Randomised controlled trials, or RCTs, are considered the ‘gold standard’ in evidence-based medicine for evaluating the effects of medical interventions.
For medical writers, RCTs are a valuable source of evidence.
Their powerful experimental design can provide important medical information on the causality of medical interventions.
This second article of our series on clinical evidence will look at what makes their design so powerful and what types of RCTs are valuable for medical writers.
While RCTs provide important information for medical writers, we will also take a look at when RCTs may also not always be the most practical source of evidence for your writing.
What are randomised controlled trials?
An RCT is a comparative, controlled experiment designed for finding useful information on efficacy (or costs) of one or two medical treatments in different patient groups.
A simple way to understand the clinical design of RCTs is by looking at the descriptive three words:
- Randomised allocation of participants to treatment groups or study arms so chance alone decides who receives treatments. This removes selection bias keeping study groups similar and making the results more reliable.
- Controlled experimental environment that compares whether an intervention works to a control condition (either placebo or an alternative comparable intervention)
- Trial design ensures intervention groups are treated the same. The analysis measures the difference (if any) of outcomes between groups
Features of a high-quality RCT
When you’re selecting RCTs from a database search to use in your writing, it helps to be reminded what to look for.
RCTs are considered to have the best study design to reduce bias.
High-quality RCTs are conducted under tightly designed and conducted centres that show that the experimental treatment caused an outcome.
Features of a well-designed RCT that minimizes bias include:
- Participants are from a large sample size and are subjected to inclusion/exclusion criteria
- Double blinded trial design
- Outcomes that are easily transferred to a similar patient demographic
- Statistically significant findings
- A control group and no differences in procedures between treatment groups so results are valid
- Planning for assessment and data collection
- An intention to treat analysis
How trial designs can influence bias
There are various types of RCT design types that help to reduce bias.
It can help to be familiar with the different designs and understand which ones offer the best evidence.
- Blinded trials can be double, which is where both participants and investigators are unaware of which treatment is given; single (only participants unaware of their treatment); or unblinded
- Factorial trials randomly assign groups to receive a particular combination of interventions
- Clustered trials randomly select participants from pre-existing ‘groups’, such as patients with type 1 diabetes
- Cross-over studies mean over the length of the trial participants receive interventions in a random sequence
- Fixed-allocation trials assign each participant with an equal probability to one treatment group over the course of the study
- Parallel-group designs are the most common where each participant in one randomly assigned group receives or does not receive the intervention
When an RCT isn’t so valuable
Even though RCTs are considered the gold standard of clinical trials, it’s important you know when not to source RCTs for your writing.
Randomisation may not be feasible, for example, if you’re writing about rare conditions that have small population sizes or if you’re after observational evidence taken over long periods of time (longitudinal studies).
In 2003 The BMJ published a great article that poked fun at why a RCT might not always be practical for exposure to an intervention at random.
The article demonstrated that observational studies have a place by looking at the usefulness of parachutes during free fall by measuring death or major trauma.
Imagine being a participant randomly allocated to the group that free fall without a parachute?
Unsurprisingly, the article couldn’t find evidence of any RCTs conducted to investigate.
Did you know?
Descriptions of the first instance of random allocation of patients to an experimental controlled condition dates back to 1747.
The study looked at treatment for 12 sailors who had scurvy. Two sailors, who randomly were assigned a treatment of lemons and oranges, recovered the quickest; suggesting a beneficial effect of citrus for scurvy.
Although today the number of participants who demonstrate a change in treatment is usually more statistically significant, initial randomisation studies have helped pave the way for the their impact on modern medicine.
Do you have any questions about randomised controlled trials? Let us know in the comments.