Clinical diagnosis of completeness of medical abortion by nurses: a reliability study in Mozambique.

BACKGROUND The provision of medical abortion continues to rely on routine use of ultrasound to confirm expulsion of pregnancy. However, the absence of ultrasound in most of the health facilities in developing countries and the additional training required to enable providers to use ultrasound is often prohibitive. The purpose of this study was to compare clinical history and physical examination with ultrasound in confirming completion of abortion. STUDY DESIGN A total of 718 women consented for medical abortion with misoprostol and were assessed for pregnancy expulsion by nurses and gynecologists. Nurses used history and physical examination while gynecologists used ultrasound to establish their diagnoses. RESULTS Nurses' clinical diagnoses for complete abortion, incomplete abortion and ongoing pregnancy were 83% (SE 0.01), 15% (SE 0.01) and 2% (SE 0.01), respectively. When gynecologists used ultrasound, the diagnoses for complete abortion, incomplete abortion, an ongoing pregnancy were 80% (SE 0.01), 17% (SE 0.01) and 3% (SE 0.01), respectively. Overall, nurses agreed with gynecologist diagnoses in 84% of cases, with a κ coefficient of 0.49 (SE 0.06) and chance-corrected first-order agreement (AC(1)) of 0.81 (SE 0.02). Agreement was very high for the diagnosis of complete abortion (AC(1) 0.89; SE 0.02), while it was moderate for ongoing pregnancy (AC(1) 0.58; SE 0.22) and incomplete abortion (AC(1) 0.45; SE 0.08). CONCLUSIONS Clinical history and physical examination alone, without the use of ultrasonography, are effective for the determination of successful pregnancy expulsion. However, greater emphasis is required on the clinical identification of ongoing pregnancy during any training of providers.