Evolution of Medical Statistics

In 1537, during the siege of Turin, a French barber-surgeon, Ambroise Paré, ran out of a boiling oil solution that was used to detoxify wounds caused by gunpowder. He improvised a balm out of egg yolks, rose oil and turpentine, only to discover that by morning,
the patients were able to sleep and in less pain than those with the oil treatment. From this moment on-wards, Paré decided to only use treatments that he deemed to be effective.
Cracked Egg

Aboard the HMS Salisbury in 1747, a pioneer of naval hygiene James Lind carried out the first recorded clinical trial the cause of scurvy, the biggest killer of the British Navy. He selected 12 men with the disease, divided them into 6 pairs and in each pair, added something different to their diets,

including seawater, vinegar, cider, a mixture of garlic mustard and horseradish, and finally two oranges and lemons. The results were conclusive, those fed the citrus fruits experienced a remarkable recovery. Whilst citrus fruits were known to have many medicinal benefits, it was by conducting this trial that Lind was able to prove it was the superior treatment. Despite his efforts, it took 40 years for lemon juice to become a regular on the menu in the Royal Navy.


The Placebo Effect 1785 – 1800s

  • A placebo is a treatment with no intended therapeutic value, such as a sugar pill, that is given to a control group in a clinical trial to allow statisticians to draw comparisons with a new treatment.
  • Today, placebos take the form of sugar pills, but back in the 18th Century physicians administered any sort of medicine deemed to not have much of an effect, for example a mild ointment.
  • John Haygarth was the first to demonstrate the placebo effect back in 1799 when he compared the expensive metal ‘Perkins tractors’, two small metal rods with pointy ends, to wooden ones. The metal tractors worked by applying the points on the aching body part, supposedly relieving the patient’s pain. He compared the results of both and found that they produced similar results, thus demonstrating a placebo effect.

Organised Doctor

Florence Nightingale was a statistician, social reformer and a pioneer in evidence-based nursing. During the 1850s, Florence was a nurse treating wounded soldiers in the Crimean War. During this time, she collected large amounts of data on the soldiers and the conditions of the hospital. The data allowed Florence, along with statistician William Parr, to show statistical evidence that the living conditions led to endemic diseases and were the primary cause of high mortality rates. As a result of Florence’s work, the department of Army Medical Statistics was set up and her legacy of evidence-based medicine led to a decline in preventable hospital deaths and greater awareness of the importance of statistics in medical research and beyond.

In the 1870s and 1880s, Polymath Sir Francis Galton is credited with the first use of  regression to describe biological phenomenon, which he discovered empirically through extensive studies in genetic heritability. His work on regression and correlation is still utilised in medical statistics today.


In 1943, the UK Medical Research Council (MRC) carried out the first double blind controlled trial for Patulin, a drug treating the common cold. This is a trial design where both doctor and patient are blinded to the whether the patient is taking the placebo or the new active treatment. The treatment was assigned by an alternation process which the statisticians deemed adequate, however the results of the trial were disappointing. It was not long after this that we see more sophisticated randomisation methods for treatment allocation being introduced.

In 1946, the MRC conducted the first Randomized Controlled Trial, testing the drug Streptomycin. The statisticians applied systematic enrolment criteria and data collection, as well as random sampling numbers for randomisation and a strictly regulated double blind design in the trial which has since after been regarded as ground breaking.

Doctor Check up


The National Health Interview Survey in the US was introduced in 1956. This was an effort by the Government to obtain comprehensive statistics on the health of the public through the use of a continuing survey which asked Americans about disease and disability. When a flu epidemic hit America in 1957, the NHIS proved an invaluable tool as it provided weekly estimates of the number of infected people. The NHIS was considered a unique survey design as it was repeated annually and the questions remained the same, allowing statisticians to compare results between years. Questionnaires are still a major part of data collected during clinical trials and are used in demonstrating the effectiveness of a new drug or treatment. 

In 1961 it was found that the drug Thalidomide, which was originally marketed as a sedative, was having tragic consequences for expectant mothers who were taking the drug to prevent morning sickness. Around ten thousand babies were born with limb malformations which was shown to be a consequence of the drug. A positive outcome of the tragedy was the introduction of much tougher drug regulations for more rigorous testing, including for teratogenic effects (harmful effects on a foetus or embryo).

Today we are still witnessing the evolution of medical statistics. New trial designs and innovative statistical methodology developed by statisticians are allowing the most effective drugs to be given to patients more quickly and for a lower cost.

Doctors Laptop

Latest News

Show all news

EventsFuture Events

  • Webinar: MCP-Mod – Theory, Implementation and Extensions - Dates: 08 – 08 May, 2019

    MCP-Mod (Multiple Comparisons & Modelling) is a popular statistical methodology for model-based design and analysis of dose finding studies. This webinar will describe the theory behind MCP-Mod (plus extensions), and how to implement it within available software. Pantelis Vlachos (Cytel) will provide a brief introduction to the methodology and illustrate the MCP-MoD capabilities in EAST 6.5. Saswati Saha (University of Brehem) will discuss new variations and alternatives to MCP-Mod and show how to implement them in R. Neal Thomas (Pfizer) will present further technical details of MCP-Mod by evaluating the method using results from least squares linear model theory.
  • PSI Toxicology SIG Workshop 2019 - Dates: 02 – 03 Apr, 2019

    This 1.5-day workshop will involve approximately 20 statisticians, focusing on discussions around “best practice” in the statistical analysis of various data types.​