Missing pieces: Known unknowns in epidemiologic methods

When I was a young Assistant Professor, hoping to spend much of my career as a methodologist, I started a project I called “missing pieces.” I set up a bulletin board on the wall by my desk at home, and on it were post-it notes, each with some aspect of epidemiologic methods that I did not yet understand and therefore would be on my to-do list of things to be studied. These were pieces of knowledge that were missing from what I hoped would eventually be a fairly complete understanding of the epidemiologic methods puzzle, especially methods for answering causal questions. Over the years as I moved to different places in the Twin Cities, the bulletin board moved with me, some post-it notes being removed, and others being added.

I slowly began to realize that there were some aspects of epidemiologic methods that none of us understood, nuts that none of us had yet cracked. I put their post-it notes in the “known unknown” portion of my “missing pieces” project.

Several years ago the editorial board of the journal Epidemiologic Perspectives & Innovations, in an editorial titled “Wishful Thinking” [1], attempted to outline some of these missing pieces. We hope to expand and update this list with the special collection in this journal, “Missing Pieces: Known Unknowns in Epidemiologic Methods.”

We invite commentaries on what are believed to be difficult, yet-to-be-solved aspects of epidemiologic methods, primarily methods for answering causal questions, that we acknowledge we do not adequately understand. Our hope is that this special collection will focus and inspire research into these areas, and perhaps even inspire funding agencies to prioritize funding this research.