Deriving modified Rankin scores from medical case-records

Area of Science:

• Neurology
• Clinical Research Methodology

Background:

• The Modified Rankin Scale (mRS) is a standard measure for assessing stroke outcome.
• Traditionally, mRS is determined through direct patient interviews (face-to-face or telephone).
• Alternative methods for deriving mRS from existing documentation are being explored.

Purpose of the Study:

• To investigate the feasibility of deriving the Modified Rankin Scale (mRS) from narrative content within patient case records.
• To evaluate the accuracy and reliability of case-record-derived mRS compared to traditional interview methods.

Main Methods:

• Two independent clinicians retrospectively assessed case records of 50 cerebrovascular patients.
• Clinicians were blinded to the original mRS assessment and rated their certainty using a 5-point Likert scale.
• Agreement between derived mRS and traditional face-to-face mRS was analyzed using attribute agreement analysis.

Main Results:

• Clinician-derived mRS from case records showed poor agreement with the standard assessment (kappa = 0.34).
• Inter-observer agreement for derived mRS was also low (kappa = 0.33).
• No correlation was found between the clinicians' certainty and the accuracy of their derived mRS grades.

Conclusions:

• Deriving accurate Modified Rankin Scale (mRS) scores solely from standard hospital case records is not reliable.
• Direct patient interviews are still necessary for precise mRS assessment, particularly in the context of clinical trials.