 |
|
A Word From Our Moderator
|
|
| Since this article does not have an abstract, we have provided the first 150 words of the full text and any section headings. |
|
|
|
To the Editor: In their Commentary, Dr Kraemer and colleagues1 raise the important issue of exploring moderators of treatment outcome in randomized trials. I agree that reliance on the P value for hypothesis testing has limited value for clinicians who need to decide if a treatment is appropriate for a particular patient. There are also 2 other points to consider.
First, the hypothesis-generating exercise that the authors promote may be valuable. However, I believe it must be accompanied by a rational discussion of possible mechanisms based on known or potential mechanisms from observational and basic science studies. If the hypotheses are supported by these data, there is a greater probability that the hypotheses will be supported by future research. In addition, the process may help future researchers design the most appropriate studies by avoiding confounding and by examining the appropriate interactions through restricted inclusion/exclusion criteria or stratification.
Second, although the . . . [Full Text of this Article]
Ian Shrier, MD, PhD
ian.shrier@mcgill.ca
Centre for Clinical Epidemiology and Community Studies
SMBD-Jewish General Hospital
McGill University
Montreal, Quebec
RELATED LETTERS
A Word From Our Moderator
Japp Deinum and Gert Jan van der Wilt
JAMA. 2007;297(2):156.
EXTRACT
| FULL TEXT
A Word From Our Moderator—Reply
Helena C. Kraemer, Ellen Frank, and David J. Kupfer
JAMA. 2007;297(2):157.
EXTRACT
| FULL TEXT
RELATED ARTICLE
Moderators of Treatment Outcomes: Clinical, Research, and Policy Importance
Helena C. Kraemer, Ellen Frank, and David J. Kupfer
JAMA. 2006;296(10):1286-1289.
EXTRACT
| FULL TEXT
|
