No clinician would argue that the first consideration in deciding to offer a preventive service is that we do no harm. This credo would be easier to adhere to if all physicians had a crystal ball in their armamentarium. In addition to projecting future risk of cervical cancer, we need to know whether a woman is destined to die of another cause before she would develop cervical cancer. Older women are a remarkably heterogeneous group and many have very long life expectancies. At present, the rates of cervical cancer in previously screened women are low. However, there are several reasons to believe that this may not always remain the case. First, changing sexual practices may increase rates of HPV infection. Second, as women age, they have a lower cell-mediated immune capacity, which increases the probability that newly acquired HPV will persist. Third, in other countries, HPV prevalence has a bimodal age distribution, with a second increase after the age of 65 years. Thus, cervical cancer rates could rise in older women over the coming decades. As, unfortunately, we cannot predict the future, we are left with making policy based on average expected results for large populations. Cost-effectiveness analysis is one tool to assist in this process. Our analyses suggest that, on average, the biennial use of combined HPV and Pap testing remains beneficial until the age of 75 years at reasonable costs, and that lifetime screening still provides more benefits than harms and is within the bounds of acceptable expenditures.