GROWTH FAILURE: STATISTICS APPLIED TO MEDICAL THERAPY

Abstract

To minimize subjective bias and to shorten control and trial-treatment periods, especially when using hazardous or scarce medications, we applied statistical methods to evaluate the influence of treatment upon the growth rates of children with growth disturbances. Utilization was encouraged by programming techniques for a desk-top calculator and demonstrating uses with six questions. The program assumes linearity of data and requires groups of paired variables, i.e. age and height. Section 1 describes growth statistically: 1. By my measurements, what is the child's growth rate? (linear regression and correlation) 2. What is my average error when measuring this child? (standard error of measurement) 3. How can I know when I have sufficient measurements to permit a therapeutic change? (F-test for regression) 4. What do I use to compare this child's growth to “normal” growth? (t test for confidence limits of slope). Section 2 compares growth periods statistically: 1. Are growth periods 1 and 2 different? (linear covariance for comparison of slopes).2. If different, how do they differ? (linear covariance discriminates change in rate from parallelism due to difference in measurement techniques or non-consecutive periods). The program can also be used to facilitate evaluation, retrospective or prospective, of a wide variety of influences on linear functions ranging from pharmacokinetics to applications in behavioral science, e.g. comparison of teaching techniques on rates of learning.

Log in or create a free account to read this content

Gain free access to this article, as well as selected content from this journal and more on nature.com

Access through your institution

or

Sign in or create an account

Continue with Google

Continue with ORCiD