American Journal of Epidemiology Advance Access originally published online on January 14, 2008
American Journal of Epidemiology 2008 167(6):644-645; doi:10.1093/aje/kwm369
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Toh et al. Respond to "Compromise or Compromising?"
1 Department of Epidemiology, Harvard School of Public Health, Boston, MA
2 Slone Epidemiology Center at Boston University, Boston, MA
Correspondence to Sengwee Toh, Department of Epidemiology, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA (e-mail: swtoh{at}hsph.harvard.edu).
Received for publication November 21, 2007. Accepted for publication November 27, 2007.
We thank Dr. Howards for her commentary (1). Our study (2) did not attempt to either credit or discredit the usefulness of automated databases in reproductive research but to raise and quantify the impact of one particular methodological concern: the ability of databases to accurately establish gestational periods. We agree with Dr. Howards that the two algorithms considered in our paper, particularly the one based on delivery date, can provide useful information, in some circumstances, on drug prescription during pregnancy. However, the validity of these algorithms in the assessment of drug safety depends on a number of factors related to gestational timing, including aspects of treatment with a specific drug (e.g., duration of treatment), the specific outcome (e.g., etiologically meaningful period of exposure), and their association with gestational age at birth (e.g., outcome associated with prematurity). Unfortunately, the necessary conditions for an algorithm to be valid for estimating risks of reproductive outcomes in relation to medication use might be unknown to the investigator and may not hold simultaneously.
Limitations arising from the absence of timing of conception may be ameliorated, as Dr. Howards (1) and we suggested (2), by linking automated databases to relevant data sources that provide more detailed and accurate gestational information. For example, Tennessee Medicaid data—a computerized claims database—have been successfully linked to vital and medical records, permitting a more accurate estimation of conception dates (3). However, the feasibility and resources required for such linkage efforts will vary according to the nature of the various databases.
Dr. Howards (1) states that it is neither possible nor desirable to develop a large cohort study or other designs to examine an exposure that is unlikely to have an effect, and she offers the compromise of using automated databases and linking them to other relevant medical records to reduce costs and effort. However, Dr. Howards notes that such an approach would be unlikely to provide data on potential confounders, and we agree. Studies of certain birth defects often require consideration of potential confounders such as smoking and folate/multivitamin exposure, and information on such exposures is unavailable in most computerized databases. In contrast, case-control surveillance of birth defects, including the Slone Epidemiology Center Birth Defects Study (4) and the National Birth Defects Prevention Study (5), is an available, productive, and cost-effective design that not only provides information on gestational timing (2) but also provides critical data on a wide range of potential confounders (e.g., cigarette smoking, alcohol intake, over-the-counter drugs including vitamin/mineral supplements, and diet). While computerized databases have great value in many areas of drug safety, their usefulness in the study of birth defects has to be evaluated in the context of alternative designs that provide information on important potential confounders as well as more accurate information on gestational timing.
 |
ACKNOWLEDGMENTS |
|---|
Conflict of interest: none declared.
|
References |
|---|
 TOP References |
|---|
- Howards P. Invited commentary: The use of imperfect data—compromise or compromising? Am J Epidemiol (2007) 167:641–3.
- Toh S, Mitchell AA, Werler MM, et al. Sensitivity and specificity of computerized algorithms to classify gestational periods in the absence of information on date of conception. Am J Epidemiol (2007) 167:633–40.
- Cooper WO, Hernandez-Diaz S, Arbogast PG, et al. Major congenital malformations after first-trimester exposure to ACE inhibitors. N Engl J Med (2006) 354:2443–51.
[Abstract/Free Full Text] - Louik C, Lin AE, Werler MM, et al. First-trimester use of selective serotonin-reuptake inhibitors and the risk of birth defects. N Engl J Med (2007) 356:2675–83.
[Abstract/Free Full Text] - Alwan S, Reefhuis J, Rasmussen SA, et al. Use of selective serotonin-reuptake inhibitors in pregnancy and the risk of birth defects. N Engl J Med (2007) 356:2684–92.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in Am. J. Epidemiol.:
- Sensitivity and Specificity of Computerized Algorithms to Classify Gestational Periods in the Absence of Information on Date of Conception
- Sengwee Toh, Allen A. Mitchell, Martha M. Werler, and Sonia Hernández-Díaz
Am. J. Epidemiol. 2008 167: 633-640.[Abstract] [FREE Full Text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||