Sleep Apnea Prediction Model Developed for African American Patients

man being weighed
man being weighed
Researchers developed a prediction model for sleep apnea in African American patients, who have a high prevalence of sleep apnea that is often undiagnosed.

A new prediction model for sleep apnea in African American patients has been developed, according to a study presented at the 2018 SLEEP Meeting held June 2-6 in Baltimore, Maryland.

Researchers from Brigham and Women’s Hospital in Boston, Massachusetts, compared results of the Jackson Heart Sleep Study and the Stop-Bang questionnaire to determine predictive factors.

The researchers included linear and nonlinear demographics, adiposity, sleep duration, sleep symptoms, and comorbidities from 723 Jackson Heart Sleep Study participants as possible predictors (n=27). Considering all combinations of the 27 predictors, the researchers chose the models with the highest C-statistics (n=175).

After analysis and cross-validation, 16 predictors were chosen for the final prediction model: age, body mass index, male sex, habitual snoring, restless sleep, waist circumference, neck circumference, witnessed apneas, depression, age2, age3, waist2, waist3, neck2, neck3, and male sex*habitual snoring. Compared with the risk for sleep apnea predicted by the Stop-Bang questionnaire, the newly developed model improved prediction (0.14; 95% CI, 0.11-0.16).

“We found a model that included demographics, adiposity, sleep symptoms, and depression to better predict sleep apnea,” wrote the researchers. “Future studies should explore the generalizability of this predictive model among a wider population of African Americans.”

Related Articles


Johnson DA, Guo N, Sofer T, Wilson J, Redline S. Sleep apnea prediction model for African-Americans, the Jackson Heart Sleep Study. Presented at: SLEEP 2018; June 2-6, 2018; Baltimore, MD. Abstract 0332.