Reliability of Hip Fracture Classification and Treatment Planning on Smartphones Versus Picture Archiving and...

To compare the reliability of classifying and recommending treatment for femoral neck and intertrochanteric fractures when reviewing radiographs on smartphones versus Picture Archiving and Communications Systems (PACS).

Retrospective cross-sectional radiographic review.

Private academic medical center.

Consecutive patients ≥18 years with isolated unilateral femoral neck or intertrochanteric fractures (AO/OTA 31A or 31B) and complete radiographs (anteroposterior pelvis, anteroposterior and lateral hip, and anteroposterior and lateral femur) who presented from January 1 to December 31, 2021, were included.

Three orthopaedic surgeons (1 orthopaedic trauma attending and 2 adult reconstruction attendings) independently evaluated radiographs on 2 separate occasions, first on their smartphone and then on PACS using their personal laptop. After reviewing patient images, a survey assessing fracture characteristics and treatment plans was completed. Smartphone images were accessed through an embedded URL link within the survey, providing zooming functionality comparable with that of standard text message. The primary outcome was intraobserver agreement (smartphone vs. PACS) for fracture classification and treatment recommendations (general and specific) using Cohen kappa (k). Additional outcomes included interobserver agreement, image quality based on subjective evaluation, and recommendation for advanced imaging.

Radiographs from 51 patients (mean age 81.1 years, range: 61-99), consisting of 60.8% women, were reviewed. This resulted in 153 PACS and smartphone assessments. Intraobserver agreement was substantial for fracture classification (k = 0.72; 19.6% discordance), general treatment (k = 0.79; 13.1% discordance), and specific treatment (k = 0.76; 19.0% discordance). Image quality was poor/unacceptable in 5.9% of PACS and 7.2% of smartphone images ( P = 0.8). Advanced imaging was recommended in 11.1% of PACS and 11.8% of smartphone reviews ( P = 1.0). Interobserver agreement for fracture classification was moderate for PACS (k = 0.58) and smartphone (k = 0.52). Interobserver agreement for general treatment was substantial (k = 0.70 PACS, 0.65 smartphone), while specific treatment agreement was slight for PACS (k = 0.20) and fair for smartphone (k = 0.25).

When compared with PACS, smartphone review of hip fracture radiographs demonstrated substantial intraobserver agreement for fracture classification and general treatment recommendations. Agreement decreased for specific treatment decisions, with a higher frequency of indeterminate classifications and treatment recommendations during smartphone review.

Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.