Materials and Methods

The Pregnancy Risk Assessment Monitoring System (PRAMS) is an ongoing, population-based surveillance system that gathers information on maternal behaviors and experiences before, during, and after pregnancy from selected states in the United States and New York City. Samples of women with recent live births are drawn from state birth certificates 2-6 months after delivery.

Data are collected by mailed questionnaires; nonrespondents are contacted by telephone. The PRAMS questionnaire in each reporting area includes core questions that appear on all PRAMS surveys and area-specific questions of interest. To produce data representative of the state birth population, data are weighted for sample design, nonresponse, and noncoverage. More detail on the PRAMS methodology has been published previously and is also available from the PRAMS web site ( http://www.cdc.gov/prams ).

We analyzed 2004-2008 data from 3 reporting areas (Missouri, New York state [excluding New York City], and New York City), the only PRAMS reporting areas that added questions on receipt of postpartum contraceptive counseling and specific contraceptive methods used postpartum to their core PRAMS surveys. Other PRAMS participating states did not collect this information.

To be included in the analyses, reporting areas must have achieved an overall weighted response rate of 65% or more for each year of data. Data were included for 2004-2008 for New York State, 2004-2007 for New York City, and 2007 for Missouri. The annual weighted response rates for these reporting areas during 2004–2008 ranged from 65% to 73%. The PRAMS project was approved by the Institutional Review Board of the Centers for Disease Control and Prevention.

Prenatal contraceptive counseling was measured by asking, “During any of your prenatal care visits, did a doctor, nurse, or other health care worker talk with you about any of the things listed below? Please count only discussions, not reading materials or videos.” One topic listed was birth control methods to use after my pregnancy.

Postpartum contraceptive counseling was measured by asking, “After your new baby was born, did a doctor, nurse, or other health care worker talk with you about using birth control?” To examine postpartum contraceptive use, women were asked, “Are you or your husband or partner doing anything now to keep you from getting pregnant?” and “What kind of birth control are you or your husband or partner using now to keep from getting pregnant?”

Respondents who answered no to the first question were classified as using no method and were not asked the second question, which included 13 response options for specific contraceptive methods and other. Respondents answering other were given the opportunity to write in a response; some responses were recoded into existing method options or new response options that were added (ie, implant, spermicide). PRAMS questions are cognitively and field tested prior to being included on the survey and evaluated every 3-4 years afterward.

More effective methods were defined as those with less than 10% of women experiencing an UIP within the first year of typical use based on published effectiveness rates and included tubal ligation, vasectomy, implant, intrauterine device, shot, pill, patch, or ring; this classification was chosen to be consistent with a prior report.

Less effective methods were those with 10% or more of women experiencing an UIP within the first year of typical use and included diaphragm, condoms, cervical cap, sponge, withdrawal, spermicide, or rhythm method. Women reporting the use of more than 1 method were classified as using the most effective of the multiple methods consistent with prior reports.

To focus on postpartum women at risk for UIP or short IPI, we excluded women who reported that they were currently pregnant (n = 70), were not sexually active at the time of the survey (n = 675), or had undergone a hysterectomy (n = 7). We also excluded respondents who answered yes to the postpartum contraceptive use core question and either did not respond to the question about the type of contraceptive method used (n = 418) or responded other to the question but the write-in response could not be recoded (n = 24).

We excluded women with missing data on the outcome or exposure variables of interest (n = 431) and women who did not receive prenatal care (n = 108) because these women did not have the opportunity to receive prenatal contraceptive counseling. Theoretically, all postpartum women had the opportunity to receive postpartum contraceptive counseling (eg, in the hospital prior to discharge), so we did not make exclusions based on the receipt of a postpartum care visit.

Separate multivariable logistic regression models were used to examine associations between prenatal, postpartum, and prenatal and postpartum contraceptive counseling and the use of any contraceptive method (yes vs no) and the use of a more effective contraceptive method (yes vs a less effective or no method). Based on a priori considerations, we adjusted for age group, race/ethnicity, marital status, education, type of insurance before pregnancy, pregnancy intention of the most recent live birth, number of previous live births, current breast-feeding, months since delivery, reporting area, and year.

To maximize the number of observations included in multivariable analyses, control variables with more than 2% missing data were recoded to include missing as a response category. We also adjusted for postpartum contraceptive counseling (when examining the effect of prenatal contraceptive counseling) and prenatal contraceptive counseling (when examining the effect of postpartum contraceptive counseling).

We examined variance inflation factors to rule out potential multicollinearity between our covariates. Because we suspected that women who received postpartum sterilization may have reported not receiving postpartum contraceptive counseling (eg, they received all counseling during pregnancy), we conducted a sensitivity analysis excluding these women (n = 863). In this subanalysis, we also examined the association between contraceptive counseling and use of long-acting reversible contraceptives (LARCs) classified as a polytomous outcome (LARC use and shot, pill, patch, or ring vs a less effective or no method).

Among the entire sample, we also examined whether the associations between receiving prenatal and postpartum contraceptive counseling and each outcome were modified by pregnancy intention or type of health insurance before pregnancy by testing interaction terms added to full models. We examined effect modification by type of health insurance before rather than during pregnancy because it is a better measure of usual insurance status because most US women have access to insurance during pregnancy. Where significant ( P < .05) effect modification was detected, stratum-specific estimates were calculated using contrast statements in a single model of the entire analytical sample. All analyses were performed on weighted data using SAS-callable SUDAAN (SAS Institute Inc., Cary, NC) to account for the complex survey design of the PRAMS.

Results

Among 11,306 women with a recent live birth in the 3 reporting areas, 10,520 (93%) were considered at risk for UIP or short IPI. Among those, 9536 (91%) were eligible for the current analysis. The majority (>50%) were aged 25-34 years, non-Hispanic white, married, and high school graduates; had private health insurance before pregnancy and at least 1 previous live birth; reported their most recent pregnancy was intended; had a postpartum check-up; were not currently breast-feeding; and were 4–12 months’ postpartum at the time of the survey ( Table 1 ).

Related to contraceptive counseling, 78% received prenatal counseling, 86% received postpartum counseling, and 72% received both ( Table 1 ). Most women (85%) reported using some method of contraception postpartum, with 53% using a more effective method (pills were most commonly reported) and 32% using a less effective method (condoms were most commonly reported).

Prenatal contraceptive counseling was significantly associated with using any vs no contraceptive method postpartum (87% vs 76% among those who did and did not receive prenatal counseling; adjusted odds ratio [AOR], 1.53; 95% confidence interval [CI], 1.29–1.82), and using a more vs less effective or no method postpartum (56% vs 39% among those who did and did not receive prenatal counseling; AOR, 1.51; 95% CI, 1.30–1.75) ( Table 2 ).

Stay updated, free articles. Join our Telegram channel

Join