use and other safe sexual practices, with optimal contraceptive protection accorded secondary priority.

The characteristics of the women in international samples differed too substantially from country to country to permit easy generalizations. In the United States, the 1995 National Survey of Family Growth points to two groups of Norplant users. The largest consists of predominantly young, single, minority women of lower socioeconomic status and educational levels, with one or more children, less likely to live in rural areas or in the northeastern portion of the country, and using the method primarily for spacing. A smaller group of older women of higher parity appears to be adopting the method as a long-term reversible alternative to tubal ligation.

Contraceptive implants produce side effects for many women, as described in the product labeling. By far the most common are changes in menstrual patterns, predominantly prolonged or irregular menstrual flow or increased bleeding. These tend to be frequent during the first 6 to 9 months of use, stabilizing by the end of the first year at a level that becomes acceptable to a majority of continuing users. The method also has non-menstrual side effects which manifest with different frequency and relative importance in different populations and at different time points, but primarily include headache, vaginal discharge, weight gain, acne, pelvic pain, and mood alterations. The studies presented suggest that the simple presence of menstrual side effects does not reliably predict decisions to continue or discontinue implant use, but there are differences between continuers and discontinuers with respect to non-menstrual side effects. Discontinuation rates associated with non-menstrual side effects seem higher than those for menstrual side effects but no single side effect, menstrual or non-menstrual, is consistently associated with decisions to discontinue use.

In studies to compare method use, implant continuation rates tend to be high relative to those of other reversible contraceptives. The overall pattern is that continuation rates are generally high through the first 2 years of use and not strikingly dissimilar from sample to sample, except in those studies that found discontinuation correlated with negative media coverage. Although there are great differences by country and although the data for the United States are scanty (partly because of low utilization), by the end of Norplant's approved 5-year term of use, approximately one-half of those who originally chose it were