Effectiveness of Fertility Awareness–Based Methods for Pregnancy Prevention

Rachel Peragallo Urrutia, Chelsea B. Polis, Elizabeth T. Jensen, Margaret E. Greene, Emily Kennedy, Joseph B. Stanford

To summarize best available prospective data on typical and perfect use effectiveness of fertility awareness-based methods for avoiding pregnancy. We conducted a systematic review of studies published in English, Spanish, French, or German by June 2017 in MEDLINE, EMBASE, CINAHL, Web of Science, and ClinicalTrials.gov. We reviewed 8,755 unique citations and included 53 studies that contained 50 or greater women using a specific fertility awareness-based method to avoid pregnancy, calculated life table pregnancy probabilities or Pearl rates, and prospectively measured pregnancy intentions and outcomes. We systematically evaluated study quality. Of 53 included studies, we ranked 0 high quality, 21 moderate quality, and 32 low quality for our question of interest. Among moderate-quality studies, first-year typical use pregnancy rates or probabilities per 100 woman-years varied widely: 11.2-14.1 for the Standard Days Method, 13.7 for the TwoDay Method, 10.5-33.6 for the Billings Ovulation Method, 4-18.5 for the Marquette Mucus-only Method, 9.0-9.8 for basal body temperature methods, 13.2 for single-check symptothermal methods, 11.2-33.0 for Thyma double-check symptothermal methods, 1.8 for Sensiplan, 25.6 for Persona, 2-6.8 for the Marquette Monitor-only Method, and 6-7 for the Marquette Monitor and Mucus Method. First-year perfect use pregnancy rates or probabilities among moderate-quality studies were 4.8 for the Standard Days Method, 3.5 for the TwoDay Method, 1.1-3.4 for the Billings Ovulation Method, 2.7 for the Marquette Mucus Method, 0.4 for Sensiplan, 12.1 for Persona, and 0 for the Marquette Monitor. Studies on the effectiveness of each fertility awareness-based method are few and of low to moderate quality. Pregnancy rates or probabilities varied widely across different fertility awareness-based methods (and in some cases, within method types), even after excluding low-quality studies. Variability across populations studied precludes comparisons across methods.