Exploring the reliability of the modified rankin scale

Area of Science:

• Neurology
• Clinical Trials
• Biostatistics

Background:

• The modified Rankin Scale (mRS) is a primary outcome measure in stroke clinical trials.
• Variability in mRS grading can impact trial results and interpretation.
• Previous studies on mRS properties used varied methods, potentially not reflecting current trial populations.

Purpose of the Study:

• To investigate inter- and intraobserver variability of the mRS using a mock clinical trial design.
• To assess the impact of a structured interview on mRS grading reliability.
• To evaluate the predictability of mRS scores from initial patient observation.

Main Methods:

• A mock clinical trial design involving 100 paired mRS assessments by research nurses and stroke physicians.
• Randomized assignment to standard mRS assessment or a structured interview.
• Blinded reassessment by a second interviewer and video-recorded reevaluation after 3 months.

Main Results:

• Moderate interobserver agreement (k=0.57) and good intraobserver agreement (k=0.72) were observed.
• No significant difference in mRS grading between standard and structured interview methods.
• Initial patient observation was unreliable for predicting mRS scores (k=0.16).

Conclusions:

• Substantial interobserver variability in mRS grading exists even among experienced clinicians.
• Current training and structured interview methods are insufficient to eliminate mRS variability.
• Further strategies are necessary to enhance the reliability of the mRS in stroke research.