The objective of the study, conducted in the general wards of Flinders Medical Center in South Australia, was to determine whether the risk of needlestick injuries from hypodermic needles could be reduced by the use of needle covers that are applied with a one-handed technique immediately after use of the needle. Baseline rates of needlestick injury were collected for a year prior to the use of needle covers, and incident cases of needlestick injury were collected after needle covers were introduced. The odds ratio, derived from pre- and post-intervention data, for injuries over which the use of needle covers could have had no influence, was used to show that the net effect of other influences on needlestick rates was not significant. The use of needle covers was associated with a significant reduction in needlestick injuries to nurses. The odds ratio for needlestick injury for nurse injuries that needle covers could not have prevented was 0.92. (95% CI 0.40-2.03). The odds ratio for needlestick injury for nurse injuries that needle covers could have prevented was 0.40 (95% CI 0.18-0.82). Although the odds of ratio for needlestick injury to cleaners was 0.19, the 95% exact confidence interval was 0.00-1.64.