Table 2. Effects of Low-Dose Aspirin on hCG-Detected Pregnancy Among Women With Adherence to the Assigned Treatment^a.

Method	Machine learning	Risk difference estimate (SE) [95% CI]	Risk ratio estimate (SE) [95% CI]
Intention-to-treat analysis	No	0.04 (0.03) [−0.01 to 0.10]	1.07 (0.04) [0.98 to 1.16]
Per-protocol analysis adjusted for baseline covariates
AIPW	Yes	0.08 (0.03) [0.03 to 0.14]	1.11 (0.04) [1.03 to 1.19]
TMLE	Yes	0.08 (0.03) [0.03 to 0.13]	1.10 (0.03) [1.04 to 1.17]
g-Computation	No	0.07 (0.03) [0.02 to 0.13]	1.10 (0.03) [1.02 to 1.17]
IPW	No	0.07 (0.03) [0.02 to 0.13]	1.10 (0.04) [1.02 to 1.18]
Unadjusted per-protocol analysis	No	0.08 (0.03) [0.03 to 0.14]	1.11 (0.04) [1.04 to 1.20]

Abbreviations: AIPW, augmented inverse probability weighting; hCG, human chorionic gonadotropin; IPW, inverse probability weighting; TMLE, targeted maximum likelihood estimation.

^a Assigned treatment was taking 5 of 7 pills (70%) per week during at least 80% of person-weeks of follow-up.