Agreement Between Prospective and Retrospective Reports of Maternal Exposure to Chemicals During Pregnancy Christine Till, MA Gideon Koren, MD, FRCPC Joanne F. Rovet, PhD Learning Objectives • Compare how well, in previous studies, retrospectively and prospectively collected information on exposure of mothers to chemicals or drugs have agreed with medical records data. • Recall the reported agreement— or lack thereof— between self-reported exposure to organic solvents at the time of pregnancy and at follow-up 3 to 7 years later. • Describe the implications of these findings and how more valid maternal exposure data might be acquired. Abstract As part of a prospective study of solvent exposure and child behavior, it occurred that women’s responses about exposure history during pregnancy differed from the information documented during the postpartum period. The objective of the study was to examine the levels of agreement in 29 self-reports of exposure information obtained before and 3- to 7-years after pregnancy. Percent agreement was low for duration of exposure (41%), protective barrier use (48%), and symptomatology (41%). When reports were not in perfect agreement, women tended to report longer durations of exposure (r = 0.67), increased use of protective barriers (r = 0.39), and more symptoms at time of follow-up (r = 0.57). Agreement of report was not substantially associated with time since pregnancy or concurrent child behavior. Low levels of agreement may reflect response biases in the data collected at time of pregnancy or recall biases at time of follow-up. These variations in self-reports are of concern because they can severely affect estimates of human teratolog- ical risk. ( J Occup Environ Med. 2002;44:708 –713) M uch of our knowledge about repro- ductive hazards associated with chemical exposure during pregnancy is based on retrospective data collec- tion. Typically, these data are as- sessed using self-administered ques- tionnaires or personal interviews with the mother during the postpar- tum period. We depend on such methods because the woman who was pregnant is often the only readily available source of informa- tion. While these methods appear to ask for simple recall of work history, both accuracy and reliability of ex- posure data are not always demon- strated. 1–3 Studies assessing the reliability of maternal recall for pregnancy infor- mation have shown that accuracy of report varies according to type of information assessed. Specifically, maternal recall for type of employ- ment, 4 past miscarriages, number of pregnancies, and hospitalizations are accurately reported by women months to years after the child’s birth when compared with data collected during pregnancy. 1,5– 8 However, ret- rospective ascertainment of exposure to chemicals, 1–3 medications, 9 –11 al- cohol or drugs, 12–14 and tobacco smoking 15–18 during pregnancy has shown poor agreement with medical records. These results draw attention to the limitations of using retrospec- tive data to study potential reproduc- tive hazards associated with expo- sures during pregnancy. A serious problem with retrospec- tively collected exposure informa- tion is the potential for recall bias. From the Department of Psychology, University of Toronto, Toronto, Ontario, Canada (Ms Till); Department of Pediatrics, University of Toronto, Toronto, Ontario, Canada (Dr Koren, Dr Rovet); The Motherisk Program, The Hospital for Sick Children, Toronto, Ontario, Canada (Dr Koren, Dr Rovet); Division of Clinical Pharmacology, The Hospital for Sick Children, Ontario, Canada (Dr Koren). This research was funded in part by the Workers Safety Insurance Board (WSIB), (Ontario, Canada) and the Physicians Services Incorporated (Ontario, Canada). Christine Till was supported by the National Science and Engineering Research Council of Canada (NSERC). Gideon Koren is a Senior Scientist of the Canadian Institute for Health Research (CIHR), and holder of the Research Leadership for Better Pharmacotherapy During Pregnancy and Lactation. Christine Till has no commercial interest related to this article. Address correspondence to: Christine Till, Department of Psychology, The Hospital for Sick Children, 555 University Ave., Toronto, ON M5G 1X8; christine.till@sickkids.ca. Copyright © by American College of Occupational and Environmental Medicine DOI: 10.1097/01.jom0000026048.24145.a1 CME Article #1 708 Maternal Report of Exposure Information • Till et al