The “NIH-DC Initiative to Reduce Infant Mortality in Minority Populations” is a congressionally mandated research collaboration between four major academic institutions in Washington, DC (Children’s National Medical Center, Georgetown University, George Washington University, and Howard University), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), the National Center for Minority Health and Health Disparities, and RTI International. As part of this collaboration, a randomized controlled trial was conducted to evaluate the benefits of an integrated behavioral cognitive intervention delivered during routine prenatal care (PNC) in reducing cigarette smoking, ETSE, depression, and IPV during pregnancy and improving pregnancy outcomes. These analyses used data collected from this trial.
Women presenting to six community-based clinical sites serving minority women (African-Americans and Hispanics) in the District of Columbia were screened for recruitment between July, 2001 and October, 2003. Women were deemed demographically eligible if self-identified as belonging to a minority group, being ≥ 18 years, <29 weeks of pregnant, a DC resident and English speaking. Verbal pre-screening for other characteristics was discouraged by the study protocol in order to avoid bias. Demographically eligible women were invited and consented to participate in the screening. Participants were screened for the four risk factors using an audio-computer assisted self interview (A-CASI) which also confirmed their pregnancy status and demographic eligibility. See El-Khorazaty et al. for more details on recruitment and retention activities and results [29
]. After initial screening confirmed eligibility, mothers were scheduled for a telephone interview. More information on sociodemographics, reproductive history and behavioral risks was collected during a baseline interview, conducted on average at 19 weeks of gestation, nine days after A-CASI screening. Follow-up data collection by telephone interview occurred during the second and third trimesters of pregnancy (22–26 and 34–38 weeks EGA, respectively). Intervention and follow-up activities continued until July 2004.
For purposes of recruitment, gestational age (GA) was based on women’s reporting and verified by medical record abstraction. Upon delivery, infant’s GA was abstracted from the medical records prioritized in the following order: ultrasound, exam and menstrual history. Analysis for GA was based on the infants’ GA assessment. Following the baseline interview, consent to participate was obtained and women were randomized to the clinical trial.
The sample size was determined to ensure adequate statistical power to test the hypotheses that our cognitive-behavioral intervention would result in reductions in the targeted risks. Assuming a 5% level of significance, 80% power would allow the detection of 10–20% reductions in risk-specific factors among women in the intervention group from a 100% prevalence at recruitment time, a sample of 1,050 women needed to be retained at the end of the follow-up period. This number was estimated to be sufficient to detect a 25% reduction in PTB and LBW combined in the intervention group as compared to that for the usual care (estimated at 20%). Based on a declining birth rate in D.C., the recruitment period was extended four months to reach the required sample size.
A total of 2,913 women were screened and 1,398 met eligibility criteria. There were 1,070 women, who were reached for baseline telephone interview and consented to participate in the randomized trial. This number exceeded the required sample size of 1,050 needed for 80% power and 5% level of significance. These demographically eligible pregnant women reporting one of the four risk factors were randomized to the intervention group (IG) or the usual care group (UCG) after completing the baseline interview. Of the1,070 women, 1,044 were African-American and still pregnant at the time of the interview. Pregnancy outcomes data were available on 918, 88% of these participants. Mothers with a live birth represented 95% (n=870). Only mothers of a singleton live born with infant medical abstraction data available were included in these analyses (n=819, 94% of mothers with a live birth). This final number represented 89% of all women with a known pregnancy outcome. (See .)
Profile of Project DC-HOPE Randomized Controlled Trial
The intervention of this RCT was designed for delivery during PNC. The intervention was evidence-based and specific to the designated psycho-behavioral risks [30
]. The intervention for smoking and ETSE was based on Smoking Cessation or Reduction in Program Treatment (SCRIPT). This intervention is based on social cognitive theory [31
] and tailored to a woman’s stage of readiness for behavioral change [32
]. The intervention for depression was based on a cognitive behavioral theory program developed by Miranda and Munoz [33
] and proven successful with low income minority women seeking primary care services. The original group therapy model was adapted to an individual treatment format. The intervention focused on secondary prevention of symptoms of depression and emphasized strategies for mood management as well as establishing and maintaining positive social interactions. Sessions emphasized skill development towards revising negative cognitions. A structured intervention developed by Parker and colleagues [34
] and based on Dutton’s Empowerment Theory [35
] which emphasized safety behaviors, was adopted for the IPV intervention. This brochure based intervention provided information about the types of abuse, the cycle of violence, a danger assessment component, in addition to the development of a safety plan. Individualized counseling sessions provided an integrated approach to multiple risks responsive to a woman’s specific risk combination. Intervention sessions were conducted privately in a room proximate to the PNC clinics and occurred immediately before or after routine PNC, for an average of 35±15 minutes. The intervention was designed to be delivered during PNC visits with eight discrete sessions. The intervention specialists were trained to apply the content of these sessions during one or more visits as needed. In order to avoid influencing the utilization patterns of PNC (an outcome compared between IG and UCG), no effort was made to facilitate or encourage participation in routine PNC in either group. Delivery of the intervention was dependent on the mother’s decision to attend a scheduled PNC visit.
Site- and risk-specific block randomization to IG or UCG was conducted. Investigators and field workers were blinded to the block size. A computer generated randomization scheme considered all possible risk combinations within each of the recruitment sites. Recruitment staff at each site called in the details of the risk profile for a new recruit, and the assignment was generated centrally by the data coordinating center.
Validated instruments were used for the A-CASI screening, and the baseline and follow-up telephone assessments of women with respect to risk factors. Telephone interviewers and their supervisors were blinded to the participants’ randomization group. At the time of recruitment and upon delivery, maternal and infant medical records were abstracted and infant and pregnancy outcomes were recorded.
To preserve the randomization, participant data were analyzed according to their care group assignment, regardless of whether they received any intervention sessions, using an intent-to-treat approach. Statistical analyses were conducted using SAS version 9.1.3 (SAS Institute Inc., Cary, NC). Bivariate analyses compared women with and without adverse pregnancy outcomes (LBW, VLBW, PTB, and VPTB) with respect to sociodemographic characteristics and psycho-behavioral risks. Similar analyses compared baseline characteristics and outcomes among women assigned to IG vs. UCG. Chi-square tests and t-tests were used for these analyses. Bivariate analyses using exact chi-square tests compared reduction in the number of risk factors at baseline and the last follow up interview prior to delivery for women assigned to IC vs. UCG.
To evaluate the impact of sociodemographic and psycho-behavioral risks on adverse pregnancy outcomes, we used logistic regression to model LBW, VLBW, PTB, and VPTB. Unadjusted analyses were conducted to assess the impact of care group assignment on the four outcomes. Absolute risk reduction, percent reduction in risk and number needed to treat were calculated. Adjusted logistic regression models controlled for variables including maternal age, care group, depression, IPV and illicit drug use during pregnancy. Variables were selected for inclusion in the logistic regression models if they reached significance in the bivariate analyses, and predictors significant at p<0.05 were retained in the final models.