"Their cases were as similar as I could have them": the origins of the modern clinical trial
In 1747, British naval surgeon James Lind carried out an experiment.1,2 He selected 12 sailors with scurvy, an often-fatal disease that was common at sea, and divided them into six pairs. Each pair received a different daily supplement to their usual diet; for example, one received cider, another seawater. Two fortunate sailors were each given two oranges and a lemon. Their health improved markedly, bolstering Lind’s theory that citrus juice was an efficacious treatment for scurvy.
Lind's trial was not without flaws: the sample size was small and the trial short-lived (fruit ran out after only 6 days). But by using what we would now call ‘control groups’, Lind was an important pioneer. Of the participants, he wrote that: “Their cases were as similar as I could have them. They all in general had putrid gums, the spots and lassitude, with weakness of the knees. They lay together in one place...and had one diet common to all.” He sought, in short, to conduct a ‘fair test’ – changing just one variable at a time, while keeping other variables constant. This principle remains at the heart of clinical trials today, and when critically analysing trials, we need to be alert to the possibility that more than one variable may have been changed.
"We wanted to muddle people up": double-blinded, randomized, multicentre trials
Jump forward another two centuries and we reach another major development in the conduct of clinical trials. In 1943–4, a trial orchestrated by the British Medical Research Council (MRC) investigated patulin as treatment for the common cold.2,3 Neither doctor nor patient knew what treatment was being administered, therefore making it the first double-blind trial. “We wanted to muddle people up,” one of the investigators later recalled.4 This was also the first multicentre trial, taking place in 11 factories and three Post Offices across the UK, under the supervision of a single researcher. A subsequent MRC trial in 1947–8, this time evaluating streptomycin in patients with tuberculosis, took another methodological step forward: this trial was the first to randomize patients, with the participants assigned different treatments according to a sequence of randomly ordered envelopes.2,5
These trials helped to establish what remains the gold standard for medical research: the double-blinded, multicentre, randomized controlled trial. The best modern trials adopt this approach, even if they look very different in their size and sophisticated methodologies. Today’s practice-changing cardiovascular trials often include tens of thousands of patients from many hundreds of centres across dozens of countries worldwide. State-of-the-art randomization involves central, independent randomization at a clinical trial coordinating centre, using complex computer-generated methods.
Powering and follow-up
Robust trials have many additional characteristics. Studies need to be adequately powered and contain enough participants for any differences between the sample groups to have a statistical significance and not be attributable merely to chance. For example, to adequately study hard cardiovascular endpoints, sample sizes as large as 15,000–20,000 patients may be required, as reflected in recent cardiovascular outcomes trials. Trials also need to keep track of outcomes for all participants. Missing data from the follow-up period can seriously confound a trial’s results.
Even today, all evidence is not equal
We’ve come a long way since 1747. We understand better the many varied and complex ways in which bias and other confounding factors can creep into clinical investigations, and we have learned much about effective study design and statistical analysis. However, not all contemporary trials meet the highest standards. Trials may have small sample sizes because they are investigating very rare diseases, or it may not be possible to blind a group to some treatments, such as surgery. Funding limitations also place constraints on how trials can be conducted.
No trial is perfect, and data from clinical trials which do not meet the highest contemporary standards should not necessarily be disregarded out of hand, but there is a need to tread carefully when evaluating their results. All trials are not equal, and we should not assign equal weight to their findings. If trials generate unexpected results, we may still regard these data to be hypothesis-generating, but the outcomes should always consider the conduct of the trial and be treated with an appropriate degree of caution.
Trials in context: the importance of the whole body of evidence
Individual trials do not stand alone, but take their place alongside other evidence. This is reflected in many clinical guidelines, where a single randomized controlled trial, however methodologically sound, is not enough to grade a recommendation at the highest evidence level.6 Long-term clinical trial programmes that include multiple, robustly designed trials contribute to establishing a consistent body of evidence. AstraZeneca’s long-term programmes, comprising clinical trials conducted according to best practice, highlight our commitment to consistently delivering high-quality evidence.
James Lind’s findings were also just one part of a body of evidence. He was not the first person to observe the benefits of citrus fruits. Nor was his famous 1747 study the last that he conducted. It is, in fact, instructive to remember that his initial confidence in the efficacy of citrus juice may have been shaken by his later practice of boiling it, unknowingly destroying the active ingredient, vitamin C.7 Anomalous results such as these risk steering us off-course if we fail to carefully evaluate the study findings.
1. Lind J. A treatise of the scurvy. Edinburgh: A. Kincaid and A. Donaldson 1753.
2. Bhatt A. Evolution of clinical research: a history before and beyond James Lind. Perspect Clin Res 1:6–10.
3. Medical Research Council. Clinical trial of patulin in the common cold. Lancet 1944;244:373–5.
4. Medical Research Council. Available at: https://mrc.ukri.org/news/blog/the-road-to-randomisation-patulin-and-the-common-cold/ (accessed November 2019).
5. MRC Streptomycin in Tuberculosis Trials Committee. Streptomycin treatment of pulmonary tuberculosis. BMJ 1948;2:769–83.
6. Cosentino F, Grant PJ, Aboyans et al. 2019 ESC guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD. Eur Heart J 2019 [Epub ahead of print].
7. Hughes RE. James Lind and the cure of scurvy: an experimental approach. Med Hist 1975;19:342–51.
Veeva ID: Z4-21468
Date of Preparation: November 2019
Date of Expiry: November 2020