Here are the hard facts. There are a lot of potential questions to ask. Studies can only take on one, maybe two questions at any given time. They have to be big because it takes lots of healthy people being tested to find a few cancers and, in turn, really make a difference in preventing an even fewer number of cancer deaths. And the studies have important details to attend to like how to determine why people died, and then deciding if that is all that matters.
Testing a cancer screening test actually involves studying the combined effect of early detection and early treatment. From the individual's perspective now you ask ,, "Is a strategy of looking for and treating cancers before they become noticeable better than simply treating cancers at the point they become noticeable?"
This question implies that two conditions must be met for screening to work: (1) the test must find cancers that will ultimately warrant treatment (i.e., not pseudodisease) and (2) treatment at the time of detection must be more effective than treatment at the time symptoms appear. One of my closest colleagues makes the analogy that an early detection strategy is like a machine with a lot of dials. Each dial has multiple possible settings: there's a dial to set the age to start screening, there's another for when to stop. There's a dial for how often to screen.
There's a set of dials for technical controls (e.g., how many views? at what magnification?). Another set specifies how to interpret the test (e.g., what constitutes abnormal? when should a test be repeated? when should a different test be ordered?). Finally, there's a set of dials for treatment.
You can probably already guess that the "right" position for these dials is not up to the max. If you set the starting age dial too young, mostly what you do is worry lots of young people with no possible benefit. If you set the stopping age too old, mostly what you do is detect a lot of pseudodisease and do lots of unnecessary treatment in the elderly. If you test too frequently, you drive people including doctors crazy
. If one makes the test too sensitive, you get a lot of noise: false positives and pseudodisease. And if you treat too aggressively, the treatment becomes worse than the disease. If you turn all the dials up to the max, every person becomes a patient some literally, all emotionally.
You end up doing more damage than good. So the "right" position is somewhere in the middle just like setting brightness, hue, color, and contrast on your TV. The difference is your TV gives instant feedback on your settings. You will soon fine-tune the settings to improve the picture. Choosing the best settings for a screening test, in contrast, takes time, money, and a lot of volunteers.
How many healthy people need to be entered into a randomized trial to discover whether an early detection strategy (with a particular set of settings) reduces death? The short answer is, a lot. The longer answer is, it depends. It basically depends on two things: how likely a cancer death is and how much of a difference you need to detect.
To obtain a feel for this, let's start with a relatively easy case (from the researchers' perspective): screening cigarette smokers for lung cancer. It's an easy case because death is common, and when death is common it is easier determine whether intervention reduces the chance of death. Cigarette smokers face a relatively high risk of cancer death: among 1,000 middle-aged smokers, about 40 will die of lung cancer within the next 10 years. If you wanted to see whether lung cancer screening early detection and early treatment can cut that death rate in half (that is, from 40 to 20 per 1,000), the statisticians would tell you that you need to enroll 3,200 smokers in your study.
Now, cutting a death rate in half is no picnic. It's more like climbing K2. No cancer test has ever achieved such a feat. But if you wish to make sure you can detect a smaller effect, you have to study even more people.
And the number needed rises exponentially. To see a 25% reduction in lung cancer deaths (from 40 to 30 per 1,000), for example, requires over 14,000 subjects; to see a 10% reduction in lung cancer deaths (from 40 to 36 per 1,000) requires almost 100,000. 1 What if we move to a common cancer that affects smokers and nonsmokers alike? Take colon cancer: among 1,000 middle-aged women and men, about 5 will die of colon cancer in the next 10 years. Now assume you want to know whether screening using, say, virtual colonoscopy is effective.
Astudy designed to detect a 50% reduction in colon cancer mortality requires studying 25,000 people; detection of a 25% reduction requires studying 120,000 people; and a 10% reduction, 800,000 people. The numbers are a bit lower for breast cancer and quite a bit higher for the rarer cancers, such as ovarian cancer, pancreatic cancer, and lung cancer in nonsmokers.
Our website is not responsible for the information contained by this article. Articleinput.com is a free articles resource thus practically any visitor can submit an article. However if you notice any copyrighted material, please contact us and we will remove the article(s) in discussion right away.
Note: This article was sent to us by: Fiona Landman at 08152010
1. The Facts About Pleural Mesothelioma
© 2009 ArticleInput.com.