J Strote1, G Chen. 1. Division of Emergency Medicine, University of Washington Medical Center, Seattle, WA 98122, USA. strote@u.washington.edu
Abstract
OBJECTIVES: Pregnancy tests are often performed routinely for female emergency department (ED) patients of reproductive age. One major reason is a perception that patients are unreliable in predicting their own pregnancy status. We hypothesised that patients could reliably predict that they were not pregnant. METHODS: The study used a prospective cohort design, in an urban academic ED, from January 19 to May 19, 2004. All patients for whom a pregnancy test was ordered were asked about their sexual history as well as two additional questions: "Do you think you might be pregnant?" and "Is there any chance you could be pregnant?" Patients with already documented pregnancies were excluded. RESULTS: A total of 474 patients had pregnancy tests performed that met inclusion criteria. Eleven (2.3%) tests were positive. Among patients who answered no to both questions (337), one test (0.3%) was positive (negative predictive value (NPV) 99.7%, likelihood ratio (LR) 0.13 (95% CI, 0.02 to 0.82)). The other historical factor with a high NPV (100%) was not being sexually active (LR not calculable). All pregnancies occurred in patients with gastrointestinal or genitourinary as the chief complaint: this comprised only 56% of the presentations for which tests were ordered. CONCLUSION: Sexual history and self assessment can be used as a highly effective predictor of a patient not being pregnant. Given the risks of missed pregnancy in the ED, and low monetary and time cost of pregnancy tests, frequent testing is still recommended in most instances.
OBJECTIVES: Pregnancy tests are often performed routinely for female emergency department (ED) patients of reproductive age. One major reason is a perception that patients are unreliable in predicting their own pregnancy status. We hypothesised that patients could reliably predict that they were not pregnant. METHODS: The study used a prospective cohort design, in an urban academic ED, from January 19 to May 19, 2004. All patients for whom a pregnancy test was ordered were asked about their sexual history as well as two additional questions: "Do you think you might be pregnant?" and "Is there any chance you could be pregnant?" Patients with already documented pregnancies were excluded. RESULTS: A total of 474 patients had pregnancy tests performed that met inclusion criteria. Eleven (2.3%) tests were positive. Among patients who answered no to both questions (337), one test (0.3%) was positive (negative predictive value (NPV) 99.7%, likelihood ratio (LR) 0.13 (95% CI, 0.02 to 0.82)). The other historical factor with a high NPV (100%) was not being sexually active (LR not calculable). All pregnancies occurred in patients with gastrointestinal or genitourinary as the chief complaint: this comprised only 56% of the presentations for which tests were ordered. CONCLUSION: Sexual history and self assessment can be used as a highly effective predictor of a patient not being pregnant. Given the risks of missed pregnancy in the ED, and low monetary and time cost of pregnancy tests, frequent testing is still recommended in most instances.