Is there a role for expectation maximization imputation in addressing missing data in research using WOMAC questionnaire? Comparison to the standard mean approach and a tutorial
2011

Using Expectation Maximization to Handle Missing Data in WOMAC Questionnaire

Sample size: 2062 publication Evidence: moderate

Author Information

Author(s): Ghomrawi Hassan MK, Mandl Lisa A, Rutledge John, Alexiades Michael M, Mazumdar Madhu

Primary Institution: Weill Cornell Medical College

Hypothesis

Does the Expectation Maximization (EM) imputation method provide better results than the standard mean imputation method for the WOMAC questionnaire?

Conclusion

The EM method is a more accurate alternative to the WOMAC imputation method, effectively creating a complete dataset for analysis.

Supporting Evidence

  • The EM method imputed scores for all subjects, while the WOMAC method did not.
  • Mean subscale scores were similar for both methods but the EM method showed more consistency with the true score.
  • The study demonstrated that the EM method is particularly beneficial as the number of items in a subscale increases.

Takeaway

This study shows that a new way to fill in missing answers in a health survey is better than the old way, helping researchers get more accurate results.

Methodology

The study analyzed WOMAC data from 2062 total hip replacement patients, comparing the EM method and the standard mean imputation method for handling missing values.

Potential Biases

The WOMAC method may produce biased estimates if the rate of missing values exceeds a certain threshold.

Limitations

The sample may not be representative of the total knee arthroplasty population, and the assumption of normality for the data may not hold true.

Participant Demographics

50% of participants were female, with a mean age of 62 years, and 92% were white.

Digital Object Identifier (DOI)

10.1186/1471-2474-12-109

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication