Is it surprising to find that general surgeons at the beginning or end of their careers perform fewer operations and have higher complication rates than middle-career surgeons in their prime? This seems so intuitively true that one might seriously question the validity of a study that failed to show it. Indeed, Studnicki et al1 confirm this correlation using data from a Florida discharge database and online practitioner profiles. However, the same methods reveal other, more puzzling associations. For example, compared with earlier cohorts, late-career surgeons appeared to treat more Hispanic patients, and they performed more cardiovascular procedures and fewer gastrointestinal procedures, with lower complications for the cardiovascular procedures but higher complications for the gastrointestinal procedures. Are these findings also valid? Maybe, but they more likely reflect the limitations of the study design and data interpretation. The authors acknowledge difficulty in adjusting databases for case mix, case complexity, and other potentially confounding factors. Several relationships probably are surrogate: it seems more plausible that the ethnicity of patients relates to the ethnicity of physicians rather than their age. An inherent limitation of the study's cross-sectional design is the inability to evaluate the effect of changing training paradigms and practice patterns, such as subspecialization, recertification, local credentialing, and duty-hour restrictions. The study takes a snapshot of four 10-year cohorts during a short period (2004-2006) rather than studying a single cohort during a career. The discrepant case ratios and outcomes between late-career surgeons and the other cohorts, even if real, may no longer be the case in 2011, after most of the late-career cohort has ceased to practice.