Hospital Or Neighborhood Control: Which One Could Have More Influence On The Assessment Of Esophageal Cancer Risk Factors?

گوارش، مجله انجمن متخصصين گوارش و كبد ايران

دوره 15 - شماره Supple.1

Article Type: Original Article
Abstract: Introduction: Case-Control studies are commonly used in studying less common cancers. While case definition is usually straightforward, appropriate control selection can be difficult. The aim of our study was to compare the results of using Hospital or Neighborhood control groups. Method: We compared the results of two different case-control studies of esophageal squamous cell carcinoma (ESCC) conducted in the same region by a single research group for several known or suspected risk factors. Case definition and enrollment was the same, but control selection differed. For the first, we selected two age- and sex-matched controls from inpatient subjects in hospitals, while for the second we selected two age- and sex-matched controls from each subject’s neighborhood of residence. We used the test of heterogeneity to compare the results of the two studies. Results: We found no significant differences in exposure data for tobacco-related variables such as cigarette smoking, chewing nass and hookah usage, but the frequency of opium usage was significantly different between hospital and neighborhood controls. Consequently, the inference drawn for the association between ESCC and tobacco use did not differ between studies, but it did for opium use. Comparing the prevalence of opium consumption in the two control groups and a cohort enrolled fromthe same geographic area suggested that the neighborhood controls were more representative of the study  base population for this exposure. Conclusion: In our two studies, we found that hospital and neighborhood controls did not lead us to the same conclusions for a major hypothesized risk factor for ESCC in this population. Our results show that control group selection is critical in drawing appropriate conclusions in observational studies.
Keyword: