Deconstructing Databases II: What Numbers Can Tell Us
"Errors using inadequate data are much less than those using no data at all."|
Charles Babbage (1792-1871)
"May you live in interesting times."|
Data collections are the numerical subtext we require for decision making in the modern world, be they guides for everything from 401 K's and hedge funds to Government budgets, corporate and small business investment, or health care allocations.
Decision making in health care investment and spending involves so many areas that only a partial list can be enumerated here: screening for disease, treatment decisions involving appropriate and inappropriate drugs and surgery for hundreds of illnesses, research, optimal diagnostic strategies, etc. Specific examples follow, but these are merely the tip of the medical iceberg.
Increasing controversy has raged over the value of screening healthy men for prostate cancer by measuring prostate-specific antigen (PSA) in blood. In five large well-controlled studies in Europe and the U.S. reported over two years ago, the PSA test was shown not to save lives. These studies finally resulted in a draft recommendation released last month by the U.S. Preventive Services Task Force that healthy men of all ages without symptoms of prostate cancer should no longer receive a PSA blood test to screen for prostate cancer. The task force members surely braced themselves for some serious criticism from urologists and men as well as their families who fervently believe the test saves lives. Moreover, opponents of healthcare reform will scream about one more attempt to ration medical services.
Few people realize the astounding number of clinical disasters resulting from widespread PSA screening. Positive test have far too often led to treatments that needlessly cause pain and suffering, loss of urinary control, and impotence in hundreds of thousands of healthy men. Between 1986 and 2005 one million men received surgery, radiation therapy or both. The vast majority of these men would never have been treated except for a PSA test result. Among them, 5,000 died soon after surgery, and up to 70,000 suffered serious complications. Up to 300,000 men suffered impotence, incontinence or both. Last year 218,000 men received the diagnosis of prostate cancer predominantly the result of screening. Yet the death rate of this cancer is essentially unchanged from 50 years ago. Autopsy studies show that 33% of men 40 to 60 years of age have microscopic prostate cancers, rising to 75% of men after age 85, yet the disease is rare before age 50, and most deaths occur over age 75.
In other words, thanks to PSA screening we're killing or crippling thousands of healthy men in the belief that we're preventing them from dying of prostate cancer. Professor Richard J. Albin who first described the PSA. antigen 40 years ago, has called its widespread use in prostate screening a "public health disaster."
We spend over $5 billion annually on mammograms to screen for breast cancer. For the 39 million women screened we still don't know what percentage-If Any-of 230,000 women diagnosed with invasive breast cancer –that's 1 in 170 mammograms-will have their lives "saved" by early diagnosis. In 1930 three women in 10,000 died of breast cancer; in 2008 the figure was two in 10,000. This suggests that fewer than 1 in a thousand healthy women screened over a decade will have a cancer found at just the right moment for successful treatment. It remains an article of faith that this reduction in mortality is a result of screening rather than improvement in treatment. A landmark article on the subject was published just two weeks ago. The authors used the National Cancer Institute's software for analyzing Surveillance Epidemiology and End Results (SEER), data covering 26% of the U.S. population, to estimate the 10-year risk of diagnosis and the 20-year risk of death—"...a time horizon long enough to capture the downstream benefits of screening. Using a range of estimates on the ability of screening mammography to reduce breast cancer mortality...," the authors estimated the risk of dying from breast cancer in the presence and absence of mammography in women of various ages (ages 40, 50, 60, and 70 years). Their summary consisted of two sentences: "Most women with screen-detected breast cancer have not had their life saved by screening. They are instead either diagnosed early (with no effect on their mortality) or over diagnosed."
And what about the women who "will undergo a decade's worth of radiation for no medical benefit, or worse yet, undergo unnecessary treatments, that can be harmful to eliminate tumors that would never have killed them."?
Gradually we are beginning to learn about the perils and pitfalls, if not the triumphs of bariatric surgery for the treatment of morbid obesity. There are over 23 million people in the U.S. with a body mass index exceeding 40. In 1995 15,000 American had some type of gastric bypass surgery, now succeeded by adjustable gastric banding done laparascopically. In the last three years, the number of surgeries has climbed to 200,000 and is rising rapidly. There are a bouquet of risks and benefits, all too numerous to detail, associated with this surgery. Significantly, long term follow-ups-over 2-5 years- of large groups of patients are hard to come by. See this comment which criticizes one report claiming a 15 year follow up of 1,400 patients. The authors' claimed 97% follow-up included only the previous TWO years. Results did not specify which patients were included nor are their charts published. I invite my readers to find what I could not, namely, large series demonstrating long term cost vs. benefit in bariatric surgery; in particular, how many patients maintained the claimed average of 100 pounds loss over more than two years.
Barrett's esophagus is the growth of intestinal-type cells above the border of the esophagus and stomach. It is a rare complication of acid reflux disease, "GERD," which is assumed to lead to the development of esophageal cancer. Yet, only 5 percent of patients who had surgery for esophageal adenocarcinoma were known to have Barrett's esophagus before operation, highlighting the fact that current screening techniques are relatively ineffective. Still, The American College of Gastroenterology states that "...patients with chronic symptoms of GERD are those most likely to have Barrett's esophagus and should undergo upper endoscopy." Yet only one-third of esophageal cancer occurs in the lower esophagus where Barrett's is seen. Note there are an estimated 10-20 million chronic GERD sufferers in the U.S. with symptoms three or more times a week! Should all these people go through endoscopic screening for Barrett's every few years?
A recent study published this year in the Journal of the National Cancer Institute which looked at data from 11,000 people gathered over 20 years in Ireland, found the risk of esophageal cancer in people with Barrett's was 0.13% or about 1 in 770 people, not 1 in 200, as previously thought... My own estimate over 5 years ago was 1 in 375 based on known cancer statistics. With the present numbers a physician would have to evaluate almost 80 patients with known Barrett's for 10 years each to have a chance of finding a single cancer.
Upper Endoscopy (EGD)
Upper gastrointestinal endoscopy, a procedure with known complications of 1 in 1,500 and a death rate of 1 in 3,500- 5,000 is performed over 10-12 million times a year in this country on people with upper GI complaints. Another, simpler and much cheaper method, virtually without risk or complications is the use of X-rays with barium to outline the esophagus, stomach and intestine, the esophagogram and upper GI series. This routine study has been virtually abandoned by the medical profession over the past 30 years. The actual number of GI series is no longer included in Government statistics (NCHS and NHDS) for hospitals, yet, the upper GI series gives invaluable information regarding upper GI function, such as swallowing, gastric emptying, stomach irritability, transit time to the small intestine, information which EGD cannot give, since it only visualizes structure, not function. Furthermore, because of the low demand for X-Ray studies of the GI tract, fewer radiologists under 50 have developed the requisite skill and experience to perform them. I have already discussed this problem in a previous newsletter.
Thanks to the lack of follow-up statistics on the results of EGD, let alone the upper GI series, we have no clear understanding why an invasive Diagnostic Procedure with real risks that costs our healthcare system up to $25-$30 billion a year has replaced a generally superior procedure without risks that would cost less than $1-2 billion a year.
"Outcomes Research" is the up and coming mantra, one which I enthusiastically embrace. When we begin systematically to follow up our patients whom we too often over-test, misdiagnose, or wrongly over-treat, we will finally gain an insight into why our costly medical system is rapidly emulating the world financial crisis. It will never be easy to follow sufficient numbers of patients for sufficient time to acquire the statistical databases that are critical in making informed judgments about the results of some of our medical handiwork. Still, it seems to me we are witnessing encouraging signs of a creeping sanity involving our healthcare confreres. We can only hope this awareness will ultimately infect our political leadership.
Martin F. Sturman, MD, FACP
Copyright 2011, Mathemedics, Inc.
EasyDiagnosis is an automated online service that analyses existing medical symptoms and predicts likely causes and conditions. Click here to find out more about this unique service and click here to try one of our modules for free.