Greetings from the National Library of Medicine and MedlinePlus.gov
Regards to all our listeners!
I'm Rob Logan, Ph.D., senior staff, U.S. National Library of Medicine (NLM).
Here is what's new this week in To Your Health, a consumer health-oriented podcast from NLM, that helps you use MedlinePlus to follow up on weekly topics.
A recommendation to lower P values from .05 to .005 may not be a well-tailored solution to enhance the quantitative rigor of medical research, suggests a viewpoint recently published in the Journal of the American Medical Association.
John Ioannidis M.D., Stanford University School of Medicine, explains recent recommendations from the American Statistical Association, as well as a coalition of statistical methodologists, proposes lowering the P value to establish statistical significance from .05 to .005. The P value is a predesignated threshold researchers use to establish whether some research findings are statistically significant, or unusual enough to declare a finding to be empirically meaningful.
Dr. Ioannidis' viewpoint notes the use of .05 P values is ubiquitous in medical research; about 96 percent of medical and public health research articles (that report P values in their abstract, full text, or both) include some values of .05 or less. Hence, a proposal to reduce it to establish more quantitative rigor would attract widespread attention — if not high anxiety — among medical researchers.
As Ioannidis explains, (and we quote): 'Moving the P value threshold from .05 to .005 will shift about one-third of the statistically significant results of past biomedical literature to the category of just suggestive' (end of quote).
However, Dr. Ioannidis counters (and we quote): 'by itself a P value does not provide a good measure of evidence' (end of quote). Ioannidis adds (and we quote): 'a P value, or statistical significance, does not measure the size of an effect or the importance of a result' (end of quote).
At best, Ioannidis notes (and we quote): 'Lowering the threshold of statistical significance is a temporizing measure. It would work as a dam that could help gain time and prevent drowning by a flood of statistical significance, while promoting better, more durable solutions' (end of quote).
Dr. Ioannidis concludes some of these solutions might include abandoning P values or tests of statistical significant altogether. He writes (and we quote): 'Because the proposed threshold … is imperfect, other more difficult but more durable alternative solutions should be contemplated' (end of quote). Among these, Ioannidis explains more focus on effect sizes and their uncertainty might be more illuminating.
Instead of a simple solution to lower statistical thresholds, Ioannidis notes (and we quote): 'a fresh start and a rapid acceleration of better practices is always possible' (end of quote).
We should note Dr. Ioannidis' viewpoint was listed first in trending (or the most highly viewed article) by JAMA viewers soon after its publication and perhaps is generating what he intended to achieve; widespread discussion and commentary among the physicians and scientists whose research results are the foundation of evidence-based medical care.
Meanwhile, an introduction to understand clinical research and clinical trials (from the National Institute of Child Health and Human Development) is available within the 'start here' section of MedlinePlus.gov's understanding medical research health topic page.
Clinicaltrials.gov, an NLM website, also explains clinical trials and provides a glossary of common research terms. To find this helpful information, please click on 'about studies' which is on the right side of clinicaltrials.gov's home page.
To find MedlinePlus.gov's understanding medical research health topic page, please type 'medical research' in the search box on MedlinePlus.gov's home page, then, click on 'understanding medical research (National Library of Medicine).'
Before I go, this reminder... MedlinePlus.gov is authoritative. It's free. We do not accept advertising .... and is written to help you.
To find MedlinePlus.gov, just type 'MedlinePlus.gov' in any web browser, such as Firefox, Safari, Chrome, or Explorer on any platform.
We encourage you to use MedlinePlus and please recommend it to your friends. MedlinePlus is available in English and Spanish. Some medical information is available in 48 other languages.
A written transcript of recent podcasts is available by typing 'To your health' in the search box on MedlinePlus.gov's home page.
The National Library of Medicine is one of 27 institutes and centers within the National Institutes of Health. The National Institutes of Health is part of the U.S. Department of Health and Human Services.
A disclaimer — the information presented in this program should not replace the medical advice of your physician. You should not use this information to diagnose or treat any disease without first consulting with your physician or other health care provider.
It was nice to be with you. Please join us here next week and here's to your health!