The QM Bibliography can be reached by clicking here. This is a bibliography of 120 key papers in heath and medicine. Each paper has a brief description and a non-pay-walled link. You can read them yourself. Don’t be put off by the jargon. It is easily mastered. In fact, here’s a cheat sheet.

HOW TO READ MEDICAL RESEARCH

In our opinion, doctors write better than physicists. Nonetheless, the readability of the papers runs the gamut from lucid to incomprehensible. However, unlike physics, statistics tend to pervade all the papers, which can be confusing, sometimes deliberately.

The jargon can be a bit imposing at first, but here is a short translation guide:

• Cohort – Group of people.
• Epidemiology – a record set on a large group of people, health and lifestyle records typically.
• Epidemiological Study – analysis of an epidemiology.
• Efficacy – success, ability to produce desired result.
• Prospective Study – prospective studies follow a group of people (a cohort) for development of disease and attempts to capture or control for factors such as lifestyle that lead to the disease of study.
• Retrospective Study – A retrospective study is similar, but looks at people that have already contracted a disease and attempts to associate that with some possibly causative historical factors; more or less going back over people’s medical records and personal histories to see if the contributing or causative factors can be captured.
• Cross sectional study – cross sectional studies are a one shot event typically. A group of people could be asked to fill out a medical history and current eating and drinking habits to attempt to determine causes of lung cancer.
• Case-control studies – Here the outcomes are gathered, and the histories examined for possible causes. Dr. Nichols could divide his cohort into a high triglyceride group (the “Cases”) and a normal triglyceride group (the “Controls”), and explore what prior lifestyle choices had been made.
• Hazard Ratio (HR), Relative Risk (RR) – odds of an event if the experimental group got 15% less cancer than the control group, the HR would be 0.85. Technically these are slightly different, HR being instantaneous and RR the average over the experiment.
• Confidence interval – a confidence interval, 95% for instance, will be stated with a pair of numbers representing the interval. It means that 95 percent of the results fell within the interval. Sometimes the interval is absurdly wide.
• Confounding factors – other things that could affect the results. Heavy smokers are often heavy drinkers. When they say ‘controlling for confounding factors’, it means they took such things into account.
• Controlled for age – when comparing disease between two different groups, results may be skewed because one group is younger. Controlling for age adjusts this.
• Quintiles – splitting a cohort into five groups based on some attribute, perhaps smoking. The lowest quintile would include the non-smokers and the highest quintile would cover the chain smokers. They would then likely compare something else in the quintiles, like lung cancer.
• Pre-prandial and post-prandial – before and after a meal.
• Fasting glucose (or whatever) – the level 12 hours after the most recent meal.

Here are a few more tips;

• If the cohort is small, say under a couple of hundred, the paper will probably not talk in statistical terms, though it might.
• Beware of papers that just talk about a single item without addressing side effects. These are numerous. Frequently the possible side effects will be mentioned but summarily dismissed. It is best to select papers that state all-cause-mortality, as this tends to catch all relevant side effects.
• Many researchers are not statisticians, and may simply plug numbers into an available stat program. This can be hazardous.
• We are not mice. If a drug works on mice, it isn’t necessarily going to work on us. Resveratrol would be a good example of this. If the drug doesn’t even work on mice, it won’t likely see the light of day.
• In vitro versus in vivo. This is cute talk for cultures in a petri dish versus a living animal. Again, just because something kills cancer in a lab cell culture doesn’t mean it would work in humans. It is a lot harder for a drug to get into a cell somewhere in our body than to get to a cell in a petri dish.
• Look for obvious conflicts of interest, such as employment by a pharmaceutical or packaged food company, or by funding of any of these.
• If there is a lot of money involved, look very hard indeed.
• If results are strong, you don’t need 10,000 cases. If 24 of 25 people responded to some therapy, that is a strong result.
• ft the benefit is more or less a toss-up, you will see studies involving tens of thousands. Conversely, if a study involves tens of thousand, expect a toss-up. Vitamins would be a good example of this.

The bibliography covers these topics. Simply search or scroll to find those of interest. Again – click here for the bibliography.

Click here for the bibliography.

The final exam will be in three weeks.