Predictive Models for Opioid and Substance Use Disorder Treatment

Stability Scores

Urine Tests Used to Fit Model

> 100,000

Patients Used to Fit Model

> 7,500

Prediction Accuracy

Opioid and Medication Use

89%

Non-Opioid Substance Use

86%
Model
AUC
Accuracy
Sensitivity
PPV
Specificity
NPV
OUD Stability Scores
0.88
0.89
0.96
0.90
0.57
0.77
Non-Opioid Use
Drug Use score
0.92
0.86
0.85
0.84
0.86
0.87

Treatment Twin Prognostic Models

SUDx’s Treatment Twin™ models represent the most likely paths through treatment (i.e., treatment trajectories) observed for groups of patients receiving medication for opioid use disorder (MOUD) treatment.

SUDx’s Treatment Twin prognostic models predict a single patient’s most likely path through treatment after just a single appointment.

Urine Tests for Model Fit

> 75,000

Patients for Model Fit

> 3,000

Unique Patient Treatment Events

> 3,500

Accuracy and Performance of Models

Our prognostic models are able to identify patients who are at the highest risk for drug use and medication non-adherence with 89% accuracy after only their first appointment.

Our models rapidly improve as more data is collected on each patient, reaching an accuracy of 92% by appointment two and 98% accuracy by appointment nine.
* Red Line = model accuracy. As more data is collected after each attended appointment, our models quickly "hone-in" on a patient's most accurate treatment twin

** Note: the following results only describe the performance of predicting subgroup 4. I think this is probably the most important group to providers etc. and it just happens to have the best performance out of the 4 TT’s

First Appointment

89%

Second Appointment

92%

Ninth Appointment

98%
Area Under the Curve (AUC) and Area Under the Precision Recall Curve (AUPR) after each attended appointment
Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (PPV) after each attended appointment

Get In Touch

Connect with the SUDx team to learn more about our toxicology services.

Request More Information
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.