A randomized trial means people are randomly assigned. People participating in the study are randomly assigned to be in the treatment group, or the no-treatment group. Random assignment allows researchers to conclude that any effect or difference between the two groups is due to the treatment itself, and not due to some other reason. For example, let's suppose the researcher did not randomly assign people to one group or the other. Instead, the researcher decided who would be placed in treatment and the no-treatment groups. We might reasonably conclude that the researcher influenced the outcome of the research. This is called bias.
Random assignment is very important because some types of bias may not be obvious or intentional. Let's say the researcher places the first 100 people who show up for the study into the treatment group. Then, the researcher places the second 100 into the non-treatment group. Well that seems sensible enough. What could be wrong with that? It could be that the people who arrived first were highly motivated. Maybe they are eager for treatment, and ready for change. Likewise, perhaps the ones who arrived later were simply interested in the $20 offered to participants. If the first group got well and the second group did not, it might have nothing to do with the treatment. Instead, it may have to do with the fact the researcher placed all the motivated, first-to-arrive folks in the treatment group. The unmotivated, late-comers were placed into the no-treatment group. Incidentally, individual motivation powerfully influences treatment effectiveness. In another section, we discuss the powerful effect of motivation.
While RCTs are considered the "gold standard" for establishing an evidence-based practice, they are not without some problems. Most RCTs are funded by someone other than the researcher. In the United States, the federal government is the largest funder of addiction research. Research is very expensive. So, funding is typically discontinued after the study participants have been followed six months post-treatment. Consequently, although a treatment may be effective, we know little about whether the results hold up beyond six months. These treatments can be listed in NREPP.
Even if we measured results over longer periods, there are still some problems. There might be considerable disagreement about what would be a good measure of effectiveness across time. This is called the "outcome measure." Furthermore, when long-term studies are conducted the studies typically use an outcome measure of quantity and frequency of substance use. There is nothing particularly special about quantity and frequency of use. These are just measures of substance use that are easy to count and measure. Naturally, researchers like things that can be easily measured. But this provides us limited information. Sometimes quality is more meaningful than quantity. Unfortunately, qualitative effects are more much more difficult to measure. Here are some examples of qualitative measures: Did someone's productivity increase? Are they more involved with their community? How has their social network changed? Has their enjoyment of life increased? Are they more satisfied with their life?
Another limitation of RCTs is that it is difficult to compare treatment studies against each other. Each study is free to choose its own methods and outcome measures. For instance, suppose one study uses abstinence as an outcome measure, but another study uses gainful employment. We could not compare these two studies and decide which is "best" or most effective. All we could say is one treatment was good at helping people to achieve abstinence. The other treatment was good at helping people become responsible wage earners. Researchers are aware of these problems and make an effort to use similar methods and measures in their studies. However, as the field progresses and new methods and measures become available, it is also valuable to use them. Consequently, it often takes many studies before we can firmly establish conclusions.