A brief primer on big data for surgeons

Abstract

With the advent of the age of massively‐parallel assays like Next Generation Sequencing (NGS), our capacity to generate large‐scale data has out‐paced clinicians’ ability to interpret these data in a way that is computationally and statistically robust, let alone relevant to patient care. The addition of high‐dimensional assays to more traditional methodologies carries the promise of a depth of biologic insight that was previously unattainable, but it is important to recognize the divide between the way big data scientists think and the way we think as clinicians. Increasing the dimensionality of the dataspace we are querying requires an adjustment in how we interpret data. Failure to account for the peculiarities of large‐scale data, or innocent application of scientific methodologies not intended for this dataspace will, at best, lead to the great expenditure of effort and resources in chasing false leads and, at worst, could result in misguided patient management. For those of us who spend the lion’s share of our energy in the clinical realm, acquiring the skill set required to responsibly interpret results derived from high‐ dimensional assays may seem daunting. This primer is intended to highlight some key concepts in analyzing big data and provide a framework for the critical assessment of conclusions drawn from the same.