Prediction of obstructive sleep apnea using visual photographic analysis

Area of Science:

• Sleep Medicine
• Anesthesiology
• Otolaryngology
• Internal Medicine

Background:

• Obstructive sleep apnea (OSA) is underdiagnosed and linked to severe perioperative risks.
• Current guidelines recommend OSA screening before surgery, yet screening rates remain low.

Purpose of the Study:

• To assess the accuracy of physicians in visually predicting OSA from photographs.
• To determine if prediction accuracy varies among anesthesiologists, otolaryngologists, and internists.

Main Methods:

• A prospective case series involving 56 patients undergoing polysomnography.
• Physicians (anesthesiologists, otolaryngologists, internists) rated patient photographs for OSA presence.
• Ratings were compared to polysomnography results (obstructive apnea hypopnea index [oAHI] ≥15).

Main Results:

• Overall prediction accuracy was 61.8% without comorbidity knowledge and 62.6% with it.
• Accuracy was significantly related to body mass index (BMI) but not age, gender, or race.
• No significant difference in prediction accuracy was found between physician specialties.

Conclusions:

• Visual prediction of OSA from photographs is only marginally better than chance.
• Physician specialty and knowledge of comorbidities did not influence prediction accuracy.
• Improved methods are needed for OSA screening in surgical patients.