Recent models for radon-induced lung cancer assume that at high levels of cumulative exposure, as experienced historically by many underground miners of uranium and other ores, the risk of lung cancer follows an inverse dose-rate (protraction enhancement) pattern. That is, for equal total dose, a greater risk is incurred by those whose total dose is accumulated at a lower rate over a longer duration than at a higher rate over a shorter duration. This inverse dose-rate effect is hypothesized to be the consequence of multiple traversals of the nucleus of a target cell by alpha particles. It has recently been concluded, however, that for low total doses, as in most residential settings, the inverse dose-rate effect should diminish and perhaps even disappear, since at very low doses the probability that more than one alpha particle would traverse a cell is small and there would be no possibility for interactions from multiple hits. Pooling original data from 11 cohort studies of underground miners, including nearly 1.2 million person-y of observation and 2,701 lung cancer deaths, we evaluate the presence of an inverse dose-rate effect and its modification by total dose. An inverse dose-rate effect was confirmed in each cohort, except one, and overall in the pooled data. There also appears to be a diminution of the inverse dose-rate effect below 50 Working Level Months (WLM), although analyses were necessarily hampered by a limited range of exposure rates at low total WLM. These data support both the presence of an inverse dose-rate effect, as well as its diminution at low total dose. As a consequence, assessment of risks of radon progeny exposure in homes (on average 15-20 WLM for a lifetime) using miner-based models should not assume an ever-increasing risk per unit dose. Rather, it is more appropriate to apply risk models that take into account protraction enhancement and its diminution.