IJE Advance Access published online on June 20, 2008
International Journal of Epidemiology, doi:10.1093/ije/dyn120
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letter to the Editor |
Author's response
1 Division of Cancer Epidemiology & Genetics (DCEG), National Cancer Institute (NCI), NIH, DHHS, Rockville, USA.
2 Regional Cancer Center, Trivandrum, India.
* Corresponding author. Senior Investigator, 6120 Executive Blvd, Rm 3046, EPS, Rockville, MD 20852, USA. E-mail: sinhar{at}nih.gov
We thank Dr Babu for highlighting several of the issues that we also were concerned about as we analysed the data presented in our paper.1
With respect to the under-representation of the vast majority of the population, it is true that the population-based cancer registries in India cover only a small proportion of the total population and include largely urban areas, although several rural areas also are represented. Rather than simply aggregating the data across the registries, such as is done in the United States, we used the national estimates from Globocan 2002 that were generated by experts at the International Agency for Cancer Research and that attempted to take into account the urban/rural and regional differences in cancer incidence rates in India.2–5
The various population-based registries in India have met quality criteria regarding the completeness and accuracy of case ascertainment and reporting, and substantial efforts have been made to minimize underreporting of cases. The proportion of cases reported solely by death certificate ranged from 1% to 9% in the various Indian registries, somewhat higher than in the UK or United States, but comparable to many other international registries. This suggests that some non-fatal cases also may have been missed. Rates for several cancers, however, such as the oral cavity and pharynx, esophagus, larynx among males and cervix uteri, were higher in India than elsewhere; the true rates for these cancers then must be even higher.
It is true that a cancer that is never diagnosed will not contribute to the incidence statistics; however, we were not in a position to quantify this or to evaluate whether this was a more severe problem in India than elsewhere. To take into account differences in age distributions across the study groups, all rates were directly adjusted using the World population standard. With respect to the population groups, we noted in our manuscript the heterogeneity in definitions of East Asians used in the various areas that we studied, but to our knowledge more precise data were not available.
Many of the potential difficulties in data completeness, accuracy and representativeness pertain not only to India but also to Singapore, the UK and the United States. As well as showing possible differences that may be present in cancer rates in different countries, we wanted to highlight the need to improve cancer research among East Asians in India as well as places where substantial migration has occurred. Such ecological studies have provided clues for further study of certain cancers and/or exposures. Thus, we attempted to gather and analyse the best data available and interpret the findings subject to a number of caveats. We feel that a particular strength of our analysis was the presentation of data for a wide range of cancers, revealing a variety of patterns by gender, cancer and geographical area that could be considered in light of potential diagnostic and reporting biases as well as suggesting real differences in risk that might be of etiological significance.
We appreciate the Commentary on our article provided by Dr Sankaranarayanan, who agreed that difficulties may be encountered but that migration studies may provide important clues to the role of environmental and ethnic factors in disease etiology and stimulate further in-depth epidemiologic studies and cancer control interventions.6 We agree that additional research is needed to document the incidence patterns, especially in rural areas of India. It was our hope that our analyses would reveal incidence patterns that would provoke thoughtful discussion and stimulate additional study on the role of environmental and lifestyle factors as well as possible diagnostic and screening practice differences and thus further our understanding of cancer causation and ultimately its prevention.
References
1 Rastogi T, Devesa S, Mangtani P, et al. Cancer incidence rates among South Asians in four geographic regions: India, Singapore, UK and US. Int J Epidemiol (2008) 37:147-a–60.
2 Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2002: Cancer Incidence, Mortality and Prevalence Worldwide. In: IARC Cancer Base No. 5, version 2.0 (2004) Lyon: IARC Press.
3 Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin (2005) 55:74–108.
4 Parkin DM, Whelan S, Ferlay J, Teppo L, Thomas DB. Cancer Incidence in Five Continents Volume VIII. In: IARC Scientific Publication No. 155 (2002) Lyon: IARC Press.
5 Parkin DM. International variation. Oncogene (2004) 23:6329–40.[CrossRef][Web of Science][Medline]
6 Sankaranarayanan R. Commentary: cancer incidence among Asian Indians in India and abroad. Int J Epidemiol (2008) 37:160–61.
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in Int. J. Epidemiol.:
- Response to ‘Cancer incidence rates among South Asians in four geographic regions: India, Singapore, UK and US’
- Giridhara R Babu
Int. J. Epidemiol. 2009 38: 1157-1158.[Extract] [FREE Full Text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||