Acknowledgments: This study was funded by the National Institute on Alcohol Abuse and Alcoholism (1R01AA017104), the National Institute of Mental Health (T32MH109205), the UCLA Center for HIV Identification, Prevention and Treatment Services (P30MH58107), the UCLA Clinical and Translational Science Institute (UL1TR000124), and the UCLA Center for AIDS Research (P30AI028697).
Note: this paper compares questions about prenatal drinking from two timepoints (baseline and 5 years) of this dataset. I later learned that questions about prenatal drinking were also asked at another timepoint, which was immediately after birth, but I haven't updated these analyses or this manuscript.
Drinking during pregnancy can lead to fetal alcohol spectrum disorders (FASD), which comprise a wide variety of detrimental consequences for the health of a child (Williams & Smith, 2015). Thus, measuring how much expectant mothers drink is important for public health, but mothers may be reluctant to admit their drinking to investigators. FASD are particularly common in the Western Cape Province of South Africa, with about 15% of children affected (May et al., 2013). We studied self-reports of drinking among mothers in this region.
One strategy that might be expected to elicit more honest replies is to ask mothers about prenatal drinking long after the child is born, rather than while they are still pregnant. Mothers may feel less shame about past events, or less able to effectively deceive researchers when they have a child with obvious signs of fetal alcohol spectrum disorder. On the other hand, mothers' memories of their drinking behavior should become less accurate over time, and their shame may be more acute once they have a visibly unhealthy child. Hence, it is not obvious whether mothers will be more honest, or will say they drink more, when asked contemporaneously or retrospectively.
Studies from Cleveland and Detroit hospitals asking mothers about prenatal drinking both during pregnancy and from 1 to 14 years after birth have found that mothers' reports of how much they drink are much more likely to increase than decrease (Ernhart, Morrow-Tlucak, Sokol, & Martier, 1988; Jacobson, Chiodo, Sokol, & Jacobson, 2002; Hannigan et al., 2010). In other words, within subjects, mothers report drinking more when interviewed retrospectively. However, past results are less consistent when drinking data is dichotomized into any drinking versus no drinking, rather than represented as ounces of alcohol. Comparing the sober consistency rate (the proportion of mothers who said they abstained retrospectively, among those who had also said they abstained contemporaneously) to the drinking consistency rate, Hannigan et al. (2010) found greater sober consistency (80% vs. 70%), Ernhart et al. (1988) found greater drinking consistency (61% vs. 78%), and Jacobson et al. (2002) found very similar rates (78% vs. 80%). It is thus unclear whether retrospective report makes mothers more or less likely to admit drinking during pregnancy. Furthermore, it is unclear how well these results generalize outside the United States.
Why would one use a dichotomous measure of prenatal drinking rather than a continuous one? A dichotomous measure is easy to administer: it requires no prop beverage containers, definition of a standard drink, distinguishing between types of beverages because of their different concentrations of alcohol, or discernment of mothers' drinking patterns. Besides reducing potential variability in how mothers interpret questions, a dichotomous measure is less dependent on memory. A priori, people seem likely to misremember exactly how much, when, or what they drank, but should have little trouble remembering whether they drank at all. Finally, while there are certainly dose-dependent effects of alcohol on feti, the consensus of researchers is that no amount of alcohol during pregnancy is safe (Williams & Smith, 2015), so preventing prenatal drinking entirely, rather than merely reducing frequencies or amounts to a known nonzero safe level, is typically the goal of interventions.
In this study, we compared contemporaneous reports of prenatal drinking with 5-year retrospective reports among mothers in the Western Cape. We also compared dichotomous to continuous measures of drinking. Finally, we compared how well these various ways of measuring drinking (contemporaneously or retrospectively, and dichotomously or continuous) could predict head circumference soon after birth and child intelligence 5 years after birth. The overall goal was to better understand the strengths and weaknesses of these strategies, for the benefit of future investigation into prenatal drinking.
We used data from a cluster-randomized controlled trial of a mentor intervention for new mothers and their babies in the area of Cape Town, South Africa (see le Roux et al., 2013; ClinicalTrials.gov registration number NCT00972699). Mothers in both conditions received comprehensive healthcare at clinics. Mothers in the experimental condition additionally received regular home visits from another mother who was a trained community health worker and a positive deviant, having had a healthy baby herself. Conditions were randomly assigned per neighborhood. In this paper, because we do not wish our results to be affected by the experimental intervention, we consider only the mothers in the control group.
Our analysis draws primarily from assessments at two timepoints, baseline and 5 years after birth. At the time of the baseline assessment (T1), mothers were 3 to 40 weeks pregnant (mean 26 weeks). Mothers were asked "How often did you use alcohol in the month before you found out you were pregnant?" and "During the month before you found out you were pregnant, counting all types of alcohol combined, how many drinks did you USUALLY have on days when you drank alcohol?". (The second question was skipped if the answer to the first was "Never". We asked mothers about the month before they recognized their pregnancy because we expected they would be more candid about this period, and that they would notice their pregnancies relatively late.) At the 5-year follow-up (T2), mothers were asked the same questions, except the word "pregnant" was replaced with "pregnant with" followed by the child's name. To define for the mother how large a drink was, the interviewer showed prop beverage containers including a beer bottle, a 250-mL wine glass, and a shot glass; we aimed to represent the American standard drink of 0.6 fl oz (14 g) of ethanol. Table 1 shows the response options for the two questions and how they were coded. For a dichotomous measure of drinking, we coded mothers as sober if they answered "Never" and drinking otherwise. To compute fluid ounces of absolute alcohol drank per day (AA/day), a measure we chose to aid comparison with previous research, we multiplied 0.6 by the numeric frequency code by the numeric amount code.
|Less than once a month||0.02|
|Once a month||0.05|
|2 to 3 times a month||0.10|
|Once a week||0.20|
|2 times a week||0.50|
|3 to 4 times a week||0.50|
|Nearly every day||0.70|
|Number of drinks|
|1 or 2||1.50|
|3 or 4||3.50|
|5 or 6||5.50|
|7,8 or 9||8.00|
|10 or more||10.00|
We consider two outcomes. At an assessment of the mothers immediately after birth (between T1 and T2, when the children were up to 4 months old; prenatal drinking was not measured during this assessment), we measured the circumference of children's heads, and calculated their sex- and age-specific z-scores using WHO norms. At T2, children were administered the Kaufman Assessment Battery for Children (Lichtenberger & Kaufman, 2010). This paper uses the Mental Processing Index (MPI), which measures general mental processing ability (i.e., intelligence) and excludes an assessment of acquired knowledge. Like traditional intelligence measures, the MPI is normed to have a mean of 100 and an SD of 15.
Data-analysis code can be found at http://arfer.net/projects/philani.
|Dichotomous (drinking vs. sober)|
|Mothers reporting, either timepoint||500|
|Percent drinking, T1||26%|
|Percent drinking, T2||22%|
|Mothers reporting, both timepoints||366|
|Percent consistent, among sober at T1||92%|
|95% Jeffreys CI||[88%, 95%]|
|Percent consistent, among drinking at T1||63%|
|95% Jeffreys CI||[53%, 72%]|
|Mothers reporting, either timepoint||500|
|Mothers reporting, both timepoints||365|
|Percent increasing from T1 to T2||18%|
|Percent decreasing from T1 to T2||12%|
|95% bootstrap CI||[0.07, 0.18]|
Table 2 compares reports of drinking at T1 and T2. Coding drinking dichotomously, we see that mothers reported drinking at a slightly higher rate at T1 than T2, and mothers who stated they were sober at T1 were much more likely to repeat their answer at T2 than mothers who stated they drank at T1. By contrast, when we code drinking in terms of AA/day, means are higher at T2 than T1, and mothers were slightly more likely to increase their reported amount of drinking than decrease it; the mean change was positive, at .12 AA/day.
|T1, dichotomous||-.0824 [-.16, .00]||-.0538 [-.15, +.04]|
|T1, continuous||-.0817 [-.16, .00]||-.0504 [-.14, +.04]|
|T2, dichotomous||-.0616 [-.15, +.02]||-.0711 [-.17, +.03]|
|T2, continuous||-.0673 [-.15, +.01]||-.0668 [-.16, +.03]|
Table 3 shows how retrospective versus contemporaneous reports and dichotomous versus continuous coding affect the correlation of drinking with children's head circumference immediately after birth and intelligence 5 years after birth. The correlations are all negative, as expected, but small. The sample correlation is largest for the dichotomous T1 measure for head circumference and the dichotomous T2 measure for intelligence, but the differences are very small, and all the confidence intervals are largely overlapping.1
- O'Connor and Paley (2006) found that using retrospective reports, mothers' statements about their prenatal drinking were more strongly correlated with child head circumference and number of physical anomalies than statements about their drinking after birth or currently. This provides evidence that retrospective reports are not merely reflections of current drinking habits.
Efron, B., & Tibshirani, R. (1993). An introduction to the bootstrap. New York, NY: Chapman & Hall. ISBN 978-0-412-04231-7.
Ernhart, C. B., Morrow-Tlucak, M., Sokol, R. J., & Martier, S. (1988). Underreporting of alcohol use in pregnancy. Alcoholism, 12(4), 506–511. doi:10.1111/j.1530-0277.1988.tb00233.x
Hannigan, J. H., Chiodo, L. M., Sokol, R. J., Janisse, J., Ager, J. W., Greenwald, M. K., & Delaney-Black, V. (2010). A 14-year retrospective maternal report of alcohol consumption in pregnancy predicts pregnancy and teen outcomes. Alcohol, 44(7–8), 583–594. doi:10.1016/j.alcohol.2009.03.003. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2889143/
Jacobson, S. W., Chiodo, L. M., Sokol, R. J., & Jacobson, J. L. (2002). Validity of maternal report of prenatal alcohol, cocaine, and smoking in relation to neurobehavioral outcome. Pediatrics, 109(5), 815. doi:10.1542/peds.109.5.815
le Roux, I. M., Tomlinson, M., Harwood, J. M., O'Connor, M. J., Worthman, C. M., Mbewu, N., … Rotheram-Borus, M. J. (2013). Outcomes of home visits for pregnant mothers and their infants: A cluster randomised controlled trial. AIDS, 27(9), 1461–1471. doi:10.1097/QAD.0b013e3283601b53
Lichtenberger, E. O., & Kaufman, A. S. (2010). Kaufman Assessment Battery for Children—Second Edition (KABC-II). In C. S. Clauss-Ehlers (Ed.), Encyclopedia of cross-cultural school psychology (pp. 557–564). New York, NY: Springer. ISBN 978-0-387-71798-2.
May, P. A., Blankenship, J., Marais, A.-S., Gossage, J. P., Kalberg, W. O., Barnard, R., … Seedat, S. (2013). Approaching the prevalence of the full spectrum of fetal alcohol spectrum disorders in a south African population-based study. Alcoholism, 37(5), 818–830. doi:10.1111/acer.12033
O'Connor, M. J., & Paley, B. (2006). The relationship of prenatal alcohol exposure and the postnatal environment to child depressive symptoms. Journal of Pediatric Psychology, 31(1), 50–64.
Williams, J. F., & Smith, V. C. (2015). Fetal alcohol spectrum disorders. Pediatrics, 136(5), e1395–e1406. doi:10.1542/peds.2015-3113
We also examined how well the four drinking measures could predict the two outcomes with linear regression models, evaluated with cross-validated mean square error. None of the estimates of predictive accuracy meaningfully improved upon baseline. We omit details for space.