Internal construct validity of the Rivermead Post-Concussion Symptoms Questionnaire

Area of Science:

• Neuroscience
• Psychometrics

Background:

• Mild traumatic brain injury (mTBI) can lead to persistent post-concussion symptoms.
• The Rivermead Post-Concussion Symptoms Questionnaire (RPQ) is commonly used to assess these symptoms.
• Internal construct validity of the RPQ in mTBI populations requires rigorous evaluation.

Purpose of the Study:

• To investigate the internal construct validity of the RPQ.
• To apply Rasch analysis to a national cohort of mTBI patients.
• To assess the RPQ's suitability for measuring post-concussion symptoms in this population.

Main Methods:

• Rasch analysis using the partial credit model was performed on data from 2,523 mTBI patients.
• Analysis included rating scale structure, local dependency, differential item functioning (age, gender), dimensionality, and model fit.
• Item and category modifications were explored to improve model fit.

Main Results:

• Rating scale categories were not consistently functioning; collapsing categories improved threshold ordering.
• Local dependency between items was identified and addressed by combining item pairs.
• The RPQ demonstrated no differential item functioning based on age or gender.
• Rasch analysis indicated multidimensionality (3+ dimensions), with the primary factor explaining 47.7% of the variance.
• Person targeting was poor (mean measure -2.16), with person reliability at 0.71 and a person separation index of 1.56.

Conclusions:

• The RPQ may not be optimal for assessing post-concussion symptoms in mTBI patients.
• Unidimensionality was not achieved, questioning the validity of summing RPQ item scores.
• The RPQ is unbiased concerning age and gender, offering some utility.
• Further refinement or alternative instruments may be necessary for this population.