In a 33-month prospective analysis of needlestick injuries, venepuncturists working under Centers for Disease Control (CDC) guidelines for handling used needles were shown to incur a needlestick injury for every 3,175 to 4,006 needle-handling procedures. On the other hand, users of a simple device designed to reduce the risk of injury when recapping used needles were shown to incur a needlestick only once in every 16,100 venepunctures performed (P<0.00l). This represents a fourfold reduction in the rate of needlestick injuries. We thus question the effectiveness of the CDC nonrecapping policy.