Patient Sociodemographic Characteristics Associated With Saturday Breast Imaging Clinic Utilization.

OBJECTIVE: To determine patient sociodemographic characteristics associated with breast imaging utilization on Saturdays to inform potential initiatives designed to improve access and reduce disparities in breast cancer care. METHODS: This was an IRB-approved retrospective cross-sectional study. All adult women (aged ≥18 years) who received a screening or diagnostic examination at our breast imaging facility from January 1, 2016 to December 31, 2017 were included. Patient characteristics including age, race, primary language, partnership status, insurance status, and primary care physician status were collected using the electronic medical record. Multiple variable logistic regression analyses were performed to evaluate patient characteristics associated with utilization. RESULTS: Of 53 695 patients who underwent a screening examination and 10 363 patients who underwent a diagnostic examination over our study period, 9.6% (5135/53 695) and 2.0% (209/10 363) of patients obtained their respective examination on a Saturday. In our multiple variable logistic regression analyses, racial/ethnic minorities (odds ratio [OR], 1.5; 95% confidence interval [CI]: 1.4-1.6; P < 0.01) and women who speak English as a second language (OR, 1.1; 95% CI: 1.0-1.3; P = 0.03) were more likely to obtain their screening mammogram on Saturday than their respective counterparts. CONCLUSION: Racial/ethnic minorities and women who speak English as a second language were more likely to obtain their screening mammogram on Saturdays than their respective counterparts. Initiatives to extend availability of breast imaging exams outside of standard business hours increases access for historically underserved groups, which can be used as a tool to reduce breast cancer-related disparities in care.

An explainable deep-learning algorithm for the detection of acute intracranial haemorrhage from small datasets.

Owing to improvements in image recognition via deep learning, machine-learning algorithms could eventually be applied to automated medical diagnoses that can guide clinical decision-making. However, these algorithms remain a 'black box' in terms of how they generate the predictions from the input data. Also, high-performance deep learning requires large, high-quality training datasets. Here, we report the development of an understandable deep-learning system that detects acute intracranial haemorrhage (ICH) and classifies five ICH subtypes from unenhanced head computed-tomography scans. By using a dataset of only 904 cases for algorithm training, the system achieved a performance similar to that of expert radiologists in two independent test datasets containing 200 cases (sensitivity of 98% and specificity of 95%) and 196 cases (sensitivity of 92% and specificity of 95%). The system includes an attention map and a prediction basis retrieved from training data to enhance explainability, and an iterative process that mimics the workflow of radiologists. Our approach to algorithm development can facilitate the development of deep-learning systems for a variety of clinical applications and accelerate their adoption into clinical practice.

Text Message Reminders Reduce Outpatient Radiology No-Shows But Do Not Improve Arrival Punctuality.

PURPOSE: To assess whether text-based appointment reminders are a cost-effective strategy to decrease patient no-shows and improve arrival punctuality in the setting of outpatient radiology imaging. METHODS AND MATERIALS: From July 2016 through October 2016, all patients scheduled for MRI imaging at two outpatient locations were randomly assigned to a texting or nontexting arm based on the day. On texting days, patients scheduled for MRI received both the traditional phone call reminder as well as a text-based reminder of their MRI examination. On nontexting days, patients scheduled for MRI received only the traditional phone call reminder. All patients were evaluated based on whether they attended the MRI appointment and, if attended, whether they arrived 30 minutes before the MRI appointment as requested in the text message. Potential associations between the text reminder and examination attendance and punctuality were assessed by χ2 test with associations considered significant at P ≤ .05. RESULTS: A total of 6,989 patients were eligible for analysis, 3,086 in the texting group and 3,903 in the nontexting group. In the texting group, 67.5% (2,083/3,086) of patients were successfully texted with an appointment reminder, with the other 32.5% not having text accessibility. The percent of no-shows was significantly decreased for the texting group compared with the nontexting group (3.8% versus 5.1%, P = .02, odds ratio 0.75, 95% confidence interval 0.59 to 0.94). There was no significant difference between the percent of patients arriving the requested 30 minutes before the MRI appointment between the texting and nontexting groups (60.0% versus 58.5%, P = .25). CONCLUSION: Text message appointment reminders are an effective strategy for decreasing radiology no-shows, even in the presence of traditional phone reminders, but do not improve patient arrival punctuality.