In reply
We agree with Polsky et al that differential loss to follow-up contributes to our finding that the percentage uninsured declined between 1992 and 2000, as we discussed in our article. Our analysis therefore underestimates the percentage uninsured during the later years of our study. In addition, we have found that individuals who experience a major decline in health are also more likely to be lost to follow-up (unpublished data, Health and Retirement Study, 1992-2000). These individuals may be more likely to lose health insurance as a result of their health decline, leading us to further underestimate the percentage uninsured during the time of the later interviews. Thus, the magnitude of the problem of lack of health insurance among the near elderly is even greater than our estimate, reinforcing the main message from our analysis: we must redouble efforts to expand access to health insurance for this vulnerable group.