Avoiding Mean Score Imputation in Patient Reported Outcomes Validation
Author Information
Author(s): Hardouin Jean-Benoit, Conroy Ronán, Sébille Véronique
Primary Institution: University of Nantes
Hypothesis
The study aims to compare different methods for handling missing data in the validation of Patient Reported Outcomes using a Rasch model.
Conclusion
The study concludes that several imputation methods, especially the Personal Mean Score, should be avoided as they can introduce bias in psychometric evaluations.
Supporting Evidence
- Imputation methods that do not consider the ability of the individual can lead to biased results.
- Using a random process in imputation methods can help reduce bias.
- Methods like NOIMP and LD showed less bias under MCAR and MAR conditions.
Takeaway
When researchers have missing answers from patients, they should be careful about how they fill in those gaps, because some methods can make the results look better than they really are.
Methodology
A simulation study was performed to evaluate the impact of sixteen different methods for handling missing values in the context of the Rasch model.
Potential Biases
Methods that do not incorporate a random process tend to overestimate psychometric quality.
Limitations
The study focused only on simple imputation methods and did not evaluate multiple imputation methods or other advanced techniques.
Participant Demographics
The study involved 500 individuals responding to a questionnaire.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website