This article is a form of thinking about the future properly regarded as conditional forecasting. It begins by reminding readers of the enormous changes in American medicine since World War II. The second part revisits critically an earlier effort at conditional forecasting for 1995 that Paul Starr and I published in the early 1980s. Besides reviewing the prescience of our forecasts, the second part outlines the earlier trends in progress we identified and the four combinations of political and economic settings we explored. On that basis, the final part takes up the challenge of anticipating sensibly some possible medical futures in the America of the early twenty-first century, a task which excludes simple extrapolation.