Contraceptive method

The survey measured the current contraceptive method by asking women to indicate the method of contraception used within the last month. If more than 1 choice was selected, participants were then asked, “Which method do you consider to be your primary method?” The outcome measure is the current use of a combined hormonal contraception, which was defined as combined birth control pills (containing both estrogen and progestin), the contraceptive patch, and the vaginal ring.

For women using birth control pills, the survey response choice was birth control pills and did not specify combined pills or progestin-only pills. To make this distinction, pharmacy claims data were obtained from Highmark Health and used to determine whether the most recent pharmacy claim prior to the woman’s baseline survey was for combined or progestin-only oral contraceptive. All other contraceptives were considered nonestrogen contraception, which included male and female condoms, the medroxyprogesterone acetate injectable, intrauterine devices, diaphragm, cervical cap, contraceptive foam, jelly, cream or suppository, the sponge, contraceptive implant, rhythm or natural family planning, withdrawal, and no method of contraception.

Medical contraindications to combined hormonal contraception

We identified study participants with the following category 3 (theoretical or proven risk usually outweigh the advantages) and category 4 (unacceptable health risk) contraindications to combined hormonal contraception, according to the Centers for Disease Control and Prevention Summary Chart of US Medical Eligibility Criteria for Contraceptive Use : (1) hypertension (category 3 if adequately controlled or category 4 if poorly controlled), (2) smokers over the age of 35 years (category 3 if <15 cigarettes/day or category 4 if ≥15 cigarettes/day), (3) a history of venous thromboembolism (category 3 if lower risk for recurrent venous thromboembolism or category 4 if higher risk for recurrent venous thromboembolism), (4) diabetes with complications (category 3 if microvascular complications or category 4 if vascular disease or diabetes >20 years), (5) coronary artery disease (category 4), (6) systemic lupus erythematosus with antiphospholipid antibodies (category 4), (7) breast cancer (category 3 if previous breast cancer with no evidence of disease for 5 years or category 4 for current breast cancer), and (8) migraine with aura (category 4).

Health conditions were determined by a series of questions asking, “Has a doctor, nurse, or other health professional ever told you that you had or have any of the following?” Smoking was assessed by asking, “Do you now smoke cigarettes every day, some days, or not at all?” In several cases, our survey tool was unable to distinguish category 3 from category 4 contraindications. For example, women over the age of 35 years who smoked every day were considered to have a contraindication, but we were unable to specify whether it was category 3 or 4 because the number of cigarettes smoked per day was not ascertained.

Of note, not all 18 category 3 and 4 health conditions were ascertained by the survey, so if women in the sample had other category 3 and 4 contraindications (severe liver cirrhosis, gallbladder disease, liver tumors, peripartum cardiomyopathy, organ transplant, thrombogenic mutations, and valvular heart disease), they may have been misclassified.

Contraceptive provider characteristics

Contraceptive provider characteristics were determined by a series of questions in which the respondent was asked about “the most recent health care visit where you received any contraceptive or women’s health care services.” The clinic type (private office vs other clinic), specialty (obstetrics and gynecology vs other specialties), and provider type (physician vs nonphysician provider) in which the contraceptive or women’s health care services occurred were recorded.

Statistical analysis

All variables were summarized with frequencies and percentages for categorical variables or with means, medians, and SDs for continuous variables prior to any analysis. χ ² tests for general association, with Fisher exact tests substituted as needed, were used to compare combined hormonal contraception use, contraceptive health provider characteristics, and sociodemographics between women with and without contraindications to estrogen.

This same approach was then applied to only the subgroup of 130 women with contraindications to combined hormonal contraception, making bivariate comparisons of health provider characteristics and sociodemographics between women with combined hormonal contraception use and women without combined hormonal contraception use. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).