Losing weights: Failure to recognize and act on weight loss documented in an electronic health record

Background: Involuntary weight loss is associated with higher mortality. When this weight loss is unrecognized, opportunities for timely diagnosis of significant conditions may be missed. Objective: To use electronic health record (EHR) data to estimate the frequency of unrecognized involuntary weight loss and its implications. Methods: We performed a retrospective analysis of the weights recorded in an EHR of 100,000 adult patients seen in outpatient clinics over a five-year period using a novel data visualization and review tool. We reviewed charts of a random sample of 170 patients experiencing weight loss periods. Our outcomes included determinations of whether weight loss 1) was voluntary vs. involuntary; 2) was recognized and documented; and 3) possible explanations identifiable at the index visit or within the subsequent two years. Results: Of 170 randomly-selected weight loss periods reviewed, 22 (13%) were involuntary, 36 (21%) were voluntary and 112 (66%) were indeterminate. Sixty-six (39%) weight loss periods were recognized by clinician at the index visits and an additional 3 (1%) at the next PCP visits. Possible explanations for weight loss emerged in the subsequent two years including medical conditions in 60 (45%), psycho-social conditions in 19 (14%), erroneous data entry in 9 (7%), voluntary weight loss in 8 (6%), and postpartum weight loss in 6 (4%). No possible explanations were found in 32 (24%). Conclusions: Periods of weight loss were common, often involuntary and frequently not recognized or documented. Many patients with involuntary weight loss had potential explanations that emerged within the subsequent two years