MCHB/EPI Miami Conference — December 7 - 9, 2005

Low Birth Weight: The Unbearable Lightness of Being — Transcript

CASSANDRE LARRIEUX: Good afternoon everybody. I do thank you for being here and for being awake still, because I know how it is after a very filling lunch. It can be kind of difficult. I'll try to keep it as lively as I can, but I make no promises.

My presentation today will look at the effects of maternal WIC participation on the incidence of very low birth weight in Michigan 's Medicaid population. The things that I would like for you to come away from this presentation with is to first get an idea what the incidence of very low birth weight is in one of the most vulnerable sections of the population in Michigan, those women who household incomes are so low that they do qualify for Medicaid. Second, I would like to illustrate again the benefits of linking data from various programs in order to answer epidemiological questions. And finally, I want you to see the effects of WIC participation, maternal participation had on birth weight in this particular high-risk population.

Very low birth weight infants are at very high risk for adverse health outcomes. They, by definition, weigh less than 1500 grams or about 3.3 pounds. They're usually pre-term, they are more likely to die compared to normal birth weight infants, and if they survive are more likely to experience various morbidities, like cerebral palsy.

Now, I know most of you have heard of WIC, the special supplemental nutrition program for women, infants, and children. Participants, I do want to stress, do select themselves to enter the program and to receive its benefits. Those benefits include: supplementing their diet with nutrient dense foods, offering breastfeeding and nutritional counseling, and referring participants to other health and social services that they may be eligible for. Over 6,000 pregnant and postpartum women participate in Michigan 's WIC program annually.

Several states have already looked at the birth and health outcomes of WIC participations within their states. In 1993, North Carolina found that prenatal WIC participation was associated with significant reductions in low birth weight and very low birth weight. A 1984 study in Massachusetts found small improvements in mean birth characteristics. A study in 2004 in New York State found that longer lengths of WIC participation in the prenatal period was significantly associated with positive birth outcomes. Even Michigan , a couple of years ago, well more than a couple of years ago, had performed various studies--had performed a study evaluating the WIC program. The 1998 study by Alowla et al, found that longer lengths of WIC enrollment was associated with reduction in the odds of small for gestational age infants.

Now, you might be wondering why is it that I'm presenting this afternoon about WIC participation in the Medicaid population when, you know, there have, they did do a study several years, not too long ago on a similar subject. What I did and what they did in 1998 is different. There were different exclusion criteria. I included only preemie para-Medicaid participants who delivered a live born singleton infant. They excluded women who had delivered pre-term, women with multiple births, and restricted their analysis to term infants only. We also identified the Medicaid participants differently. They used the variable in the vital records data to determine who was on Medicaid and who was not. I was able to access Medicaid's information myself and identify those women and link them to their WIC information, if they had such. And third, the analysis that we did was slightly different.

The particular question that we were looking at this time around was does prenatal WIC participation reduce the odds of very low birth weight infants in the Medicaid population. Like I said before, we used information from Michigan 's Medicaid and WIC program added databases as well as vital records to answer this particular question. These data sets were linked internally in Michigan 's data warehouse. Like I said before, only preemie para-Medicaid women who delivered a live born singleton infant in 2002 were included in the analysis. And among the estimated 130,850 infants born in 2003 in Michigan , almost half, 42.2 percent were participants. So it's a large chunk of infants that we're talking about their outcomes and their mother's participation.

We performed various uni-variant and bi-variant analyses before calculating the adjusted odds ratio via logistic regression. We used the information from the uni-variant and bi-variant analysis really to guide our multi-variant analysis.

We performed a forward selection logistic regression in SASS, the initial analysis. In the initial analysis we included variables for prenatal WIC participation, maternal age, race ethnicity, adequacy of prenatal care utilization, and smoking during pregnancy. All the variables except for adequacy of prenatal care utilization were used in the final logistic regression model. The reference group for the variables in the final model included women from 18 to 24 years old for maternal age, Hispanic, ethnicity of any race for race ethnicity, and no smoking during pregnancy for the smoking variable.

Of the 16,941 women we included in the study, 263, which is 1.55 percent, delivered a very low birth weight infant. After adjusting for maternal characteristics in the final logistic regression model, there was a statistically significant association between not enrolling in WIC and experiencing a very low birth weight outcome. Those women who did not enroll were 51 percent more likely to have delivered a low birth weight infant. Other co-variants found to be significantly associated with very low birth weight included smoking, non-Hispanic/black race, maternal age less than 18 years old, and maternal age greater than or equal to 25. Women who did not participate in WIC were younger than 18, older than 25, were of black race, or smoked during pregnancy, had higher odds of delivering a very low birth weight infant compared to their counterparts.

As with all research, this one had its share of limitations. First, maternal nutrition impacts fetal growth more strongly than it does length of gestation. If I had to do it over again I would try to figure out how to do the weight for gestational age, which probably would have been a better indicator. Several studies have documented the differences eligible women who participate in WIC and those who don't. I wasn't able to adjust for some of these confounders, which, you know, can include parental motivation, childcare, transportation issues. And because I couldn't adjust for them, that is a limitation of this study. Adverse maternal health behaviors, including smoking and drinking are known to be underreported in vital records and that is something that is another limitation and I couldn't, you know, also not only because in vital records they ask you whether you smoked or you don't. Like I said, most women even if they do smoke will say that they don't. Another thing too that's a problem with that particular variable is that they don't adjust for how much the women smoke. We expect to see that heavier smokers would be more likely to experience an adverse birth outcome, but because they don't have that information, that was something I couldn't adjust for. And because we limited our analysis to Medicaid participants who do have access to healthcare and who are of low income we could not generalize this to the entire state of Michigan .

WIC encourages and we still should continue to encourage women, eligible women to enter the program as soon as they realize that they--as soon as they think they're pregnant. WIC can be a vehicle for other programs to capture those women and influence adverse maternal behavior. So, you know, if we could partnership with a smoking cessation program or drinking program because we do have the women, but because of funding and because of the scope of WIC--we provide health information. Our funding does not include performing medical services or very intensive interventions. That is something that a partnership might be able to ameliorate.

Also, I would like, if I could do this, you know, could do my own study, I really want to get some information about those characteristics, those that differentiate women who enroll and women who don't. I would really like to have adjusted for that, because that is a confounder and it tends to be a confounder for most WIC studies because we just--we have programmatic data, we don't ask why did you not enter earlier. And so that is something else that if I could do my own study independently and had, you know, the funds to do so, that's the kind of information that I would like to collect in addition to what's already there to adjust for that.

And I do thank you for listening and being here and I will answer, like I said, like Milton said, a, you know, very pressing question right now, and then some more a little bit later on. Thank you.