Blinded gait assessment in idiopathic normal pressure hydrocephalus: reliability and correlation with clinical and patient-reported outcomes.

Abstract

Background: Idiopathic normal pressure hydrocephalus (iNPH) predominantly manifests with gait disturbances, yet clinical assessments are vulnerable to confirmation bias, particularly post-shunt surgery. Blinded video evaluations are a method to enhance objectivity in gait assessment, but their reliability has never been systematically investigated. The aim was to evaluate the inter-rater reliability of blinded gait assessments in iNPH patients and to investigate how these assessments correlate with the Hellström iNPH scale and patient-reported health status following shunt surgery.

Methods: Thirty-nine patients (mean age 75.5 years) diagnosed with iNPH between 2019 and 2023 were recorded performing Timed Up and Go (TUG) test before and after shunt surgery. Patients who required a walking aid were excluded. Four specialized raters, blinded to timepoint, evaluated gait pattern and graded improvement. Inter-rater agreement was quantified by Krippendorff's α; Spearman's ρ assessed correlations between graded improvement, Hellström iNPH scale changes, and EuroQol 5-Dimension 5-Level Visual Analogue Scale (EQ-VAS) differences.

Results: Agreement on video graded improvements was strong (α = 0.80, 95% CI: 0.76-0.84), whereas agreement on specific gait patterns was moderate (α = 0.53, 95% CI: 0.43-0.62). Graded improvement scores correlated moderately with changes in the Hellström iNPH scale (ρ = 0.67, p < 0.01) and showed fair correlation with EQ-VAS (ρ = 0.37, p < 0.01).

Conclusions: Blinded video assessments reliably captured postoperative gait improvements in iNPH and showed strong inter-rater agreement. While specific gait pattern ratings were less consistent, combining structured video scoring with clinical scales can improve outcome evaluation. More refined tools are needed to better detect subtle changes in gait and to reflect patient-perceived recovery.