Twenty-seven years ago I found what seemed to be the only functioning storm-drain in Tanzania, in East Africa, and fell down it, severely injuring a knee in the process. The journey to the mission hospital in the back of a pick-up truck over sixty miles of rutted laterite road was one of the more agonising experiences of my life.

I had an arthroscopy when I returned home several weeks later — I could not even hobble until then — and the orthopaedic surgeon told me that unless I did physical therapy every day for a very long time it was inevitable that I should be crippled by arthritis within twenty years.

It was equally inevitable that I would not do physical therapy every day for a long time; and here I am, twenty-seven years later, without so much as a twinge from my knee. My faith in the predictive powers of orthopaedic surgeons has been somewhat dented.

That was why I read with interest a paper in a recent edition of the New England Journal of Medicine comparing physical therapy with surgery for meniscal tears in the knees of people with osteoarthritis. To cut a long story short, there was no difference in outcome, an important finding, since 465,000 people undergo operations for precisely this situation every year in the United States alone.

Actually, the uselessness of operation had been established before — the uselessness from the patients’ point of view, that is. Two previous trials had compared real with sham operations, and with no operations at all, and found no difference in the outcome two years later. One might suppose that, in the light of these findings, the 465,000 operations still performed annually constituted something of a scandal.

The clinical trial reported in the NEJM is, like all such trials, not definitive. The follow-up period was only 6 months, relatively few patients were recruited to it, and some patients initially allocated to physical therapy had an operation nonetheless for reasons that are not entirely clear. Moreover, the trial is only that of operation versus physical therapy; strictly speaking, there should also be a comparison with patients who had no treatment at all.