American Journal of Epidemiology Advance Access originally published online on August 18, 2006
American Journal of Epidemiology 2006 164(7):627-628; doi:10.1093/aje/kwj262
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Response to Invited Commentary |
Green et al. Respond to "Clues to the Etiology of Inflammatory Bowel Disease"
1 Public Health Branch, Manitoba Health, Winnipeg, Manitoba, Canada
2 Department of Community Health Sciences, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
3 Department of Internal Medicine, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
4 Inflammatory Bowel Disease Clinical and Research Centre, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
Correspondence to Dr. Charles N. Bernstein, Faculty of Medicine, University of Manitoba, 804F-715 McDermot Avenue, Winnipeg, Manitoba, Canada R3E-3P4 (e-mail: cbernst{at}cc.umanitoba.ca).
Received for publication March 28, 2006. Accepted for publication March 31, 2006.
Dr. Moayyedi has made several insightful comments (1
) on our paper (2). On the basis of his comments, we are encouraged to proceed with a number of follow-up studies, including modeling of the spatial relation between multiple sclerosis and enteric infections.
We do, however, have the following comments in response. Firstly, Dr. Moayyedi suggests that migration of individuals between small geographic areas over time brings into question our conclusions concerning the relation between high levels of hygiene in earlier life and subsequent development of inflammatory bowel disease in later life (the "hygiene hypothesis"). While we agree that our study design may not have accurately captured information on the environmental characteristics of individuals in early life, we argue that this issue may not be as large a problem as it first appears. A number of recent studies have suggested that mobility rates have declined since the 1980s, with the majority of moves taking place over very short distances as people adjust their housing arrangements to accommodate life events such as marriage, divorce, and the addition of new children (3, 4). This would suggest that, barring dramatic changes in socioeconomic or life circumstances, most people when moving tend to migrate to adjacent neighborhoods with characteristics similar to those of the ones they just left. It is also plausible that exposure to enteric infections in both childhood and young adulthood may provide protection against autoimmune diseases later in life. Since there may be a relatively short period of time between the protective effect of enteric exposure and the peak incidence of inflammatory bowel disease, which occurs shortly thereafter in the third decade of life, the effect of migration on our study results may have been quite minimal. While the failure to control for the effects of migration in our study most certainly reduced spatial variability in disease rates and the strength of observed relations (4), the study produced marked and statistically significant results. Conceivably, if we had been able to successfully control for the migration bias in our study design, the spatial patterns and statistical relations we observed may have been more pronounced.
Secondly, Dr. Moayyedi suggests that our method of spatial smoothing may have been biased, since the expected number of cases in each geographic area was on average too small (<20). Although we agree that the smaller the number of observations in a study area the lower the confidence one has in smoothed estimates, we argue that Dr. Moayyedi's concerns are somewhat overstated, since he bases his conclusions on a study (5) which assessed the impact of numerator size on the performance of Bayesian smoothing, not the adaptive mean nearest-neighbor smoothing approach (6, 7) explicitly used in our study. In our study, we confirmed the validity of our smoothed spatial patterns with the spatial scan statistic (8, 9), a technique considered to be very robust to the problem of small numbers.
Thirdly, Dr. Moayyedi suggests that our study may suffer from ecologic bias. While we agree that the ecologic design of our study may have limitations related to the ecologic fallacy, we would argue that in studying the relation between infectious diseases and chronic diseases, where many cases of infectious disease are never reported, an attempt to relate cases of chronic and infectious diseases at the individual level would be seriously confounded by these "missing cases." Analyzing trends at the population/ecologic level, however, minimizes this problem significantly, since the units of analysis that are used are geographically defined populations at risk as opposed to individuals at risk.
Finally, in a previous study, Bernstein et al. (10) found statistically significant associations between subjects with ulcerative colitis and Crohn's disease and a number of chronic immunoinflammatory diseases. Thus, while these autoimmune diseases do cluster on an individual basis, we hope to conduct further ecologic studies in search of associations worth pursuing on an individual basis. We fully agree with Dr. Moayyedi (1) that the pursuit of the etiology of inflammatory bowel disease and the strength of the associations we have reported could be enhanced greatly by similar studies from other jurisdictions.
 |
ACKNOWLEDGMENTS |
|---|
Conflict of interest: none declared.
|
References |
|---|
 TOP References |
|---|
- Moayyedi P. Invited commentary: clues to the etiology of inflammatory bowel disease—a return to John Snow? Am J Epidemiol 2006;164:624–6.
[Free Full Text] - Green C, Elliott L, Beaudoin C, et al. A population-based ecologic study of inflammatory bowel disease: searching for etiologic clues. Am J Epidemiol 2006;164:615–23.
[Abstract/Free Full Text] - Hurley SE, Reynolds P, Goldberg DE, et al. Residential mobility in the California Teachers Study: implications for geographic differences in disease rates. Soc Sci Med 2005;60:1547–55.[CrossRef][Web of Science][Medline]
- Rogerson PA, Han D. The effects of migration on the detection of geographic differences in disease risk. Soc Sci Med 2002;55:1817–28.[CrossRef][Web of Science][Medline]
- Richardson S, Thomson A, Best N, et al. Interpreting posterior relative risk estimates in disease-mapping studies. Environ Health Perspect 2004;112:1016–25.[Web of Science][Medline]
- Waller L, Gotway C. Visualizing spatial data. Applied spatial statistics for public health. Hoboken, NJ: John Wiley and Sons, Inc, 2004:68–117.
- Centers for Disease Control and Prevention. Spatial smoothing of heart disease mortality rates. Appendix B: methodological and technical notes. In: Women and heart disease. (Online atlas). Atlanta, GA: Centers for Disease Control and Prevention, 2002. (http://ftp.cdc.gov/pub/Publications/womens_atlas/09-append-b.pdf).
- Kulldorff M. Geographic information systems (GIS) and community health: some statistical issues. J Public Health Manag Pract 1999;5:100–6.[Medline]
- Kulldorff M, Athas WF, Feurer EJ, et al. Evaluating cluster alarms: a space-time scan statistic and brain cancer in Los Alamos, New Mexico. Am J Public Health 1998;88:1377–80.
[Abstract/Free Full Text] - Bernstein CN, Wajda A, Blanchard JF. The clustering of other chronic inflammatory diseases in inflammatory bowel disease: a population-based study. Gastroenterology 2005;129:827–36.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||