Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003
CHARLAN KROELINGER: Good morning, my name is Charlan Kroelinger. I’m from the University of South Florida and I’ll be talking about partner influence on women’s perceptions of pregnancy. First of all, we have the essential research question, does partner support impact certain stressors that occur during pregnancy, specifically pregnancy wantedness and physical and/or verbal abuse. How did we define partner support? Well we had two different measures. We had simply a measure of presence or absence. Was partner support present throughout the course of the pregnancy or was it absent? And then we used a scale to assess the magnitude of support if it were present. How did we assess the stressors? Well, we asked the question for women in their third trimester, “When you first found out you were pregnant, did you feel like you wanted to have the baby?” And for abuse, we had two questions during interview.
So to tell you a little bit about the population from which we sampled, we had two sample sites in the state of Alabama, Tuscaloosa County and Mobile County. We went to all area clinics within the county and here are some basic demographic characteristics to give you an idea of how the counties compare with the state. You’ll see that Mobile County is disparate compared to Tuscaloosa and Alabama. Tuscaloosa pretty much follows the trend for the state itself, but you’ll see that in terms of income, Mobile County is much lower, in terms of poverty level, a little bit higher, in terms of birth, birth rate, the total number of live births, quite a bit more in comparison to the total population for the county. Now, how did we conduct our study? We conducted a prospective cohort and we wanted to interview women who were pregnant. We interviewed women in their first trimester of pregnancy at their initial prenatal visit. They had to be between the ages of 13 and 34, so they weren’t at high risk in terms of age. They were either black or white because we were interested in differences between the experience of black women and white women, and they had to be receiving Medicaid, so we’re talking about a specific low-income population.
We followed each participant forward through childbirth and collected data on birth outcome and we interviewed each participant twice during the pregnancy; at the initial visit, which had to be during the first trimester, and a second time during the third trimester. When we were all finished, we had a sample size of around 500, a little information on the sampling technique. We reviewed charts to see whether or not participants could, in fact, be included. Did they meet those prior criteria that I talked about on the previous slide, if so, we approached them during the prenatal visit. And for those of you that have worked at a health department clinic or Medicaid waiver clinic, those first visits are often a few hours long. So, we had enough time to approach them, speak with them, receive informed consent, and conduct the interview. Each of our study sites took approximately four years to collect the data that we collected on these pregnant women.
So here’s a little bit what the model looked like. I’m more of a picture person than I am a word person, so I like to put things into pictures and this is kind of what this particular presentation is all about. And this is kind of a component of my dissertation, a small component but we’re looking at the effect of partner support, how it buffers these problematic outcomes. And also, we’re interested in if whether you’re in the black or white ethnic group, if that effects how or interacts with social support in anyway. So a little bit about the sample, this is what the sample looked upon completion. You’ll see that we did have--in terms of ethnicity--a few missing. In terms of educational level, at least 2/3 had less than a high school education--between high school and middle school, and most of the women in the sample were single. In terms of other demographic characteristics, we had an average age of 23, and average weight of about 152, and a BMI of about 25.5.
Here’s what partner support looked like. About 23% of the sample said that they didn’t receive any partner support at all, the partner was not in the picture, and partner here is defined as either the father of the baby or current partner during pregnancy. In terms of this scale, we had an average score of about four. Our responses to the stressors, well, this is interesting in that about 42% of those we had responses for said that the pregnancy was unwanted, and in terms of abuse we had about 35% say that at some point they felt they had been abused. And you’ll notice that in the beginning I defined abuse as verbal and/or physical, so we’re talking about above and beyond physical, so that may be why the number seems a bit high for this population. On to the analysis, assessing whether or not partner support is present or absent, whether that buffers the effect of pregnancy wantedness, is here. We ran a logistic regression adjusting for age, educational attainment, marital status, pre-pregnant weight, and BMI, and we found out there is a significant association between partner support and wanting pregnancy.
In fact, women who lack partner social support are about two times more likely to say their pregnancy is unwanted. Does that differ between our two groups, between black and white women? Well, when we ran that analysis we found that it did not, so across the board partner support is important in terms of wanting the pregnancy. Now, to assess the magnitude of partner support, here is an analysis using the scale with pregnancy wantedness, and you’ll see again that there’s a significant association with the magnitude of support in wanting the pregnancy. For every one-point increase in that support scale, a woman is about 16% more likely to say her pregnancy is wanted compared to those who scored lower on the partner support scale. In terms of interaction, does being black or white interact with that social support mechanism and again we found that it did not, just across the board the magnitude of social support was the same for both groups. Now, in terms of abuse, when we ran the model with abuse alone, we found that social support in and of itself was not significantly associated with physical or verbal abuse during pregnancy. However, what’s interesting is when we ran the sub-group analysis looking at both black and white women, we found that there’s an indication of some kind of interaction there, so we wanted to explore that a little bit more.
When we looked at the two groups separately, we looked at white women and we found that white women who lack support are about 30% more likely to report physical abuse, because remember, we interviewed these women. We didn’t check them physically for any signs of abuse. We asked them, “Have you been abused?” So, it’s kind of a measure of perceived abuse during pregnancy, and also, much more likely than black women who had support. When you look at black women alone, black women lacking social support are less likely to report abuse than black women with social support. So we have, kind of, these opposing effects occurring in terms of ethnic groups, so we thought we’d look at that a little more closely. So we asses the magnitude of social support in terms of abuse, and again, when we ran it by itself, there were no significant associations, but when we looked at the interaction of ethnicity, there was an indication that that’s occurring, that there’s a difference between the two groups. And similarly, to presence or absence of support, there is this difference between white and black women, in that white women who lack support are more likely to report abuse, and you’ll see that black women are less likely to report abuse.
So, white women who lack support are about 2 1/2 times more likely to report abuse compared to socially supported white women, and over two times more likely to report it compared to socially supported black women. Again, when we look at black women alone, those who lack support are less likely to report abuse. So, what does that mean exactly? It means that we have some kind of qualitative different between the two groups. First of all, referring back to wantedness it’s a pretty straightforward result that presence of support is significantly associated with wanting the pregnancy, and the magnitude of that support is significantly associated with wanting the pregnancy. The findings were consistent across both ethnic groups and consistent with what we know.
Now, kind of back to the conundrum here, partner social support during pregnancy is significantly associated with abuse only in the presence of ethnicity, only through the interaction of ethnicity, and the findings are inconsistent between the two groups. White women with no support are more likely to report abuse whereas black women with no support are less likely to report abuse. So again, it’s qualitative. It’s, kind of, going like this. What are the possible explanations for that? Well it could be that type of abuse plays a role in reporting. Now we had measured verbal and physical abuse but in the end we combined them and just measured abuse overall. Perhaps there’s a problem in doing that in analysis. Also, perceived abuse may differ among our two groups. The questions themselves might be problematic and there might be cultural differences in the role and definition of social support. So how one particular group defines partner support may differ from how another group defines partner support. Social support measures may be masking the associations on sub-scales. Well you could break that scale into emotional and instrumental or material support and we kept the scale whole.
Perhaps if we’d broken the scale into two parts we would have found that there’s something going in terms of emotional versus material support. And the role of the partner providing social support may differ among groups. It may be, in fact that among one group, particularly white women, the partner is key in providing support during pregnancy. However, in the group of black women it may be that other familial members are more important in providing support. Maybe mothers or fathers, and partners don’t play that important a role in the pregnancy. So how are we going to address that? Well, first of all I think that next time we run the analysis we’ll be looking at abuse specifically in terms of physical and verbal. We’ll separate the two out and see if that has an effect on this interaction. We’ll look at social support in terms of emotional versus instrumental or material, and we’ll look at the roles that other family members play in social support, and is the mother more important, or as important as the partner in one group over the other. Then we’ll also be looking at the role of bias in the study. Was there a response bias? Was one group less likely to report abuse than another group? That’s very important I think and also interviewer bias, did some interviewers perhaps coax participants more than others, could their demeanor have affected the reporting itself.
So, in conclusion, assessing the presence or absence of partner supporting pregnancy is key to decreasing the stressors we’ve talked about. Further, the magnitude of support plays a role in buffering the affects of the stressors. If you remember the model that I showed you in the beginning, there is a buffering effect. So what can practitioners do? Well they can take a proactive role in primary prevention, which I know is what we’re always thinking about and they may alleviated existing stressors by intervening in current relationships if they believe the relationship may, in fact, be abusive or affecting the pregnancy. So, again, ongoing analysis will enable further explanation of these findings, especially the very interesting finding of interaction with support and abuse, and other stressors should be analyzed in terms of these outcomes. Also, I think it’s important to focus on specific culturally defined sub-groups to study that buffering effect. If there is a true difference between these black and white women in the study, we need to separate them and look at them exclusively. Okay, thank you.