LakeMountD
Doctor Science
- Awards
- 1
I am currently reading a book called Overdo$ed America, which I recommend to everyone who truly cares about their health, sees a physician frequently and takes prescription medications, and to people who enjoy reading PubMed and other internet/paper sources.
I am quoting a specific section of this book on pages 13-16 and am typing this up because I felt it was very significant and disturbing evidence of the misleading details added into pharmaceutical studies.
"In August 2000, I was reading the New England Journal of Medicine over a typical lunch in my office when I noticed an article titled "Pravastatin Therapy and the Risk of Stroke." This caught my attention because stroke is a fairly common problem among my elderly patients, and though most strokes are minor, some are tragic, leaving patients trapped in a permanent state of severe disability.......(skipping story about one of his stroke victims named Mrs. Rose)........
I was certainly interested in decreasing other patients' risk of suffering the same fate. But as I started to read the abstract, I became a bit suspicious when I realized that the title itself was misleading. The study did indeed examine the effect of pravastatin, a cholesterol-lowering statin drug, in decreasing the risk of a stroke, but it only included people who had already suffered a heart attack or unstable angina. The reduction in the risk of stroke reported in the study did not apply to (most of my previous patients) or to any other patients who did not already have heart disease, as might be reasonable assumed from the article's title. Still, the abstract went on to report a statistically significant reduction in the risk of strokes among post-heart attack patients given Pravaschol compared with those given a placebo, and concluded that "pravastatin has a moderate effect in reducing the risk of stroke."
In the few minutes I had left before it was time to see my afternoon patients, I noticed that when the pre-study differences in conditions that predisposed a person to stroke (such as diabetes, high blood pressure, smoking, atrial fibrillation, and previous strokes) were taken into account, the reduction of stroke in the people who took Pravaschol was not even nlarge enough to be statistically significant. This was the first time I had seen an article published in a major journal that reported results that did not reach statistical significance. I paused for a moment to register this fact and then read on.
The NEJM article reported an impressive sounding 19 percent reduction in the risk of stroke in people who had taken Pravachol compared with those who had been given the placebo. This came down to 16 percent when the results were corrected for pre-study differences in factors that predisposed stroke. This way of expressing the effect of taking Pravachol is called the "relative risk reduction". But relative risk tells only part of the story and often conveys an exaggerated impression of the benefit of the new drug or therapy. The far more important result is the amount of disease that is prevented by a given drug or therapy, called the "absolute risk reduction."
The research skills I had learned as a Robert Wood Johnson Fellow served me well in critically reading articles in medical journals. Occasionally I would take out a pencil and calculator to do the arithmetic when the findings were confusing or their importance was not clear. And that is what I did with this study.
Over the course of this six-year study, 4.5 percent of the patients who took the placebo had strokes, compared with 3.7 percent of the patients who took Pravachol. The ratio between 4.5 and 3.7 percent provides the relative risk reduction of 19 percent. however, the actual number of strokes prevented by taking Pravachol, or the absolute risk reduction, is calculated by subtracting 3.7 from 4.5 percent. So over the six years of the study there were 0.8 percent fewer strokes among the people who took Pravachol. In other words, if the 1000 post heart-attack patients took Pravachol for a year, there would be about one less stroke. This didn't sound to me like a major finding. I called the local pharmacy to find out the cost of Pravachol 40mg per day and calculated that the cost of each stroke thus prevented was $1.2 million, not even including the cost of the extra blood tests and doctor visits to monitor for potentially dangerous side effects of medication.
My curiosity about the rest of this article now piqued, but it was time to start the afternoon clinic session. I took the article home to continue the time-consuming process of careful analysis. One of the benefits of a faculty appointment at Harvard Medical School is access to the school's digital library, which provides access to virtually all the important medical journals. I wanted to dig around a little to see if I could find more information about statins and strokes.
The better I understood the details of this article, the more misleading its conclusion appeared to be. The biggest problem was the disparity between the people included in the study and the people most at risk of stroke. The people in the study averaged 62 years of age, but the age at which most strokes occur in the general population is much older: half of the men who have strokes are 71 years or older, and half of the women who have strokes, like Mrs. Rose, are 79 or older. This is important because the patients in the study age 70 and older who had been treated with Pravachol had 21 percent MORE strokes than the patients given the placebo.
I kept going.
Eighty-three percent of the people included in the study were men, but three out of five stroke victims in the general population are women. The fact that fewer than one out of eight people in the Pravachol study were women turns out to be important because the women in the study who were given Pravaschol experienced 26 percent more strokes than the women who were given the placebo.
And five out of six patients in the study were taking aspirin routinely to decrease their risk of having more heart trouble. But in the general population, the vast majority of people, like Mrs. Rose, do not take aspirin routinely. This is important because among the people in the study who were not taking aspirin, those taking Pravachol had 20 percent MORE strokes than those taking placebos.
My patient, Mrs. Rose, was a typical stroke victim: female, in her early 80's, and not taking aspirin. According to the results of the study, her risk of stroke would have been increased, not decreased, by treating her with Pravaschol."
I am quoting a specific section of this book on pages 13-16 and am typing this up because I felt it was very significant and disturbing evidence of the misleading details added into pharmaceutical studies.
"In August 2000, I was reading the New England Journal of Medicine over a typical lunch in my office when I noticed an article titled "Pravastatin Therapy and the Risk of Stroke." This caught my attention because stroke is a fairly common problem among my elderly patients, and though most strokes are minor, some are tragic, leaving patients trapped in a permanent state of severe disability.......(skipping story about one of his stroke victims named Mrs. Rose)........
I was certainly interested in decreasing other patients' risk of suffering the same fate. But as I started to read the abstract, I became a bit suspicious when I realized that the title itself was misleading. The study did indeed examine the effect of pravastatin, a cholesterol-lowering statin drug, in decreasing the risk of a stroke, but it only included people who had already suffered a heart attack or unstable angina. The reduction in the risk of stroke reported in the study did not apply to (most of my previous patients) or to any other patients who did not already have heart disease, as might be reasonable assumed from the article's title. Still, the abstract went on to report a statistically significant reduction in the risk of strokes among post-heart attack patients given Pravaschol compared with those given a placebo, and concluded that "pravastatin has a moderate effect in reducing the risk of stroke."
In the few minutes I had left before it was time to see my afternoon patients, I noticed that when the pre-study differences in conditions that predisposed a person to stroke (such as diabetes, high blood pressure, smoking, atrial fibrillation, and previous strokes) were taken into account, the reduction of stroke in the people who took Pravaschol was not even nlarge enough to be statistically significant. This was the first time I had seen an article published in a major journal that reported results that did not reach statistical significance. I paused for a moment to register this fact and then read on.
The NEJM article reported an impressive sounding 19 percent reduction in the risk of stroke in people who had taken Pravachol compared with those who had been given the placebo. This came down to 16 percent when the results were corrected for pre-study differences in factors that predisposed stroke. This way of expressing the effect of taking Pravachol is called the "relative risk reduction". But relative risk tells only part of the story and often conveys an exaggerated impression of the benefit of the new drug or therapy. The far more important result is the amount of disease that is prevented by a given drug or therapy, called the "absolute risk reduction."
The research skills I had learned as a Robert Wood Johnson Fellow served me well in critically reading articles in medical journals. Occasionally I would take out a pencil and calculator to do the arithmetic when the findings were confusing or their importance was not clear. And that is what I did with this study.
Over the course of this six-year study, 4.5 percent of the patients who took the placebo had strokes, compared with 3.7 percent of the patients who took Pravachol. The ratio between 4.5 and 3.7 percent provides the relative risk reduction of 19 percent. however, the actual number of strokes prevented by taking Pravachol, or the absolute risk reduction, is calculated by subtracting 3.7 from 4.5 percent. So over the six years of the study there were 0.8 percent fewer strokes among the people who took Pravachol. In other words, if the 1000 post heart-attack patients took Pravachol for a year, there would be about one less stroke. This didn't sound to me like a major finding. I called the local pharmacy to find out the cost of Pravachol 40mg per day and calculated that the cost of each stroke thus prevented was $1.2 million, not even including the cost of the extra blood tests and doctor visits to monitor for potentially dangerous side effects of medication.
My curiosity about the rest of this article now piqued, but it was time to start the afternoon clinic session. I took the article home to continue the time-consuming process of careful analysis. One of the benefits of a faculty appointment at Harvard Medical School is access to the school's digital library, which provides access to virtually all the important medical journals. I wanted to dig around a little to see if I could find more information about statins and strokes.
The better I understood the details of this article, the more misleading its conclusion appeared to be. The biggest problem was the disparity between the people included in the study and the people most at risk of stroke. The people in the study averaged 62 years of age, but the age at which most strokes occur in the general population is much older: half of the men who have strokes are 71 years or older, and half of the women who have strokes, like Mrs. Rose, are 79 or older. This is important because the patients in the study age 70 and older who had been treated with Pravachol had 21 percent MORE strokes than the patients given the placebo.
I kept going.
Eighty-three percent of the people included in the study were men, but three out of five stroke victims in the general population are women. The fact that fewer than one out of eight people in the Pravachol study were women turns out to be important because the women in the study who were given Pravaschol experienced 26 percent more strokes than the women who were given the placebo.
And five out of six patients in the study were taking aspirin routinely to decrease their risk of having more heart trouble. But in the general population, the vast majority of people, like Mrs. Rose, do not take aspirin routinely. This is important because among the people in the study who were not taking aspirin, those taking Pravachol had 20 percent MORE strokes than those taking placebos.
My patient, Mrs. Rose, was a typical stroke victim: female, in her early 80's, and not taking aspirin. According to the results of the study, her risk of stroke would have been increased, not decreased, by treating her with Pravaschol."
