Saving My Knees: How I Proved My Doctors Wrong and Beat Chronic Knee Pain by Richard Bedard Page A

suffered from chronic knee pain. Despite that, their defects improved at about the same rate as those of their pain-free counterparts.
    The authors did identify factors that led to an increase in cartilage defects over time. Being overweight or female boosted your risk. Being over forty did too. But there was also good news buried in the age-related statistics. Subjects over forty and under shared roughly the same chance of having faulty cartilage improve somewhere in their knee. That suggests the ability of the tissue to heal doesn’t flip off, like a toggle switch, once you reach a certain age.
    A huge risk factor for worsening cartilage was having bone spurs, outgrowths that form when the body tries to restore stability to a decaying joint. The bony protrusions often restrict normal leg movement and damage or interfere with the remaining cartilage. From that revelation, I took the message that knee pain sufferers should take early action, before the joint undergoes significant structural changes.
    Overall, the study seemed to support much more hope than despair. Those subjects whose cartilage got better didn’t even follow any knee-friendly exercise regimen, as with the Swedish experiment.
    At this point, a skeptical reader might protest. There’s an unacknowledged elephant in the room: measurement error. MRI technology isn’t perfectly accurate. Also the resulting images are literally black-and-white, but their interpretation isn’t. Opinions can vary among those reading the same MRI.
    The Australian researchers took pains to address the potential for human error at least. The same person graded all the MRIs at the beginning and end of the study. Also he evaluated the final set of MRIs without knowing any of the initial scores, to avoid prejudicing him.
    Even so, some errors probably crept in. Say the human grader initially scores a cartilage lesion a “three,” but it’s actually a “two.” Two years later, he rechecks it. The defect is still a “two.” He correctly marks it as such but incorrectly records the lesion as having improved. The problem lies with his early mismeasurement.
    Of course committing such errors can skew the figures both ways. He also may incorrectly measure and fail to capture an improvement.
    It’s unclear how eliminating measurement error would have affected the results. That’s not so for a related issue: the problem of how precise the measuring tool is. Once you factor this in, it becomes clear that the study probably undercounted the number of defects that improved—perhaps by a lot.
    That’s because the precision of a measuring tool can matter a good deal when quantifying change. This is an important point to grasp. Here’s a simple illustration of what I mean:
    Imagine you’re monitoring the growth of teenagers, year to year. There’s only one catch: you’re using a very crude ruler. Whereas normal rulers display not only inches, but also eighths or sixteenths of an inch, this ruler measures only in feet. So for example, anyone who is more than four feet tall, but no more than five, is considered five feet in height.
    Say Carol is five-three, Bill five-eight, and Jim five-eleven. They’re all marked down as six feet tall (more than five feet, less than six). Now let’s suppose the purpose of this experiment is to quantify the percentage of teenagers that grow from year to year.
    If, one year later, Carol has spurted three inches, Bill has gained two, and Jim one and a half, what will our statistics show, after we’ve measured everyone with our foot-long rulers? That’s right: only one-third of our subjects have grown. Carol and Bill are still six feet tall. Jim is now seven feet.
    The problem: our measuring device lacks sufficient precision.
    Now apply that idea to the cartilage defect study.
    The researchers employed a commonly used zero-to-four scale to grade the quality and thickness of cartilage. What if they could have been more precise than that five “slice”