Measurement of the quality of facial lesion reconstruction with observer-graded affect display

Maria Phillis, Masaru Ishii, Jacob Dey, Jennifer Goines, Patrick J. Byrne, Kofi D.O. Boahene, Lisa E. Ishii

Research output: Contribution to journalArticlepeer-review

6 Scopus citations


IMPORTANCE: The impact of facial reconstructive surgery on improving affect display for patients with facial lesions is incompletely understood. OBJECTIVES: To measure the impact of surgically reconstructing facial lesions on observer graded affect display. DESIGN, SETTING, AND PARTICIPANTS: This was a prospective randomized controlled experiment at an academic tertiary referral center; 120 naïve observers completed one of 4 surveys of 20 images of faces. MAIN OUTCOMES AND MEASURES: A total of 80 different images of faces were used: 32 faces with lesions preoperatively and postoperatively, and 16 normal faces. The 32 lesion faces were categorized into 4 categories of lesion with 8 faces in each category: small peripheral, small central, large peripheral, and large central. Participants viewed 20 images of faces: 8 preoperative, 8 postoperative, and 4 normal. Observers rated faces by selecting all terms that applied from 8 affects of the Derogatis Affects Balance Scale (DABS), the Ekman scale's 6 basic emotions, and a neutral term. Latent class analysis and regression were used to determine number and characteristics of affect classes and significance of covariates: operation status, size, and location. Postoperative faces were compared with normal faces via bootstrap analysis to determine statistically significant differences in affect display. Analysis took place in August 2014 and was verified in December 2015. RESULTS: A total of 120 observers rated 80 facial images via 4 surveys. The mean (SD) age of the observers was 32.2 (13.5) years (range, 18-72 years). Both scales collapsed into 3 latent classes: positive, neutral, and negative for normal faces and faces with lesions. Using the DABS, faces with lesions preoperatively were more likely to be in the negative affect class (small peripheral [SP]: mean, 0.384; 95 % CI, 0.325-0.444; small central [SC]: mean, 0.505; 95% CI, 0.433-0.577; large peripheral [LP]: mean, 0.583; 95% CI, 0.519-0.647; large central [LC]: mean, 0.475; 95% CI, 0.412-0.539) than normal faces (mean, 0.235; 95% CI, 0.190-0.281) or postoperatively (SP: mean, 0.141; 95% CI, 0.106-0.176; SC: mean, 0.207; 95% CI, 0.160-0.254; LP: mean, 0.272; 95% CI, 0.219-0.325; LC: mean, 0.189; 95% CI, 0.149-0.228) and less likely to be in the positive affect class (SP: mean, 0.213; 95% CI, 0.168-0.257; SC: mean, 0.204; 95% CI, 0.164-0.243; LP: 0.130; 95% CI, 0.096-0.164; LC: mean, 0.213; 95% CI, 0.172-0.254) than normal faces (mean, 0.459; 95% CI, 0.399-0.519) or postoperatively (SP: mean, 0.384; 95% CI, 0.323-0.445; SC: mean, 0.410; 95% CI, 0.350-0.469; LP, 0.298; 95% CI, 0.243-0.354; LC: mean, 0.415; 95% CI, 0.358-0.472). Postoperatively, faces with SP, SC, and LC lesions were restored to similar or improved levels of positive and negative affect as normal faces. Faces with LP lesions were improved in positive and negative affect from preoperatively, and restored to similar levels of negative affect as normal faces, but still had lower levels of positive affect compared with normal faces. Results using the Ekman scale were similar. CONCLUSIONS AND RELEVANCE: Facial reconstructive surgery was effective in restoring affect display in patients with facial lesions. This provides new data supporting the positive impact of reconstructive surgery on patients' quality of life. LEVEL OF EVIDENCE: NA.

Original languageEnglish (US)
Pages (from-to)467-473
Number of pages7
JournalJAMA Facial Plastic Surgery
Issue number6
StatePublished - Dec 1 2016

ASJC Scopus subject areas

  • Surgery


Dive into the research topics of 'Measurement of the quality of facial lesion reconstruction with observer-graded affect display'. Together they form a unique fingerprint.

Cite this