Chances are in the last six months, you have heard or read the term “randomized controlled trial.”
Once an obscure and highly technical term, it first entered common parlance when, in mid-March, President Trump suggested hydroxychloroquine as a possible treatment for COVID-19. Physicians and scientists said that randomized controlled trials involving the drug would be needed before we could know whether it was an effective treatment. More recently, media coverage of a possible coronavirus vaccine has often mentioned the term, noting that vaccines must go through such trials before the Food and Drug Administration will approve them.
Indeed, it has become prominent in the news since the pandemic began. Typing the term into the search engine Google and selecting “News” will yield more than 146,000 results. A search of the Washington Examiner finds 20 articles using the term since March. We often write that it is “considered the gold standard in medical research.”
But what, exactly, is a randomized controlled trial?
In summary, it is a scientific experiment in which patients are randomly assigned to receive either a specific treatment or a placebo. Sometimes, patients are assigned to a third group that receives the best available treatment that has been tested by previous research. That way, the new treatment can be compared to the most effective established treatment.
Ultimately, it will inform researchers about the efficacy and safety of a specific treatment. Researchers will be able to determine whether the treatment is more effective than no treatment or the best established treatment. They will also be able to better understand what side effects the treatment may have.
Assigning patients randomly enables researchers to eliminate certain sources of bias that could otherwise confound the results. Compare randomized controlled trials to another very common type of study known as a “retrospective study.” A retrospective study enables researchers to examine events that have happened in the past. For example, take a recent retrospective study that compared patients who were hospitalized with COVID-19 and treated with hydroxychloroquine to patients who were not treated with it.
The study found a lower mortality rate among those patients who took hydroxychloroquine. But these types of studies have a number of limitations, the most important being bias. Specifically, physicians might be prescribing hydroxychloroquine for those patients most likely to do well with the drug and not prescribing it for those they believe it will not help. That alone may explain why the study found that patients taking hydroxychloroquine did better.
There is also the very real possibility that what actually helped those patients who took hydroxychloroquine was psychosomatic. That is, their symptoms were aggravated by anxiety over having COVID-19, and receiving a treatment reduced that anxiety.
Randomized controlled trials have since been conducted to determine the impact of hydroxychloroquine on patients hospitalized with hydroxychloroquine. With one exception, they have not found any benefit. Whether hydroxychloroquine helps COVID-19 outpatients who tend to have less severe symptoms is an open question. Randomized controlled trials of outpatients are not expected to be finished until later in the year.
The ALIFE study, outlined in an article in the German medical journal Deutsches Ärzteblatt, demonstrates how a randomized controlled trial is conducted. ALIFE, which is short for Anticoagulants for LIving FEtuses, examined the effect that the anticoagulant heparin had on miscarriages. The study compared pregnant women to those who took heparin and aspirin to those who either took only aspirin or were given a placebo.
Much preparation goes into conducting such a trial, and the most important part of the preparation may be determining the “primary endpoint.” The article in Deutsches Ärzteblatt states that a primary endpoint is a “parameter measured or observed that is recorded at a defined time and can be assumed to reflect the effect of a treatment.” In the ALIFE study, the endpoint was live births.
The ALIFE study randomly assigned women who had at least two previous miscarriages and were trying to conceive or were less than six weeks pregnant. Random assignment not only helps reduce bias, but it also reduces “confounding factors,” factors that make it difficult to distinguish what, including the treatment, actually affected the outcome. Such factors include age, weight, income, and education, to name a few. Random assignment ensures that confounding factors are distributed equally among the different groups. This way, any differences in the results can be attributed to the treatment and not the confounding factors. In the end, the ALIFE study found that neither treatment with heparin or aspirin was any better than no treatment for reducing miscarriages.
Like any human practice, randomized controlled studies have their limitations. Two such limitations are time and cost. Randomized controlled trials can take years to conduct. During that time, the general population may be deprived of effective medical treatments. They are also expensive, sometimes costing in the hundreds of millions of dollars. If the amount of time needed for a trial is long enough, it may be too expensive to conduct. For example, studying the loosening of artificial joints can require up to 15 years.
Ethics is another limitation. If during a randomized controlled trial the placebo group is doing obviously worse than the treatment group, it may be unethical to continue. If in the ALIFE women who had been given the placebo had far more miscarriages than those given heparin, ethical considerations may have ended the trial before it could be completed.