BACKGROUND Measures of contraceptive effectiveness combine technology and user-related factors. Observational studies show higher effectiveness of long-acting reversible contraception compared with short-acting reversible contraception. Women who choose long-acting reversible contraception may differ in key ways from women who choose short-acting reversible contraception, and it may be these differences that are responsible for the high effectiveness of long-acting reversible contraception. Wider use of long-acting reversible contraception is recommended, but scientific evidence of acceptability and successful use is lacking in a population that typically opts for short-acting methods. OBJECTIVE The objective of the study was to reduce bias in measuring contraceptive effectiveness and better isolate the independent role that long-acting reversible contraception has in preventing unintended pregnancy relative to short-acting reversible contraception. STUDY DESIGN We conducted a partially randomized patient preference trial and recruited women aged 18–29 years who were seeking a short-acting method (pills or injectable). Participants who agreed to randomization were assigned to 1 of 2 categories: long-acting reversible contraception or short-acting reversible contraception. Women who declined randomization but agreed to follow-up in the observational cohort chose their preferred method. Under randomization, participants chose a specific method in the category and received it for free, whereas participants in the preference cohort paid for the contraception in their usual fashion. Participants were followed up prospectively to measure primary outcomes of method continuation and unintended pregnancy at 12 months. Kaplan-Meier techniques were used to estimate method continuation probabilities. Intent-to-treat principles were applied after method initiation for comparing incidence of unintended pregnancy. We also measured acceptability in terms of level of happiness with the products. RESULTS Of the 916 participants, 43% chose randomization and 57% chose the preference option. Complete loss to follow-up at 12 months was <2%. The 12 month method continuation probabilities were 63.3% (95% confidence interval, 58.9–67.3) (preference short-acting reversible contraception), 53.0% (95% confidence interval, 45.7–59.8) (randomized short-acting reversible contraception), and 77.8% (95% confidence interval, 71.0–83.2) (randomized long-acting reversible contraception) (P < .001 in the primary comparison involving randomized groups). The 12 month cumulative unintended pregnancy probabilities were 6.4% (95% confidence interval, 4.1–8.7) (preference short-acting reversible contraception), 7.7% (95% confidence interval, 3.3–12.1) (randomized short-acting reversible contraception), and 0.7% (95% confidence interval, 0.0–4.7) (randomized long-acting reversible contraception) (P = .01 when comparing randomized groups). In the secondary comparisons involving only short-acting reversible contraception users, the continuation proba...

II-1.