Imputation of missing values within WHODAS 2.0 data collected from low back pain patients using the response function...

Area of Science:

• Rehabilitation medicine
• Psychometrics
• Health outcomes research

Background:

• The WHO Disability Assessment Schedule 2.0 (WHODAS 2.0) is a key tool for measuring disability.
• Missing data in surveys can affect the accuracy of disability estimates.
• The response function (RF) method is a potential imputation technique.

Purpose of the Study:

• To evaluate the impact of missing data and RF imputation on bias, precision, and reliability of WHODAS 2.0 disability estimates.
• To assess the performance of the RF method in handling missingness in WHODAS 2.0 data.

Main Methods:

• A simulation study using data from 284 low back pain patients.
• Rasch modeling to analyze person and item parameters.
• Imputation of missing data using the response function method.

Main Results:

• The RF method showed minimal bias and maintained precision in disability estimates.
• Increasing missing rates slightly reduced the Person separation index but it remained high (>0.94).
• Cronbach's alpha values averaged 0.99, indicating high reliability, though with increased variation.

Conclusions:

• The response function method is a reliable approach for imputing missing WHODAS 2.0 data, preserving construct validity.
• Imputation of missing items in domain 5 (work/school activities) using RF demonstrated satisfactory reliability for disability estimation.
• The RF method offers a valuable tool for clinicians and statisticians in managing missing data in disability assessments.