Interview with Statistician Jamie E. Collins, PhD

Jamie E. Collins, PhD (JEC)
Senior Statistician at OrACORe
Assistant Professor of Orthopedic Surgery
Harvard Medical School

Emma C. Lape (ECL)


ECL: Can you tell us about your new grant? What is the mechanism and goal of the award? What are the aims of your project?

JC: The Rheumatology Research Foundation has career development awards for junior investigators. The idea is to help them develop as researchers in a supportive environment. It’s great for me because this is my first grant, and hopefully it is a stepping stone to many more. As for the research focus itself, they are looking for innovative methods.

Specifically, my project aims to identify fast progressors in knee osteoarthritis. I’m using data from the Osteoarthritis Initiative (OAI), which is a large cohort study with about 5,000 subjects that was established to study the natural history of osteoarthritis (OA). The OAI collected a lot of data on these people—MRIs, x-rays, blood and urine samples, survey data. I will use statistical techniques to identify fast progressors. To give some context about the research question, there’s been a focus recently on how OA is a heterogenous disease. Not everyone with OA has the same amount or type of structural damage, or the same symptoms. And we want to know whether their disease progresses differently, as well. Can we really describe the trajectory of changes in OA with a one-size-fits-all approach? That’s the traditional approach—to model the trajectory of the group as a whole, and maybe identify baseline factors that make someone progress a little faster or a little slower. But we’re looking for subgroups whose progression is really different.

An extra wrinkle is that it’s hard to define progression, especially when some people have total knee replacement (TKR). They might have TKR precisely because they were progressing quickly, but now you’ve lost their data—there’s no knee left to study. So we have to figure out how to address that missing data, which connects to my thesis work on missing data. Maybe we can view TKR as a surrogate for poor outcomes.


ECL: So some of your previous research on missing data could end up being a crucial part of this project, too. In what other ways does this new grant connect to your previous work? Or does it take you in new directions?

JC: It actually combines two threads of my previous research. The first is my paper on pain trajectories (also in the OAI), and the second is my work on structural progression in those who undergo TKR. I’m now looking at both trajectories in the OAI and the issue of missing data due to TKR. It’s exciting because there is not a ton of work out there using latent growth curve models. So this is an interesting approach.


ECL: To shift gears a little, could you tell me about your experiences at the recent OARSI meeting this past spring? A lot of OrACORe staff were there. What did you present on?

JC: The meeting was a lot of fun. I presented on a project using MeTeOR (Meniscal Tear in Osteoarthritis Research) data. The project deals with OA progression among MeTeOR subjects who had arthroscopic partial meniscectomy (APM) versus those who had only physical therapy. There have been suggestions that APM might speed up OA progression, but there was only observational data previously. So having a randomized trial comparing the surgery to PT was a great opportunity to study progression. Our results suggested that the APM group progressed a bit faster. But what counts as fast or slow progression? We know, for example, that being in the APM group gives you a higher odds of having an additional subregion in your knee with an osteophyte (bone spur). But we will want to know whether that’s associated with outcomes that matter to patients.


ECL: I also heard that you led a young investigators mentoring session. Can you tell me more about mentorship at OARSI and the role of more established investigators?

JC: I’m part of the Young Investigators Subcommittee of OARSI, which has about 10 members. We run several workshops during the conference, once of which this year was on social media. We could tweet and view each other’s tweets on a projector to discuss using social media in research. We also ran a mentoring session with about 175 people attending. We divided up by subject area (biomechanics, etc.) and by career specifics (manuscripts, grant-writing, etc.). Then we matched up mentors and mentees and gave everyone an hour to talk. The mentors are fantastic and invested in working with young investigators. And it was especially nice to see buy-in from top OARSI leadership—Dr. Katz, the current president, and OARSI’s president elect both attended as mentors.

A final observation from the close of the conference: The Year in Review is always the last session of the meeting, and they go through some of the year’s biggest research developments. A ton of work from OrACORe was featured, probably 6 or 7 papers, including OAPol projects and analyses using MeTeOR data. That was very cool to see.