Skip to main content
Fig. 3 | BMC Medical Informatics and Decision Making

Fig. 3

From: Dirichlet process mixture models to impute missing predictor data in counterfactual prediction models: an application to predict optimal type 2 diabetes therapy

Fig. 3

Predictive distributions for missing data and treatment response conditional on the scenario for a patient at 6-months. Unobserved true values are given by the black vertical line and its prediction by the grey distribution. As more variables become known, the uncertainty in the predictions decreases. Scenario 1: Past Drugs, Current Drugs, Age known. Scenario 2: Past Drugs, Current Drugs, Age, HbA1c known. Scenario 3: Past Drugs, Current Drugs, Age, HbA1c, BMI known. Scenario 4: Past Drugs, Current Drugs, Age, HbA1c, BMI, eGFR known. Scenario 5: All variables known. Patient: Number of Past Drugs [2], Number of Current Drugs [0], HbA1c [73], eGFR [68.9], ALT(log) [2.9], BMI [30.1], Age [75]

Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us