Predictive Outcome Modeling

AI-powered recovery trajectory predictions

Model Accuracy: 87% (validated on 2,400+ patients)

Patient Profile - Jemma Wilson

Age

Stroke Severity

Moderate

Time Since Onset

2 months

Baseline SPI

0.42

Recovery Trajectory Prediction

24-week voice quality improvement forecast with 95% confidence intervals

Upper Bound (95% CI)
Lower Bound (95% CI)

Expected Improvement

82%

From baseline 45% to 82% in 24 weeks

Confidence Level

87%

Model accuracy on similar patients

Time to Goal

18wks

To reach 75% functional threshold

Personalized Milestone Timeline

Predicted recovery milestones based on current trajectory and similar patient outcomes

First Conversational Words

Week 4

85% confidence

Complete Simple Sentences

Week 8

78% confidence

Phone Conversations

Week 12

72% confidence

Social Interactions

Week 16

68% confidence

Functional Independence

Week 24

65% confidence

About This Prediction Model

Our AI model is trained on 2,400+ stroke recovery cases with validated outcomes. It analyzes demographic factors (age, gender, stroke severity, time since onset), baseline voice metrics (SPI, STD, quality scores), therapy parameters (intensity, adherence), and early session performance to predict individual recovery trajectories.

Model Accuracy

87% on validation set (n=480)

Mean Absolute Error

±6.2 percentage points

Last Updated

January 2026