Interobserver and Intraobserver Reliability of Classification Systems for Radiographic Complications After Radial Head Arthroplasty.

PURPOSE: Although several classifications are used to assess radiographs following radial head arthroplasty (RHA), including the Popovic classification for radiolucency, the Chanlalit classification for stress shielding (SS), the Brooker classification for heterotopic ossification (HO), and the Broberg-Morrey classification for radiocapitellar arthritis, little is known about the reliability of these classification systems. The purpose of this study was to determine the interobserver and intraobserver reliability of these classifications. METHODS: Six orthopedic surgeons at various levels of training reviewed elbow radiographs of 20 patients who underwent RHA and classified them according to the Popovic, Chanlalit, Brooker, and Broberg-Morrey classifications for radiolucency, SS, HO, and RHA, respectively. Four weeks after initial review, radiographic reviews were repeated. Reliability was measured using the Fleiss kappa and the intraclass correlation coefficient. Agreement was interpreted as none (<0), slight (0.01-0.2), fair (0.21-0.4), moderate (0.41-0.6), substantial (0.61-0.8), and almost perfect (0.81-1) based on agreement among attending surgeons. RESULTS: Among fellowship-trained attending surgeons, interobserver reliability was slight for SS (Chanlalit) and the categorical interpretation of radiolucency (Popovic), fair for radiocapitellar arthritis (Broberg-Morrey) and HO (Brooker), and substantial for the ordinal interpretation of radiolucency (Popovic). Residents had a higher interobserver reliability than attending physicians when using the Brooker classification. Mean intraobserver reliability was fair for SS (Chanlalit) and the categorical interpretation of radiolucency (Popovic), moderate for HO (Brooker) and radiocapitellar arthritis (Broberg-Morrey), and almost perfect for the ordinal interpretation of radiolucency (Popovic). Trainees had higher intraobserver reliability than attending surgeons using the SS (Chanlalit) classification. CONCLUSIONS: The number of Popovic zones is reliable for communication between physicians, but caution should be taken with the Brooker, Chanlalit, Broberg-Morrey, and categorical interpretation of the Popovic classifications. All the classifications had better intraobserver than interobserver reliability. CLINICAL RELEVANCE: Reliability of classification systems for radiographic complications after RHA is less than substantial except the number of zones of radiolucency; therefore, caution is required when drawing conclusions based on these classifications.

Adult Reconstruction Fellowship Interviewee Perceptions of Virtual vs In-Person Interview Formats.

BACKGROUND: Owing to COVID-19, arthroplasty fellowship programs will be required to interview virtually for the current application cycle. Unrelated to COVID-19, our arthroplasty fellowship offered the 2019-2020 interviewees the option of an in-person or virtual interview. The purpose of the present study is to compare interviewee perceptions regarding in-person vs virtual interview formats from that application cycle at a single institution. METHODS: A 17-question survey was sent to all 26 interviewees (13 in-person and 13 virtual) shortly after the rank-list submission deadline. Interviewees were asked to what extent they agreed or disagreed with several statements, ranging from whether the interview was enjoyable to whether interviewees felt they were being adequately evaluated. In this Likert scale rating system, "strongly agree" was given 5 points (more positive outlook), and "strongly disagree" was given 1 point (more negative outlook). Chi-square analyses were performed. RESULTS: Seventeen interviewees (8 in-person and 9 virtual) returned questionnaires (response rate: 65%). Both in-person and virtual interview ratings were similar when averaged across all statements (4.5 vs 4.4, P = .67). In-person and virtual ratings were also similar for each individual statement (all P > .05). On average, interviewees spent $557/in-person interview. Fifteen (88%) said virtual interviews were more convenient, and 14 (94%) said they were more cost-effective. CONCLUSION: At a single institution, perceptions on interview format, as quantified through Likert scale ratings, were similar between in-person and virtual groups. The vast majority also viewed virtual interviews as more convenient and cost-efficient. These findings have immediate implications for future fellowship application cycles.