Sensitivity and specificity of a screening questionnaire for dry eye

Area of Science:

• Ophthalmology
• Epidemiology
• Biostatistics

Background:

• Dry eye disease (DED) is a prevalent condition requiring effective screening tools for large-scale epidemiological studies.
• The Dry Eye Epidemiology Projects (DEEP) study necessitates a reliable method for identifying potential DED cases.

Purpose of the Study:

• To develop and validate a screening questionnaire for identifying individuals with dry eye disease.
• To establish the diagnostic accuracy of the questionnaire for use in the DEEP study.

Main Methods:

• A 19-question questionnaire was developed, with 14 questions used for analysis.
• Stepwise multiple regression was employed on clinic data (77 cases, 79 controls) to create a discriminator function.
• Data was randomly split to estimate unbiased misclassification rates using a test set method.
• Receiver Operating Characteristic (ROC) curves were generated to assess performance.

Main Results:

• The final questionnaire demonstrated high sensitivity (60%) and specificity (94%).
• The developed discriminator function showed negligible bias in misclassification rate estimation.
• A secondary discriminator using the sum of responses yielded similar performance characteristics.

Conclusions:

• The Dry Eye Screening Questionnaire is a valuable and accurate tool for DED screening in epidemiological research.
• The questionnaire's performance suggests its utility in the DEEP study for efficient case identification.
• Further investigation is planned to assess generalizability to the broader population.