4 Dec


We carried out a case-control study from 1996 to 1999 at Cook County Hospital and the University of Chicago Hospital. The Institutional Review Boards at both sites approved the study. Cases were 104 African-American mothers of VLBW infants (less than 1,500 g) who were recruited from the admission logbooks of the newborn intensive-care units of each hospital. Controls were 104 African-American mothers of normal birthweight (NBW) (greater or equal to 2,500 g); healthy infants (“healthy controls”); and 104 African-American mothers of NBW, sick infants who required assisted ventilation or other major life support (“sick controls”). The healthy controls were recruited from the labor and delivery logbooks and were chosen if admitted near the same date as the cases; sick controls were identified from NICU logbooks. Each case-control set came from the same hospital.

Maternal race was self-defined. We excluded mothers of twins, mothers of babies with congenital malformations, and mothers of babies who died shortly after birth. The study involved extensive in person interviews conducted in the hospital within three days after delivery. Informed consent was obtained from all mothers prior to the interviews, which were conducted by trained African-American women interviewers in privacy. The interview was based on a questionnaire that included questions in seven different categories: pregnancy; prenatal care and delivery; social and home environment; employment; discrimination and unfair treatment; pregnancy and stress; and racism, experience, and response. Interviewers also collected personal data, such as maternal age, marital status, education, household income, prenatal care, parity, and tobacco and alcohol use, during the current pregnancy. Mothers were given a stipend of $10.00 for their participation whether or not they completed the interview.

A modified “Life Events Inventory” was used to assess stress in pregnancy. Life events, such as serious health problems in the mother, death of a significant other, homelessness and job loss, among other things, were reviewed. The original questionnaire explored 24 possible stressful life events, out of which 19 items had sufficient response completeness in our data for analysis. We also asked about perceived racial discrimination at school, in getting a job, at work, in getting medical care, or in getting service at a restaurant or a store, using a format similar to that described by Krieger. Prenatal care was deemed poor if it started in the second or third trimester. Poor and no prenatal care were combined into one category.

Support in the delivery room was assessed by a series of questions during the interview and was defined as the presence of one or more person— chosen by the mother—present in the delivery room at her request. The mother’s relationship with her partner and relatives and aspects of emotional and practical support were explored through a series of 11 questions mostly taken from the Maternal Social Support Index of Pascoe. Three additional questions explored religious beliefs and practices.
canadian drugstore online

We used two control groups of NBW babies in order to evaluate the possibility of recall bias associated with having a newborn in intensive care. When the sick controls (normal BW infants in the NICU) were compared against the healthy baby controls, we found no difference for the key variables of interest. Thus, we combined all NBW infants for the subsequent analyses reported here.

Statistical analysis first examined the relationship of a number of previously reported risk factors to VLBW. Then the responses to each of the social support (including delivery room support) and religion questions were evaluated both for their direct association with VLBW and, in the case of delivery room support, for its possible ability to modify the effect of other risk factors, such as stressful life events. Confounding was evaluated by stratifica tion. Data analysis was accomplished using SAS for Windows, Version 6.12 and dEPID, version 2.11 (Centers for Disease Control).