Estimating measurement error of the Oswestry Disability Index with missing data

Area of Science:

• Orthopedics
• Rehabilitation Medicine
• Health Outcomes Research

Background:

• The Oswestry Disability Index (ODI) is a critical patient-reported outcome measure for low back pain.
• Incomplete ODI questionnaires can lead to significant measurement error and misclassification of patient disability levels.

Purpose of the Study:

• To compare the accuracy of alternative scoring (AS) and multiple imputation (MI) in estimating measurement error with missing ODI items.
• To assess the precision of AS and MI in scoring the ODI and classifying disability with increasing missing data.

Main Methods:

• Simulated 1000 datasets for 1-9 missing ODI items, calculating scores using AS and MI.
• Determined absolute percentage error (APE) and misclassification rates for each method.
• Compared APE between AS and MI to evaluate precision and misclassification.

Main Results:

• Alternative scoring (AS) demonstrated lower absolute percentage error (APE) and misclassification rates compared to multiple imputation (MI).
• For 9 missing items, AS had 12% APE and 13% misclassification, while MI had 56% APE and 58% misclassification.
• Measurement error and misclassification increased with more missing ODI items for both methods.

Conclusions:

• Alternative scoring (AS) is a more precise method than multiple imputation (MI) for handling missing data in the Oswestry Disability Index.
• Clinicians must consider the impact of missing ODI data and choose appropriate scoring methods to avoid misinterpretation.
• Accurate interpretation of ODI scores is crucial for effective patient assessment and treatment planning in low back pain management.