Objective: To compare the cost and complication rate of two alternative strategies for the diagnosis and medical management of ectopic pregnancy when ultrasound is nondiagnostic.
Design: A decision tree was constructed to compare  dilatation and curettage (D&C) followed by treatment of all ectopic pregnancies with methotrexate versus  empiric treatment of all patients with possible ectopic pregnancies with methotrexate without D&C.
Setting: University setting.
Patient(s): Ten thousand hypothetical women with nonviable pregnancies and a known incidence of ectopic pregnancy were entered into a computer model.
Main outcome measure(s): The two approaches were compared with respect to the number of missed ectopic pregnancies, complications, procedures performed, admissions to the hospital, and cost.
Result(s): The D&C group had 1% more failed managements of ectopic pregnancies and 13.4% fewer patients with a miscarriage undergo a second treatment for resolution. The D&C group had 13.7% fewer complications including 6.3% fewer hospitalizations. D&C costs $173 to $223 more than empiric use of methotrexate per patient.
Conclusion(s): Empirically treating women at risk for ectopic pregnancy with methotrexate does not reduce complications or save money. In the absence of such savings, the desire to make an accurate and definitive diagnosis, allowing objective prognosis on future fertility and risk of repeat ectopic pregnancy, supports the need to distinguish a miscarriage from ectopic pregnancy before treatment with methotrexate.