in the patient's balance test scores can be attributed to substantial variability in test delivery, as well as inconsistent performance of the individual patient. Assessment of postural instability may benefit from standardizing test delivery, e.g., using treadmill-induced perturbations.