OBJECTIVES: Our objective was to assess the predictive validity of the Adherence Estimator® (AE), a three-item instrument designed to estimate patients’ propensity to adhere to prescription medications for chronic disease. METHODS: The AE was part of a larger survey mailed to adults who had an index prescription fill for one of five chronic diseases. Persistence over time was assessed using the continuous measure of medication gaps (CMG) from pharmacy claims data. The Wilcoxon rank sum test was used to assess differences in median PDC between pairs of the adherence risk groups (low/medium/high risk). Sensitivity, specificity, and positive predictive value (PPV) of the AE in the prediction of non-persistence were calculated at 45 days, and three, six, nine, 12, and 14 months. Multiple linear regression was used to assess whether the AE was a significant predictor of persistence at different time intervals, controlling for demographics and comorbidity. RESULTS: There were 1,674 usable responses for an overall survey response rate of 24.3%. Overall, 42.4% of the respondents were classified as being at low risk for non-adherence; 35.1% at medium risk; and 22.5% at high risk. At 14 months, median therapy gaps (CMG) for the low risk group (38.2%) were significantly lower than those for the medium-risk (48.7%) and high-risk groups (59.1%). At 14 months, sensitivity, specificity, and PPV for the AE were 61.1%, 50.5%, and 73.9%, respectively. The sensitivity and the specificity of the AE remained constant over time; the PPV increased as more respondents fell off therapy. Starting at three months, the AE risk groups (low vs. medium/high risk) significantly predicted subsequent gaps in therapy as measured by the CMG. CONCLUSIONS: Patients’ propensity to adhere to prescription medications for chronic disease as measured by the AE significantly predicted patients’ adherence behavior from three to 14 months post-index fill.