April 2020 HSC Section 4 - Plastic and Reconstructive Problems

Research Original Investigation

Standardized Cosmesis and Health Nasal Outcomes Survey

within each particular domain but low between the items be- longing to different domains. For all questions, less than 3.6% of data were missing. The missing data were included into analysis. Internal consistency of each of 2 domains was examined using Cronbach α. For both domains, the Cronbach α was ex- cellent: 0.94 (95%CI, 0.92-0.95) for the nasal obstruction do- main and 0.94 (95% CI, 0.93-0.95) for the cosmesis domain. There was no need for increasing α by eliminating items. The factor structures of the domainswere evaluated using exploratory factor analysis (Table 2). For the nasal obstruc- tion domain, the parallel analysis of the scree plot suggested a single retained factor with eigenvalue of 3.13 (eFigure 2A in the Supplement ). The parallel analysis of the scree plot for the cosmesis domain suggested 2 retained factors with eigenval- ues 4.35 and 0.28 (eFigure 2B in the Supplement ). However, the second factor was substantially below the Kaiser cutoff point of retaining and, therefore, the domains were consid- ered unidimensional. Factor loadings (pattern matrix) and unique variances of items are also presented in Table 2. The loading estimateswere high for all the items varying from0.74 to 0.92 (Table 2). We then evaluated patient responses (n = 191 patients) for all individual questions, composite obstruction and cosme- sis scores, as well as the NOSE score (eFigure 3 in the Supple- ment ). Kruskal-Wallis testing showed a significance level of P < .001 when evaluating the difference between groups (pre- operative cosmetic, postoperative cosmetic, preoperative func- tional, postoperative functional, and nonrhinoplasty) for all individual questions, composite scores, andNOSE score. Cor- relations between the obstruction composite score and the NOSE scores were r = 0.943 ( P < .001), which is very strong. The obstruction and cosmesis composite scores were only weakly correlated ( r = 0.388; P < .001). Discussion Four questionnaires have been identified in a previous sys- tematic review 6 as intended to assess both functional and cos- metic aspects of rhinoplasty. These are the Rhinoplasty Out- comes Evaluation, 11 the Functional Rhinoplasty Outcome Inventory 17, 12 theRhinoplastyHealth InventoryandNasal Out- comes scale, 13 and the Expectations of Aesthetic Rhinoplasty Scale. 14 Unfortunately, none of these questionnaires were de- veloped using patient input through the form of interviews, which limits their usefulness to assessingpatient-centeredout- comes. The advantage of the SCHNOS questionnaire is that the questions retained in the initial questionnaire were identi- fied by patients as most relevant to their experience. Themost widely used and accepted PROM to assess nasal obstruction in rhinoplasty is currently the NOSE question- naire, developed by Stewart et al. 4 The nasal obstruction com- posite score of the SCHNOS showed excellent correlationwith the NOSE questionnaire, with a Pearson correlation coeffi- cient of 0.943. The SCHNOS may therefore be administered without the NOSE scale to assess the construct of nasal obstruction.

nasal cosmesis were calculated separately as sums of the re- sponses to each item from0 to 5. This way themaximum total score for the nasal obstructiondomainwas 20, and, for the cos- mesis domain, 30 points. Spearman rank correlationwas used to assess the strength of correlation between different items. All the analyses in phase 3 were conducted using Stata/IC sta- tistical software (release 14, StataCorp LP). Phase 4: Field Testing Finally, preliminary field testing of the questions and scores was performed in the same population of 191 patients. Pa- tients were divided into 4 groups depending on their preop- erative or postoperative status andwhether their surgerywas purely functional or had a cosmetic component. A fifth group of nonrhinoplasty patients, who presented to the clinic for other facial cosmetic or reconstructive reasons, was also in- corporated into analysis. For each group, the results of the Lik- ert scales of each question were compared using Kruskal- Wallis H testing using P <.05 as statistically significant. Composite scores for nasal obstructionandnasal cosmesiswere also calculated by adding up the scores of questions domain- specific questions, and compared between groups using the same method (as described herein). Results of NOSE scores were also contrasted between groups with the same method. Pearson correlation coefficients between the composite scores and NOSE score were also calculated. For phase 4, SPSS sta- tistical software (version 22; IBM Corp) was used. Results Of the 191 patients invited to answer the questionnaire, 191 of 191 answered (100%response rate): 67%(128)werewomenand 33% (63) weremen. Their mean (SD) age was 41.5 (15.8) years. A total of 18 patients were invited to participate for the inter- views and 18 of 18 accepted to participate (a 100% response rate): 9 cosmetic, 4 functional, and 4 combined cosmetic and functional rhinoplasty patients. Of these patients, 5weremale and 13 were female. Their mean (SD) age was 38 (14.8) years (range, 20-64 years). Nine patients were preoperative pa- tients, and 9 patients were postoperative. A total of 5 experts were interviewed. Regardless of the indications for rhino- plasty, most of the patients, and all of the experts, perceived cosmesis and obstruction to be equally important during the initial interview phase. The most commonly elicited and pri- oritized concepts were included in the preliminary version of the questionnaire (eFigure 1 in the Supplement ). Of the 191 patients, 107 (56%) underwent rhinoplasty for cosmetic reasons, with or without a functional indication. Functional rhinoplasty was performed in 52 patients (27%). Thirty-two patients (17%) were nonrhinoplasty patients pre- senting for a variety of other conditions (eg, facial trauma, skin cancer reconstruction). Of the operative patients, the SCHNOS was applied preoperatively in 90 patients and after the sur- gery in 69 patients. The correlation matrix for the 10 items is presented in Table 2 . The correlations between pairs of items variedwidely, from 0.13 to 0.85. As expected, the correlations were high

JAMA Facial Plastic Surgery January/February 2018 Volume 20, Number 1 (Reprinted)

jamafacialplasticsurgery.com

© 2017 American Medical Association. All rights reserved.

23