PubMed · 10871985
The source did not provide an abstract. Follow the original record for more information.
Explore connections, maps & timelines
R Karis, D Karis. 2000-06-26. Intercessory prayer.. https://doi.org/10.1001/archinte.160.12.1870-a
Cite the original work for its findings. Save a collection to share your selection of sources.
KEEP EXPLORING
A new test procedure is presented for the problem of testing whether a treatment is better than a control when there is ordered categorical data. The new test is based on the methodology developed for general 'one-sided' alternatives by Cohen and Sackrowitz. An evaluation of the test is made by comparing the power and expected p-values of the test with those of the Wilcoxon-Mann-Whitney test in examples. As predicted by the theoretical work by Cohen and Sackrowitz, the new test is seen to be preferable to the Wilcoxon-Mann-Whitney test. Computer programs to assist implementation of the new test are made available.
BACKGROUND: The authors have shown that the primary determinants of prostate carcinoma progression are tumor volume and the percent of the tumor comprised of Gleason Grade 4/5 cells. In the current study the authors evaluated six different techniques for the morphometric measurements of prostate carcinoma volume. METHODS: A computer-assisted image analysis (NIH Image, developed and maintained by the National Institutes of Health, Bethesda, MD) was used to analyze all 108 step-sectioned prostate specimens obtained between January 1 and December 31, 1997. The authors used the Stanford technique of 0.3-cm step-sections, measuring the volume of the tumor at both 0.3-cm and 0.6-cm intervals. The other 4 methods included the authors' previous method based on an earlier image program, the ellipsoidal method (pi / 6 x width x height x length), an estimation of the square area of the largest tumor, and the maximum tumor dimension (MTD). RESULTS: The authors first checked the accuracy of NIH Image analysis by measuring 24 circles of widely different sizes. The mean coefficient of variation was 1.7% and the correlation between the mean circle areas measured by the NIH Image software and true circle area essentially was perfect (correlation coefficient [r] = 1 and r(2) = 0.999; P < 0.0001). In comparison with the authors' original computer image program using 0.3-cm step-sections measured by a different observer, r(2) with the NIH Image analysis was 0.93. Using NIH Image only, the 0.6-cm step-section method missed measurable cancers in 16.7% of 108 radical prostatectomies in comparison with the 0.3-cm step-method. The mean tumor volume with the 0.6-cm section method (P < 0.0001) and the ellipsoidal method (P < 0.05) were significantly higher than with the 0.3-cm section method. r(2) from linear regressions using the 0.3-cm step section method as the standard versus the ellipsoidal method was 0.594, and was 0.89 versus the 0.6-cm step-section method, 0.652 versus the square area estimation, and 0. 527 versus the MTD method. CONCLUSIONS: The results of the current study support NIH Image as a powerful software program for the morphometric measurement of prostate carcinoma volume. Pathologic processing with 0.3-cm section slices was found to be more accurate for tumor volume than the 0.6-cm section slices. The ellipsoidal method, the square area of the largest tumor, and the MTD all were found to be inferior to computer-assisted image analysis measurements. In certain clinical situations in which only estimates of tumor volume are required, the square area of the largest tumor appears to be the best choice (r(2) 0.652).
Longitudinal studies are characterised by repeated measurements of the outcome variable on the same individuals. This implicates that observations are correlated and that statistical techniques, such as linear regression analysis cannot be used. Special longitudinal techniques correct for the dependency in repeated observations by analysing the development of a certain individual in time. Depending on the research question, the distribution of the outcome variable, and the number of repeated measurements, a certain technique should be chosen. Sophisticated longitudinal techniques like generalized estimating equations and random-coefficient models are most recommendable.