Advantages And Disadvantages Of Observational And Experimental Studies

In this chapter, we've examined both nonexperimental and experimental studies to test various purported health benefits of vitamin C and other antioxidants. Although we've seen that experimental studies have pitfalls, I've emphasized their advantages, which make them the gold standard of medical research for good reason. However, we shouldn't simply dismiss the role of prospective studies, case-control comparisons, and other kinds of nonexperimental designs in medical research. These kinds of studies can provide essential information when experimental studies are impractical or unethical. For example, no one seriously doubts that smok ing has multiple bad consequences for human health, yet much of the evidence for this belief comes from large-scale, long-term comparisons of smokers with matched groups of nonsmokers. In this case, matching subjects for as many potential confounding variables as possible has been a reasonable subst itute for randomly assigning "volunteers" to smoke or refrain from smoking for several years. Indictment of smok ing as the most significant controllable health risk has been reinforced by the fact that many different studies of diverse groups of people have all pointed to the same conclusion. Furthermore, experimental and non-experimental studies often provide complementary kinds of evidence to answer questions not only in medicine but in other biological sciences as well. I'll illustrate the process of marshaling multiple types of evidence to answer biological questions in Chapter 4, using two recent ecological examples.

Unfortunately, the culture of science sometimes leads to excessive valuation of experimental methods. There is a fascinating example of this in the case of a disease of newborns called persistent pulmonary hypertension. With conventional medical treatment (CT ), only 20% of patients survived this disease. In the late 1970s, a group at the University of Michigan (Bartlett et al. 2000) began testing a new treatment called extracorporeal membraneous oxygenation (ECMO) in which the patient's blood is passed through a heart-lung machine outside the body for several days. They were able to increase survival rate to 80%, but weren't completely confident of this seemingly dramatic success. Perhaps their pat ients differed in some unk nown way from the earlier group of pat ients, and this difference rather than their new treatment accounted for the difference in survival. Therefore, they designed a small randomized trial to compare ECMO to CT. Because the new treatment seemed so successful and they wanted to make it available to as many patients as possible, they used a randomization method called "randomized play the winner." The treatment for the first baby suffering from persistent pulmonary hypertension was to be determined completely randomly; that is, there would be 50% probability that it would receive CT and 50% probability that it would receive the new therapy. If the treatment was successful, its probability would be increased for the next baby. As it happened, the first baby was assigned the new ECMO treatment and survived whereas the second baby was assigned CT and died. By this time, according to the prespecified protocol for randomization, the probability of assigning a baby to the new treatment was 75% and to the old treatment was 25%. The next 10 babies received the new treatment and all survived; the results of the study were reported in 1985 as 11 survivors of the 11 babies treated by ECMO and 1 death of the one baby treated by CT.

This study was severely criticized by other researchers because only one patient received CT, so a follow-up st udy with a standard randomization scheme similar to that used in the vitamin C st udies was designed and carried out (Ware 1989; Royall 1991). There were nine patients in the ECMO treatment group, all of whom survived, and 10 patients in the control group that received CT, six of whom survived. This example raises a wrenching ethical dilemma. Do you think the deaths of four infants who received CT in this final randomized trial were "necessar y" in some sense to demonstrate conclusively the value of the new treatment? Or do you think the medical community should have been satisfied with the initial comparative data in which the control group was not randomly selected but was simply made up of patients born before the new treatment was developed? How would your answers differ if the benefits of a new treatment weren't so dramatic? Seemingly dry and technical aspects of scientific methodology sometimes have profound practical and ethical implications.

Quit Smoking Today

Quit Smoking Today

Quit smoking for good! Stop your bad habits for good, learn to cope with the addiction of cigarettes and how to curb cravings and begin a new life. You will never again have to leave a meeting and find a place outside to smoke, losing valuable time. This is the key to your freedom from addiction, take the first step!

Get My Free Ebook

Post a comment