The results in Table 2 support Hypothesis 3 that the SRE values will continue to predict alcohol-related adverse outcomes as subjects age. Here, SRE5 values at T15 predicted more alcohol problems and a greater probability of carrying an AUD diagnosis over T20, T25, and T30 follow-ups, with similar results regarding the ability of T20 and T25 SRE5 to predict the subsequent 10- and five-year outcomes. Similar patterns were seen for maximum drinks, although these were only significant at T20 and T25. Predictions based on SRET were all significant and were greater in magnitude than for the SRE5. Thus, not only did a lower LR per drink as indicated by the SRE maintain its ability to predict more adverse alcohol outcomes as these men grew older, but, as was seen with cross-sectional relationships of SRE scores to alcohol-related parameters, the predictions became more robust with advancing age. The latter could reflect the shorter period of time inherent in the T20 and T25 predictions compared to T15, or the greater level of experience with alcohol that went into the SRE estimates with
