Various groups of higher beings keep watch over the lives of us humble mortals, including the American Council on Science and Public Health, which here pronounces on reducing sodium intake to prevent cardiovascular disease. “A 1.3g/day lower lifetime sodium intake translates into an approximately 5mm Hg smaller rise in blood pressure as individuals advance from 25 to 55 years of age, a reduction estimated to save 150 000 lives annually.” Whether or not you question the extrapolations behind this statement, there is no doubt that this reduction could be achieved easily by lowering the salt content of processed food.

Which patients are most likely to get a deep vein thrombosis? Those in hospital, of course. And which patients outside hospital are most likely to get a DVT? Those who have just left hospital, according to this study conducted in Worcester, Mass. In fact more VTEs were diagnosed after the patients had left hospital than those diagnosed while they were still there. The other big risk factors in the community are cancer and previous DVT; oestrogen-taking is a very minor risk by comparison.

Poor reading fluency independently predicts all-cause mortality and cardiovascular death among elderly people in American cities, after adjusting for baseline health and socioeconomic status. For some reason, the authors equate reading fluency with “health literacy”, whatever that may be.

How often do you check the serum thyrotropin (TSH) levels of your patients when they come back year after year feeling tired all the time, and putting on weight though they “hardly eat anything”? In extreme desperation, I’ve been known to do it more than once in the same year. But this important study of TSH levels measured in a large network of US primary care physicians shows that they vary very little over 5 years. More than a half of patients with an isolated abnormal result are normal on repeat testing, and only 2% became abnormal within 5 years of a normal test. Do fewer.

Some guidelines on the diagnosis of diabetes still recommend a glucose tolerance test, or at least the measurement of glucose two hours after a 75g load. In fact this is liable to much more random variation when measured twice than the fasting level or the HbA1c. Don’t bother.