QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Guo, C. Articles by Wilkens, L. R. Search for Related Content PubMed PubMed Citation Articles by Guo, C. Articles by Wilkens, L. R.

---

Letter

Use of Bootstrap Procedure and Monte Carlo Simulation

Cancer Research Center of Hawaii University of Hawaii Honolulu, HI 96813

Dear Dr. Suttie:

We would like to comment on the article entitled, "Application of the Bootstrap Procedure Provides an Alternative to Standard Statistical Procedures in the Estimation of the Vitamin B-6 Requirement" (Hansen et al. 1999 ). Bootstrap is a novel approach with many applications in nutrition. We are glad to see it presented in a nutrition journal. Although the paper provides a good review of the bootstrapping methodology, we would like to point out two statistical issues to consider if the procedure is to be applied more broadly: repeated measures on each subject and multiple subjects at each dose.

In the depletion-repletion study that the authors use to illustrate the bootstrap procedure, eight young women were measured repeatedly for each of six vitamin B-6 status indicators. Longitudinal data such as these require special statistical methods because the set of observations on one subject tends to be correlated (Diggle et al. 1994 ). In the graphical display of data, repeated measurements for the same subject are often connected with lines to accentuate the longitudinal nature of the study. If we apply this to the data shown in Figure 1 of their study, there appears to be some degree of correlation among repeated measurements of urinary 4-pyridoxic acid (4-PA) excretion within individuals. Thus, although the authors report that the repeated observations for their subjects were uncorrelated, this might be due to the small size of their study. Failure to account for the correlation in repeated observations often results in a confidence interval that is too narrow and a false statistical significance. Therefore, the intrinsic correlation in repeated observations must be taken into account to draw valid scientific inference. In their estimation of the vitamin B-6 requirement, it might be more appropriate to assume that there is a positive correlation between any two measurements on the same subject and to use weighted least-squares estimation instead of the ordinary least-squares used in this application.

To perform bootstrapping in this context, each subject should be sampled with replacement rather than the individual observations. It also would be interesting to see how the bootstrap estimate would change if the mean baseline value of urinary 4-PA excretion at the end of the adjustment period was reevaluated in each bootstrap sample and applied to the inverse prediction.

We also would like comment on the design of Monte Carlo simulation. It appears that the authors did not take into account the fact that multiple subjects are measured at each level of vitamin B-6 intake in a depletion-repletion study. A valid Monte Carlo simulation study of the performance of their estimators requires that, for each dose of vitamin B-6 intake (X_i), several subjects (Y_i) must be generated from their simulation model: Y_i = ß₀ + ß₁X_i + _i. With a single subject per dose, their Monte Carlo simulation may not be able to assess the performance of estimators as well as the authors have claimed.

REFERENCES

1. Hansen C. M., Evans M. A., Shultz T. D. Application of the bootstrap procedure provides an alternative to standard statistical procedures in the estimation of the vitamin B-6 requirement. J. Nutr. 1999;129:1915-1919[Abstract/Free Full Text]

2. Diggle P. J., Liang K. Y., Zeger S. L. Analysis of Longitudinal Data 1994 Oxford University Press New York, NY.

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Guo, C. Articles by Wilkens, L. R. Search for Related Content PubMed PubMed Citation Articles by Guo, C. Articles by Wilkens, L. R.