IJE Advance Access originally published online on February 14, 2008
International Journal of Epidemiology 2009 38(1):323-324; doi:10.1093/ije/dyn006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letters to the Editor |
Author's Response
Johns Hopkins Bloomberg School of Public Health, Baltimore MD, USA.
* Corresponding author. Johns Hopkins Bloomberg School of Public Health, Baltimore MD, USA. E-mail: gchan{at}jhsph.edu
We appreciate the close attention Aaby et al. have given to our work. Deciding how to conduct the analysis of observational studies requires balancing potential biases, and we recognize the choices we made are not necessarily those Aaby et al. would have made. Although much of their letter is of a general nature, we think it useful for us to reply to their three main comments regarding our analysis.
Point 1: Risk-free survival time
Aaby and colleagues raised the concern that our study contained ‘a special form of survival bias, pre-survey survival time being added to the analysis for those who survived to the time of the first survey’. They also questioned ‘whether many of the 4,103 children have contributed risk-free time to the retrospective analysis’ (Aaby letter).
To address the concern that there was risk-free survival time, we re-ran the analysis so that:
- the timeline starts at zero at the latest of BCG receipt and first study visit for those who entered in the baseline survey;
- the timeline starts at zero at BCG receipt or end date of the baseline survey (if BCG receipt occurred before the end of the baseline survey period) for those who entered after the baseline survey.
|
Point 2: An ‘increasingly selected frail subgroup’ among those BCG vaccinated but not yet DTP vaccinated
The second concern raised by Aaby et al. addresses a potential non-measurable difference between those who received DTP vaccination and those who did not (Aaby letter). They write ‘in most situations, healthy children are vaccinated first and by implication those remaining in the previous vaccination group will increasingly tend to be unhealthy children too frail or too poor to get vaccinated’. Despite the speculative nature of this argument, it points to a fundamental and well-known shortcoming of the previous studies of the non-specific effects of DTP with respect to child mortality. All studies on this subject to date have been observational, and, as summarized by the WHO Task Force on Routine Infant Vaccination and Child Survival, ‘it is impossible to be sure that unmeasured confounding variables did not influence the findings’ in these studies.1 While the universe of potential confounders is virtually limitless, we feel that we did as much as possible with our data in controlling for the composition of its subgroups. In this study, we chose only to examine children who at least received a BCG vaccination, and therefore were sufficiently non-poor and non-frail to have some access to vaccines. Moreover, as of 6 months of age, only 68% of the infants had received a dose of DTP, implying that during the period of highest mortality, there was not a strong selective pressure; if it had been 97% coverage by 6 months it would have been more problematic.
Point 3: Sex differential effects of vaccines related to the timing of vaccinations
The goal of our analyses was to isolate the association of DTP receipt with mortality among BCG-vaccinated children, with a focus on sex differentials. This was in response to the call of the WHO Task Force for further investigation of the hypotheses that had been raised by Peter Aaby and his colleagues. Our data, however, are not extensive enough to address the further elaboration on these hypotheses as mentioned in the letter of Aaby et al.2 With only two deaths among girls not vaccinated with DTP, further stratification of the data, or the addition of covariates to examine a greater number of subgroups, cannot be sustained. As an alternative way to investigate the effects of methodological differences, Aaby and colleagues could apply our relatively straightforward approach to their data sets.
Conclusion
Our study results are consistent with the studies done in Papua New Guinea, Burkina Faso, India and Bangladesh.3–6 Although not all of these studies examined the sex-differential hypothesis as closely as we did, they all indicated a survival benefit associated with DTP receipt. Some or all of this benefit may be due to selection factors. The similarity between our SES-adjusted and unadjusted results, however, gives indirect evidence that selection may not have played a large role in our study. We recognize there may be more than one way to interpret our specific finding that boys with DTP and all girls had lower mortality than boys without DTP, depending on the extent of sex-differential selection and mortality. A relatively simple explanation of our results is that receipt of DTP among boys helps them overcome the disadvantage they have relative to girls early in life.
References
1 WHO Task Force on routine Infant Vaccination and Child Survival. (2004) Report of a meeting to review evidence for a deleterious effect of DPT vaccination on child survival.
2 Peter A, Christine SB, Jens N, Henrik R. Sex-differential non-specific effects of BCG and DTP in Cebu, The Philippines. Int J Epidemiol (2009) 38:320–23.
3 Breiman RF, Streatfield PK, Phelan M, Shifa N, Rashid M, Yunus M. Effect of infant immunisation on childhood mortality in rural Bangladesh: analysis of health and demographic surveillance data. Lancet (2004) 364:2204–11.[CrossRef][Web of Science][Medline]
4 Lehmann D, Vail J, Firth MJ, de Klerk NH, Alpers MP. Benefits of routine immunizations on childhood survival in Tari, Southern Highlands Province, Papua New Guinea. Int J Epidemiol (2005) 34:138–48.
5 Moulton LH, Rahmathullah L, Halsey NA, Thulasiraj RD, Katz J, Tielsch JM. Evaluation of non-specific effects of infant immunizations on early infant mortality in a southern Indian population. Trop Med Int Health (2005) 10:947–55.[CrossRef][Web of Science][Medline]
6 Vaugelade J, Pinchinat S, Guiella G, Elguero E, Simondon F. Non-specific effects of vaccination on child survival: prospective cohort study in Burkina Faso. Br Med J (2004) 329:1309.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||