The first medical meta-analysis: Almroth Wright’s 1904 analysis of typhoid vaccine evidence

In the late nineteenth century, armies faced a silent enemy far more deadly than any opposing soldier: typhoid fever. This bacterial infection, spread through contaminated food and water, ravaged military camps, killing more British troops in the South African War than enemy fire did. To combat this, a scientist named Almroth Wright developed a vaccine made from killed bacteria, hoping to protect soldiers before they fell ill. However, proving that a new medicine works is rarely simple, especially when the data comes from real-world chaos rather than a controlled laboratory. Scientists often face a mountain of individual reports, each with its own flaws, making it hard to see the true picture. This is where the concept of a meta-analysis enters the story. Think of it as a way to gather many separate, small studies and combine their results into one large, clearer view, allowing researchers to see patterns that might be hidden when looking at just one report at a time. Today, this method is a cornerstone of modern medicine, but its origins are rooted in a fierce debate over who could best interpret the evidence for the typhoid vaccine.

For decades, history books have credited a statistician named Karl Pearson with performing the first medical meta-analysis in 1904. The story goes that Pearson took the available data on the typhoid vaccine, rearranged it, and concluded that the vaccine offered little protection, effectively shutting down the argument for its widespread use. However, a new examination of the original records suggests a different timeline and a different hero. The author of this paper, Vivian McAlister, argues that the true pioneer was actually Almroth Wright himself, the very man Pearson criticized. Before Pearson ever touched the data, Wright had already gathered the scattered reports from hospitals and military camps, checked their reliability, and combined them to reach a powerful conclusion: the vaccine worked.

Wright's work began in the trenches and field hospitals of the late 1890s and early 1900s. He collected reports from soldiers who had volunteered for the injection and compared them with those who had not. He knew the data was imperfect; the groups were not perfectly matched, and the records were sometimes messy. Yet, he believed that by looking at the numbers together, a clear trend would emerge. He organized these findings into tables, calculated the differences in death rates and infection rates, and estimated that the vaccine reduced the risk of dying from typhoid by between 50 and 75 percent. He also found it cut the number of people getting sick by about half and reduced the chance of death for those who did get sick by roughly 52 percent. Wright presented this synthesis in 1904, arguing that the evidence was strong enough to justify giving the vaccine to every soldier.

The controversy began when a military officer, seeking ammunition against Wright's push for universal vaccination, asked Karl Pearson to review the same evidence. Pearson did not look at Wright's full collection. Instead, he selected a smaller subset of the data, rearranged the numbers, and applied a complex statistical method that was unfamiliar to most doctors. Pearson concluded that the link between the vaccine and survival was weak. His report was used to block the adoption of the vaccine, and the resulting public argument between the two men became so heated that it silenced further debate for years. Pearson dismissed Wright's objections as a lack of statistical understanding, while Wright felt Pearson had ignored crucial parts of the evidence.

Decades later, this paper revisits the original data to see who was right. Using modern computer software designed to combine study results, the author re-analyzed the exact same reports Wright had assembled. The results were striking. When the data was combined using today's standard methods, the vaccine still showed a massive benefit. The analysis confirmed that vaccinated soldiers were far less likely to die from typhoid, far less likely to catch the disease, and much more likely to survive if they did get sick. The modern re-analysis of the full dataset produced results that were almost identical to Wright's original estimates. Even when the author looked only at the specific studies that Pearson had chosen to include, the vaccine still appeared to be highly effective, contradicting Pearson's original conclusion that the protection was negligible.

The paper also highlights a crucial lesson about how we trust science. Pearson's work was mathematically sophisticated, but it relied on a narrow slice of the evidence and a method that could obscure the true effect. Wright's approach, while lacking modern statistical weighting, was more honest about the limitations of the data and more comprehensive in its scope. He understood that the vaccine worked because the pattern held true across different groups and different settings, even if the numbers varied slightly from place to place. The author notes that the high level of agreement between Wright's 1904 synthesis and the modern re-analysis suggests that Wright's intuition was sound. He had successfully performed the first true meta-analysis, assembling, appraising, and synthesizing the evidence long before the term existed.

The consequences of this historical dispute were real and tragic. Because the debate was stalled by Pearson's report, the British Army did not make the vaccine mandatory at the start of the First World War. As a result, typhoid returned with a vengeance, killing hundreds of soldiers who could have been saved. In contrast, Canadian troops, who were inoculated based on different evidence and leadership, suffered far fewer cases and deaths. This paper does not just correct a name in a history book; it restores the credit to the man who first showed us how to look at many small pieces of evidence to find a life-saving truth. It reminds us that the goal of statistics is not to create complex barriers to understanding, but to clarify the evidence so that life-saving treatments can be recognized and used. Wright's 1904 report stands as a stronger landmark in the history of medical science than the re-analysis that followed it, proving that sometimes the most powerful tool is simply the willingness to look at all the facts together.