American Journal of Epidemiology Advance Access originally published online on June 2, 2006
American Journal of Epidemiology 2006 164(2):126-127; doi:10.1093/aje/kwj195
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Response to Invited Commentary |
Hardy et al. Respond to "Beyond Frequencies and Coefficients"
1 Medical Research Council National Survey of Health and Development, Department of Epidemiology and Public Health, Royal Free and University College London Medical School, London, United Kingdom
2 Department of Social Medicine, University of Bristol, Bristol, United Kingdom
Correspondence to Dr. Diana Kuh, Medical Research Council National Survey of Health and Development, Department of Epidemiology and Public Health, Royal Free and University College London Medical School, Gower Street Campus, 1-19 Torrington Place, London WC1E 6BT, United Kingdom (e-mail: d.kuh{at}nshd.mrc.ac.uk).
Received for publication March 15, 2006. Accepted for publication March 20, 2006.
The invited commentary by Wang (1
) on our paper on the developmental origins of midlife physical performance (2) is a critical appraisal of prevailing epidemiologic approaches to life course analyses rather than the substantive findings of our study.
We share many of the interests and concerns raised by the commentary. Epidemiologists share an interest with researchers from many other disciplines in investigating the critical and cumulative (3) effects of childhood developmental factors on aging in later life. We agree that epidemiologists have focused for too long on risk factors for disease-specific outcomes rather than on underlying aging processes that may be common to a number of age-related diseases, a view shared by recent governmental reports in the United Kingdom (4). Indeed, our choice of measures of standing balance and chair rising was because these functional outcomes may be summary indicators of these aging processes and are associated with subsequent frailty, disability, and death.
We also agree that the interdisciplinary and integrative life course approach is a valuable and increasingly popular approach to the study of aging but is very challenging methodologically. Different disciplines have developed a variety of statistical approaches to handle the complexity of time-dependent covariates that may be confounders or intermediaries on the pathway linking development and aging. The statistical approaches suggested by Wang are complementary rather than alternative to the "classic risk factor- or variable-based approaches" (1, p. 124) used in our paper. Traditional regression models using, for example, multiple measures of growth can provide useful insights when dealing with temporally dependent exposures as long as care is taken in model building and the correct conditional interpretation applied (5, 6). Data quality issues, such as measurement error and missing data, also need to be considered more fully in life course analyses (5, 7), and Wang correctly raises the problem of residual confounding in relation to adjustment for socioeconomic position.
Thus, Wang and we agree that the findings that we report are valuable because, by identifying the periods of growth that have an impact on later life functioning and how they relate to socioeconomic circumstances, they provide the basis on which investigators can ask additional research questions and build further analyses. Whatever the method of analysis used, the presentation of results should be transparent and easily understood. It is reassuring if results from more complex methods are similar to those from simpler methods. If they reveal differences, it may be due to methodological differences or because the methods address different questions. Similarly, where associations between development and aging are robust across different studies using different methodologies, this provides stronger evidence on which to base public health interventions. If associations vary, then this may be due to different methodologies, or it may be that the links between development and aging are time and place dependent.
The choice of statistical method is dependent on a clear conceptual life course framework and underlying theoretical models that can generate hypotheses to test. This is a critical point that we have raised elsewhere (3, 8–10) and is relevant for not just life course epidemiologists but also life course researchers from other disciplines. Wang is right to highlight the need to test individual life course trajectories within a broader ecosocial context. We have repeatedly stated that life course models and ecosocial models need to be integrated as they are complementary and provide added value. We have proposed conceptually how one might do this (3). However, this requires data sets that have sufficient heterogeneity at different contextual levels. This may not be possible to test within a single data set and may require collaborative analyses across data sets. We are currently seeking funding to apply this integrated approach across a number of life course studies.
Longitudinal studies that started many years ago may not have all the data required for a full life course model, and researchers should focus on the particular strengths of each study. The strengths of the National Survey of Health and Development, for example, lie in the measures of body size and cognitive function across the life course utilized in our paper. The study is less able to address Wang's question regarding socioeconomic position-stratified, age-related dietary changes before and after age 7 years, as data on diet and physical activity in childhood are more limited. Those running more recent cohort studies following children into adult life need to make careful priorities about the information to be collected that will allow comparative research with older cohorts and fill the research gaps that remain.
 |
ACKNOWLEDGMENTS |
|---|
Conflict of interest: none declared.
|
References |
|---|
 TOP References |
|---|
- Wang C. Invited commentary: beyond frequencies and coefficients—toward meaningful descriptions for life course epidemiology. Am J Epidemiol 2006;164:122–5.
[Free Full Text] - Kuh D, Hardy R, Butterworth S, et al. Developmental origins of midlife physical performance: evidence from a British birth cohort. Am J Epidemiol 2006;164:110–21.
[Abstract/Free Full Text] - Ben-Shlomo Y, Kuh D. A life course approach to chronic disease epidemiology: conceptual models, empirical challenges, and interdisciplinary perspectives. Int J Epidemiol 2002;31:285–93.
[Free Full Text] - House of Lords Science and Technology Committee. Ageing: scientific aspects. London, United Kingdom: The Stationery Office Limited, 2005.
- De Stavola BL, Nitsch D, dos Santos Silva I, et al. Statistical issues in life course epidemiology. Am J Epidemiol 2006;163:84–96.
[Abstract/Free Full Text] - Cole TJ. Modeling postnatal exposures and their interactions with birth size. J Nutr 2004;134:201–4.
[Abstract/Free Full Text] - De Stavola BL, dos Santos Silva I, McCormack V, et al. Childhood growth and breast cancer. Am J Epidemiol 2004;159:671–82.
[Abstract/Free Full Text] - Kuh D, Ben-Shlomo Y, Lynch J, et al. A glossary for life course epidemiology. J Epidemiol Community Health 2003;57:778–83.
[Abstract/Free Full Text] - Kuh D, Hardy R. A life course approach to women's health: linking the past, present and future. In: Kuh D, Hardy R, eds. A life course approach to women's health. Oxford, United Kingdom: Oxford University Press, 2002.
- Ben-Shlomo Y, Kuh D. Conclusions. In: Kuh D, Ben-Shlomo Y. A life course approach to chronic disease epidemiology. 2nd ed. Oxford, United Kingdom: Oxford University Press, 2004.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||