Hypertriglyceridemia

Description

Hypertriglyceridemia risk is the likelihood that the user has abnormally high triglyceride levels (defined as being above 1.7 mmol/L or 150 mg/dL) and corresponds to the percentage of people with the user’s risk profile who have abnormally high triglyceride levels. BitDoctor determines the user’s hypertriglyceridemia risk based on blood flow patterns and demographic information.

Participants

6,414 adults (18+ years of age) recruited at a hospital health clinic in China.

Additional Data Collection

Subjects also received a blood test at the same clinic visit where their triglyceride levels were measured.

Modelling Approach

Blood flow signal was extracted and processed from facial video, and then blood flow features were extracted from blood flow signal.

Feature selection was carried out on blood flow and demographic features to identify features predictive of hypertriglyceridemia. A machine-learning based classifier was created to predict whether an individual has hypertriglyceridemia based on those features.

The model was created/trained using 5,132 subjects. The distribution of sex and hypertriglyceridemia status in this group is summarized in Table 19.

Subjects in videos	Composition
Sex distribution	48.7 % Male; 51.3% Female
Hypertriglyceridemia distribution	35%

Table 19: Hypertriglyceridemia distribution of training set

Model Performance

The Hypertriglyceridemia Risk model was then validated for accuracy on an independent portion of the dataset (n=641) that was not used in training - validation set. Then the final accuracy was obtained on an independent portion of the dataset (n=641) that was not used in training - test set. The distribution of sex and hypertriglyceridemia status in these groups is summarized in Table 20.

Subjects in videos	Composition in Validation	Composition in Test
Sex distribution	49.8% Male; 50.2% Female	47.3% Male; 52.7% Female
Hypertriglyceridemia distribution	35%	35%

Table 20: Hypertriglyceridemia distribution of validation and test sets

The accuracy on the test set calculated as area under the curve was 79.2% as sown in Figure 24.

Figure 24: AUC of Hypertriglyceridemia risk prediction

Predictions are displayed to the user as a percentage likelihood of having hypertriglyceridemia. A percentage of 45% or more suggests a risk of hypertriglyceridemia, while 55% or more suggests high risk of hypertriglyceridemia.