Twin Research

Articles

Accuracy of Mothers' Retrospective Reports of Smoking During Pregnancy: Comparison with Twin Sister Informant Ratings

Andrew C. Heatha1 c1, Valerie S. Knopika2, Pamela A. Maddena3, Rosalind J. Neumana4, Michael J. Lynskeya5, Wendy S. Slutskea6, Theodore Jacoba7 and Nicholas G. Martina8

a1 Missouri Alcoholism Research Center, Department of Psychiatry,Washington University School of Medicine, St Louis, Missouri, USA. andrew@matlock.wustl.edu

a2 Missouri Alcoholism Research Center, Department of Psychiatry,Washington University School of Medicine, St Louis, Missouri, USA.

a3 Missouri Alcoholism Research Center, Department of Psychiatry,Washington University School of Medicine, St Louis, Missouri, USA.

a4 Missouri Alcoholism Research Center, Department of Psychiatry,Washington University School of Medicine, St Louis, Missouri, USA.

a5 Missouri Alcoholism Research Center, Department of Psychiatry,Washington University School of Medicine, St Louis, Missouri, USA; Palo Alto Veterans Administration, Palo Alto, California, USA.

a6 Queensland Institute of Medical Research, Brisbane, Australia.

a7 Missouri Alcoholism Research Center, Department of Psychology, University of Missouri, Columbia, Missouri, USA.

a8 Palo Alto Veterans Administration, Palo Alto, California, USA.

Abstract

Retrospective assessment of maternal smoking or substance use during pregnancy is sometimes unavoidable. The unusually close relationship of twin sister pairs permits comparison of self-report data versus co-twin informant data on substance use during pregnancy. Information about smoking during pregnancy has been gathered from a series of mothers from an Australian volunteer twin panel (576 women reporting on 995 pregnancies), supplemented in many cases by independent ratings of their smoking by twin sister informants (821 pregnancies). Estimates of the proportion of women who had never smoked regularly (56–58%), who had smoked but did not smoke during a particular pregnancy (16–21%), or who smoked throughout the pregnancy (16–18%), were in good agreement whether based on selfreport or twin sister informant data. However, informants underreported cases who smoked during the first trimester but then quit (1–3% versus 7–9% by self-report). Women who smoked throughout pregnancy (by informant report) rarely denied a history of regular smoking (< 1%), although a small proportion of apparent false negative cases were identified where they either denied smoking during a pregnancy (9%) or denied smoking beyond the first trimester (10%). We conclude that retrospective smoking data can safely be used to identify potential associations of later child outcomes with maternal smoking during pregnancy.

Correspondence:

c1 Address for correspondence: Andrew C. Heath, D.Phil., 40 N. Kingshighway, Suite 2, St Louis MO 63108, USA.

Metrics