TABLE 3.

Logistic regression models of interest in over-the-counter progestin-only pill use among an online sample of transgender, nonbinary, and gender-expansive people assigned female or intersex at birth in the United States

Independent variable and adjustment set to control for confounding	Interested in over-the-counter progestin-only pill usea
	Odds ratio	95% confidence interval	P value
Model 1: age (y) (n=1415), no adjustment was necessary to estimate the total effect of this variable on interest in over-the-counter progestin-only pill use
18–24	1.67	1.33–2.10	<.001b
25–34 (ref)
35–44	0.85	0.60–1.21	.37
45–49	0.49	0.22–1.11	.09
Model 2: education level (n=1381), adjusted for age and race and ethnicity
High school degree or less	3.02	1.94–4.71	<.001b
Some college, trade, or technical school	1.04	0.78–1.39	.77
College degree (ref)
Some graduate or professional study	1.12	0.73–1.73	.60
Graduate or professional degree	0.70	0.51–0.96	.03b
Model 3: currently employed (n=1383), adjusted for education
Yes	0.97	0.75–1.26	.84
No (ref)
Model 4: currently a student (n=1396), adjusted for age
Yes	1.26	0.98–1.61	.07
No (ref)
Model 5: marital status (n=1387), adjusted for age and education
Single, never married	1.02	0.75–1.39	.88
Married, civil union, registered domestic partnership, or engaged (ref)
Divorced, widowed, separated	1.23	0.71–2.13	.46
Other	0.58	0.24–1.45	.25
Model 6: has health insurance (n=1,392), adjusted for employment, student status, and marital status
Yes (ref)
No or donť know	1.91	1.24–2.93	.003b
Model 7: considers self at risk for unintended pregnancy (n=1400), adjusted for age, marital status, and participant sterilization
Yes	1.33	0.96–1.85	.08
No or do not know (ref)
Model 8: current oral contraceptive use (n=1394), adjusted for marital status, participant sterilization, insurance status, and considering self at risk for future unplanned pregnancy
Yes	1.69	1.17–2.44	.005b
No or do not know (ref)
Model 9: ever progestin-only pill use (n=1413), adjusted for ever having been pregnant and wanting to avoid estrogen generally
Yes	2.32	1.70–3.17	<.001b
No or do not know (ref)
Model 10: is sterilized or has had tubes tied, ovaries removed, and/or uterus removed or other procedure that makes getting pregnant impossible (n=1385), adjusted for age, education, marital status, and ever having been pregnant
Yes	0.31	0.12–0.79	.01b
No (ref)
Model 11: ever use of testosterone for gender affirmation (n=1415), adjusted for age and wanting to avoid estrogen generally
Yes	0.72	0.57–0.90	.004b
No (ref)
Model 12: has ever used contraception for gender affirmation (n=1413), adjusted for age
Yes	1.31	0.93–1.86	.13
No or do not know (ref)
Model 13: Wants to avoid estrogen generally (n=1,415), adjusted for age and ever having used testosterone
Yes	1.32	1.04–1.67	.02b
No or do not know (ref)
Model 14: wants to avoid estrogen because they view it as a female or feminizing hormone (n=1413), adjusted for age and ever having used testosterone
Yes	1.72	1.36–2.19	<.001b
No or do not know (ref)
Model 15: ever felt that opinions about their gender identity and/or sexual orientation from healthcare staff have negatively impacted them in a healthcare setting (n=1305), adjusted for age, education, and region
Yes	1.08	0.86–1.36	.50
No or do not know (ref)

We used the directed acyclic graph approach¹⁶ to select covariates for each model. Supplemental Appendix B contains the directed acyclic graphs.

^aParticipants were considered interested in over-the-counter progestin-only pill use if they reported “yes” (vs “no” or “don’t know”) to a question asking if they would use a birth control pill that only had progestin that they could buy over the counter (without a prescription)

^bIndicates P<.05 in logistic regression.

Grindlay. Transgender interest in over-the-counter progestin-only pills. Am J Obstet Gynecol 2024.