Can a Simple Measure of Stress Predict—and Improve—Treatment Outcomes?