The Bradford Hill Criteria Don't Hold Up

In 1965, the epidemiologist, Austin Bradford Hill, who helped link smoking to lung cancer, gave a speech where he presented his viewpoints/criteria on how we can arrive at causation from correlation. This lecture was a bit of a game changer at the time given that the tobacco industry was employing statisticians, medical doctors, and even popular science writers to push the idea that the relationship between smoking and lung cancer was merely a correlation, not a causal one. Read More

Is Moderate Carbohydrate Intake the Best?

Recently, a giant paper on carbohydrate consumption and mortality was published in The Lancet. The paper discussed the findings of a prospective cohort study and a meta-analysis of several cohort studies. Studies like this are often the ones that generate the most hype, which is always bizarre to me given that higher quality randomized studies almost never receive any attention. As a result of all the noise (see below), I had to discuss the study in question. Read More

Exercise, Mental Health, and Big Data

Recently, a large cross-sectional study that investigated the relationship between exercise frequency and mental health was published in The Lancet Psychiatry and also happened to set Twitter on fire. I want to discuss the good and the not so good. First, some good! Some Good The study, which included 1,237,194 adults from the US, found a significant relationship between physical exercise and self-reported mental health burden. Here’s what the authors reported, Read More

Misuse of Standard Error in Clinical Trials

Reporting effect sizes with their accompanying standard errors are necessary because it lets the reader interpret the magnitude of the treatment effect and the amount of uncertainty in that estimate. It is magnitudes better than not providing any effect sizes at all and only focusing on statements of statistical significance. Although many authors provide standard errors with the intention of relaying the amount of uncertainty in the model, there are several misconceptions about when the standard error should be reported, and it is often misused. Read More

High Statistical Power Can Be Deceiving

Even though many researchers are now acquainted with what power is and why we try to aim for high power in studies, there are still several misconceptions about statistical power floating around. For example, if a study designed for 95% power fails to find a difference between two groups, does that offer more support for the null hypothesis? Many will answer yes, because they elicit that if such a large study failed to find a difference between two groups, then this provides evidence for no effect. Read More

Chocolate Milk Is Delicious, Not Magic

Chocolate milk is one of the few drinks in sports nutrition to have a pretty good reputation. Not only is it delicious, but it’s also rich in calcium, carbohydrates, flavonoids, and electrolytes. These are nutrients that seem to aid sports performance. The conclusions froma few studies over the years have supported its use for improving sports performance. However, there are also several contradictory findings that didn’t find it to be superior to other sports drinks that contain carbohydrates and electrolytes. Read More

Seductive Surrogates Can Be Deadly

In clinical trials, it’s not always possible to measure hard endpoints like cardiovascular disease events and cancer remission rates. Studies that use clinical outcomes often dichotomize these variables, and as a result, they need to have a large number of participants and be long in duration to detect differences between groups. Again, this type of research is expensive and not always feasible. In many scenarios, a more practical alternative is to focus on intermediate markers. Read More

Vitamin E, Mortality, and the Bayesian Gloss

Bayesian data analysis is beginning to gain traction in several fields. Some of those reasons include that it allows individuals to represent uncertainty using probability distributions and it helps them avoid losing information that’s typically lost with point estimates and dichotomization. Bayesian inference also allows for relevant background information to be incorporated into a model using a more continuous approach rather than making binary decisions about what to include. Read More

Does Protein Increase the Risk of Heart Failure?

About three weeks ago, a cohort study was published in Circulation that claimed that protein consumption was associated with heart failure. The press reacted as I expected them to. If you read some of these articles, most of them seem to conclude that a high protein diet is probably not good for you and that Americans eat too much protein. Anyway, back to this study. Given the nature and limits of these types of studies (you can read more about that here) I was a bit skeptical, but also open to the idea that there might be a possible relationship between increased protein consumption and an increased rate of heart failure. Read More

Problems with the Number Needed to Treat

The number needed to treat (NNT) is a popular statistic used in medicine and its use is even encouraged by groups like Cochrane and CONSORT. Why is it so popular? Most believe that the NNT is more understandable than effect sizes like odds ratios or risk ratios or statistics like the absolute risk reduction. The NNT is also believed to convey more meaningful information. In this blog post, I am going to discuss: Read More