Handling missing data through prevention strategies in self-administered questionnaires: a discussion paper

Area of Science:

• Epidemiology
• Health Services Research
• Data Science

Background:

• Self-administered questionnaires are cost-effective for large-scale data collection but often suffer from high rates of missing data.
• Missing data can compromise statistical power, representativeness, and generalizability of research findings.
• While imputation methods exist, they have limitations including potential underestimation of effects and reduced statistical power.

Purpose of the Study:

• To identify effective strategies for preventing missing data in self-administered questionnaires.
• To discuss the effects, methodological, and statistical considerations of these prevention strategies.

Main Methods:

• Review of existing literature on missing data prevention strategies.
• Analysis of strategies related to questionnaire administration format, follow-up procedures, and reminder systems.

Main Results:

• Electronic administration methods, such as tablets and email/telephone reminders, are associated with lower rates of missing data.
• These prevention strategies are relevant and feasible across clinical, nursing, and epidemiological research.

Conclusions:

• Prevention strategies offer practical solutions to mitigate missing data in research.
• Further research with robust designs is needed to validate these findings and establish consensus.
• Understanding and implementing these strategies is vital for collecting accurate patient and family data.