BACKGROUND: Breast implants are frequently used in cosmetic and reconstructive breast surgery. Capsular contracture, the most common long-term complication, is usually graded using the Baker classification. Despite its widespread use, the reliability of the Baker classification has never been established. The aim of this study was to determine the interobserver reliability and agreement of the Baker classification. METHODS: Sixty women who had undergone cosmetic breast augmentation were included. They were examined independently by two plastic surgeons from an observer pool. The Baker score was determined, along with firmness, dislocation, symmetry, and pain using four-point scales. Patients were asked to complete the BREAST-Q postaugmentation module. The interobserver reliability and agreement were calculated for all variables with a quadratic weighted kappa. RESULTS: The interobserver reliability of the Baker classification was poor (kappa, 0.55; 95 percent CI, 0.37 to 0.72). Interobserver reliability of the clinical parameters firmness (0.64; 95 percent CI, 0.49 to 0.79), dislocation (0.49; 95 percent CI, 0.26 to 0.73), and symmetry (0.61; 95 percent CI, 0.34 to 0.88) was also poor. Pain scores seemed more reliable (0.72; 95 percent CI, 0.56 to 0.89); however, most patients had no pain. The interobserver agreement for the Baker score was 48 percent; in 43 percent, the observers differed one category; and in 12 percent, the difference was more than one category. CONCLUSIONS: Interobserver reliability and observer agreement of the Baker classification for capsular contracture were poor. Consensus about how to adequately rate the symptoms of capsular complaints is lacking. A more reliable method of measurement or description is needed, especially for scientific research purposes, to assess the long-term problems associated with breast implants. CLINICAL QUESTION/LEVEL OF EVIDENCE: Diagnostic, IV.
BACKGROUND: Breast implants are frequently used in cosmetic and reconstructive breast surgery. Capsular contracture, the most common long-term complication, is usually graded using the Baker classification. Despite its widespread use, the reliability of the Baker classification has never been established. The aim of this study was to determine the interobserver reliability and agreement of the Baker classification. METHODS: Sixty women who had undergone cosmetic breast augmentation were included. They were examined independently by two plastic surgeons from an observer pool. The Baker score was determined, along with firmness, dislocation, symmetry, and pain using four-point scales. Patients were asked to complete the BREAST-Q postaugmentation module. The interobserver reliability and agreement were calculated for all variables with a quadratic weighted kappa. RESULTS: The interobserver reliability of the Baker classification was poor (kappa, 0.55; 95 percent CI, 0.37 to 0.72). Interobserver reliability of the clinical parameters firmness (0.64; 95 percent CI, 0.49 to 0.79), dislocation (0.49; 95 percent CI, 0.26 to 0.73), and symmetry (0.61; 95 percent CI, 0.34 to 0.88) was also poor. Pain scores seemed more reliable (0.72; 95 percent CI, 0.56 to 0.89); however, most patients had no pain. The interobserver agreement for the Baker score was 48 percent; in 43 percent, the observers differed one category; and in 12 percent, the difference was more than one category. CONCLUSIONS: Interobserver reliability and observer agreement of the Baker classification for capsular contracture were poor. Consensus about how to adequately rate the symptoms of capsular complaints is lacking. A more reliable method of measurement or description is needed, especially for scientific research purposes, to assess the long-term problems associated with breast implants. CLINICAL QUESTION/LEVEL OF EVIDENCE: Diagnostic, IV.