February 17th, 2010
The Value of Statistical Modeling?
James De Lemos, MD
In this week’s “The Expert Is In” blog on CardioExchange, I interview Rodney Hayward about his very interesting and soon-to-be controversial paper just published in the Annals of Internal Medicine. Rodney and his coauthors argue that the current lipid guidelines are misguided.
Specifically, the authors show that the NCEP ATP III guideline approach, in which lipid-lowering therapy targets specific goals, is neither evidence-based nor an efficient and effective public health strategy. They argue that a tailored strategy, in which candidates for lipid-lowering therapy are identified based on global CV risk, would more accurately identify individuals who would benefit, would treat more patients, and would improve quantity and quality of life. The study design was….not a randomized controlled trial….not an well-performed observational study….but rather a mathematical model.
It would be difficult to perform randomized controlled trials testing all of the diverse strategies for targeting lipid-lowering therapies. Thus, modeling exercises are becoming increasingly popular strategies for attempting to answer these complex questions.
What you think of this methodology? Have you learned enough about modeling and statistics to be able to interpret these types of data? Are you convinced by the logic of this argument? (As for the important practical question — should this study, and will it, change your practice? — please tell us what you think
One Response to “The Value of Statistical Modeling?”
NEJM Journal Watch — Recent Cardiology Articles- Anticoagulation in Patients with Cancer-Associated Pulmonary Embolism: How Long is Long Enough?
- Spironolactone Use After Percutaneous Intervention for Acute Myocardial Infarction
- CRISPR/Cas9 Gene Editing for Transthyretin Cardiomyopathy
- Does Colchicine Reduce Cardiac Events After Myocardial Infarction?
- First-Line Therapy for Ventricular Tachycardia After MI: What's the Best Approach?
overstepping mathematical bounds
I’m a big fan of mathematical models, but in patient care, we always need to keep it real. Clinicians should absolutely demand practical, easy-to-understand validation of any diagnostic or treatment protocols based on complex models. No model is brilliant if it can only be understood by mathematical geniuses. See http://www.ox.ac.uk/research/mathematical_physical_life_sciences/index.html on the Oxford website for how math modeling can help, and what its boundaries are. You don’t have to know how the internet works to appreciate the concrete evidence of its value on your home computer. The proof is in the pudding, not the recipe. Don’t let any mathemagicians tell you otherwise.