My Applied Research Project (ARP) through A.T. Still University’s (ATSU) Doctor of Athletic Training (DAT) program started with something I was seeing every day on the job. Working in professional baseball, I noticed that relief pitchers have a very different life than starters: they pitch more often, with less rest, and are expected to give everything they have even in short outings. And they get hurt a lot. What surprised me was that when I went looking for research on their injury risk, almost nothing focused on them specifically. Most studies either looked only at starting pitchers or lumped everyone together. That felt like a real blind spot, both for the science and for how we actually take care of these athletes. So, the question almost wrote itself: does how much a relief pitcher throws in one season affect their chances of getting injured the next?  

One thing that worked in my favor early on was that the data I needed was publicly available through the official MLB website. Pitching stats, transaction records, and injury placements—it was all there, and I was able to pull together a large dataset covering the 2021-2024 seasons, with 809 pitcher-seasons in total. Compared to a lot of research where just getting access to data is a project in itself, this part went relatively smoothly.  

What turned out to be genuinely hard was everything that came after collecting the data. These records were never designed for research. They are built for running a baseball league, not for an epidemiological study. So, I had to reorganize everything from the ground up: align exposure and outcome windows season by season, manually classify hundreds of injury records, calculate workload metrics, and merge information across multiple sources without introducing errors. It took a lot of time and a lot of careful decision-making.  

Once the data was cleaned up and I started running analyses, things began to take shape. There was an interesting pattern where pitchers who threw the most actually seemed to get hurt less often per appearance, and upper extremity injuries clearly dominated the picture. But I was not confident enough to know whether what I was seeing was genuinely meaningful or just a random finding in how the data came together. That uncertainty was a turning point. That is when Dr. Valier and I reached out to Dr. Bay and Dr. Chandran.  

Dr. Chandran helped us sharpen our focus. He suggested narrowing the population we were studying and gave us practical guidance on the statistical approach, which led us to use a model better suited to the nature of our outcome data. Dr. Bay looked at the dataset from a different angle and helped us work through the analysis in ways we had not fully considered on our own.  

In the end, we did not find a statistically significant link between how much a relief pitcher threw and how likely they were to end up on the Injured List the following season. That might sound like a disappointing result, but it actually tells us something useful. It suggests that simply counting pitches or appearances is not sensitive enough to predict injury on its own. At the same time, we found that nearly half of all pitcher-seasons were followed by an orthopedic injury, which is striking, and that upper extremity injuries on the throwing side accounted for the vast majority of time missed.  

Those findings help paint a clearer picture of what relief pitcher injury really looks like, and they point toward what future research needs to do better: account for the type and intensity of pitches thrown, workload spikes relative to a player's normal baseline, and the large amount of throwing that never shows up in any box score.