Evaluation of the dietary intake data coding process in a clinical setting: Implications for research practice

Area of Science:

• Nutrition Science
• Clinical Research Methodology
• Data Quality Assurance

Background:

• High-quality dietary intake data is crucial for establishing diet-disease relationships in clinical research.
• Source data verification (SDV) is a potential quality assurance method for dietary data.
• This pilot study evaluated SDV for dietary intake data coding in a clinical trial.

Purpose of the Study:

• To apply source data verification to assess the quality of dietary intake data coding in a clinical trial.
• To identify potential barriers affecting data quality during the dietary data coding process.

Main Methods:

• Source data verification was performed on 20 cases from a clinical trial.
• Data sources included diet history interview transcripts, paper forms, and nutrition analysis software outputs.
• Thematic analysis of in-depth interviews with dietitians was conducted.

Main Results:

• A total of 2024 discrepancies were identified, with the highest rate (57.49%) between interviews and software outputs.
• Discrepancies involved both food intake quantities and frequencies, particularly in the 'vegetable products and dishes' food group.
• Participant recall bias and potential subconscious bias in dietitians' coding were identified as contributing factors.

Conclusions:

• Dietary intake data accuracy is influenced by the required level of food detail.
• Enhanced participant support and supportive coding tools may improve data consistency and quality.
• This study presents a novel method for assessing dietary intake data coding errors, suggesting further investigation.