[Artificial neural networks applied to study allergic conjunctivitis screening questionnaire].

Area of Science:

• Ophthalmology
• Allergy and Immunology
• Biomedical Engineering

Background:

• Allergic conjunctivitis is a common condition in children.
• Accurate screening tools are needed for early diagnosis and management.
• Existing screening methods may have limitations in sensitivity and specificity.

Purpose of the Study:

• To evaluate the sensitivity and specificity of a screening questionnaire for allergic conjunctivitis.
• To compare diagnostic accuracy using multivariable analysis.
• To develop an artificial neural network (ANN) for improved future screenings.

Main Methods:

• An observational, cross-sectional study was conducted with 48 children diagnosed with allergic conjunctivitis and 54 controls.
• A screening questionnaire was administered to all participants.
• Multivariable statistical analysis and ANN modeling were performed.

Main Results:

• The ANN model achieved 100% accuracy in predicting allergic diagnosis using seven specific questionnaire items.
• Question five alone demonstrated good sensitivity (85.4%) and specificity (85.1%).
• The overall questionnaire showed low agreement with clinical examination (Kappa coefficient = 0.337).

Conclusions:

• An efficient, seven-question model was developed for screening allergic conjunctivitis.
• The ANN model shows promise for accurate, large-scale population screenings.
• This simplified model can facilitate easier application in public health settings.