Purpose: Orthopaedic surgeons are frequently presented with advertisements for orthopaedic and medical products in which companies make claims of clinical and scientific fact. This study was designed to evaluate the statements made in orthopaedic print advertisements and determine whether they are supported by scientific data.
Methods: Fifty statements from fifty advertisements were chosen at random from six peer-reviewed orthopaedic journals. The companies that placed the advertisements were contacted to provide supporting data for the statement of clinical or scientific fact. Three senior orthopaedic surgeons evaluated the data for quality and support. A high-quality study was defined as a study that could be published in the peer-reviewed literature. A well-supported statement was defined as a statement with enough supporting evidence to be used in clinical practice. The evaluating surgeons were blinded to product and company identification.
Results: The supporting data were from a published source for eighteen claims (36%), from a presentation at a public forum or a scientific meeting for twelve claims (24%), or were "data on file" only at the company for twelve claims (24%). Interobserver agreement among the surgeons evaluating the advertisements for quality and support was good (the average intraclass correlation coefficient was 0.72). Of the fifty claims, twenty-two were considered unsupported by scientific data, seventeen were classified as possibly supported, seven were well supported, and four were from companies that did not respond despite three requests. Claims that were supported by published data were significantly more likely to be rated as well supported (p < 0.001). All twelve claims that were supported purely by "data on file" at the company were considered to be poorly supported.
Conclusions: Orthopaedic surgeons should interpret claims made in orthopaedic print advertisements with caution. Approximately half of the claims are not supported by enough data to be used in a clinical decision-making process.