What Went Right? A Mixed-Methods Study of Positive Feedback Data in a Hospital-Wide Mortality Review Survey.

BACKGROUND: Toxic work culture contributes to healthcare worker burnout and attrition, but little is known about how healthcare organizations can systematically create and promote a culture of civility and collegiality. OBJECTIVE: To analyze peer-to-peer positive feedback collected as part of a systematized mortality review survey to identify themes and recognition dynamics that can inform positive organizational culture change. DESIGN: Convergent mixed-methods study design. PARTICIPANTS: A total of 388 physicians, 212 registered nurses, 64 advanced practice providers, and 1 respiratory therapist at four non-profit hospitals (2 academic and 2 community). INTERVENTION: Providing optional positive feedback in the mortality review survey. MAIN MEASURES: Key themes and subthemes that emerged from positive feedback data, associations between key themes and positive feedback respondent characteristics, and recognition dynamics between positive feedback respondents and recipients. KEY RESULTS: Approximately 20% of healthcare workers provided positive feedback. Three key themes emerged among responses with free text comments: (1) providing extraordinary patient and family-centered care; (2) demonstrating self-possession and mastery; and (3) exhibiting empathic peer support and effective team collaboration. Compared to other specialties, most positive feedback from medicine (70.2%), neurology (65.2%), hospice and palliative medicine (64.3%), and surgery (58.8%) focused on providing extraordinary patient and family-centered care (p = 0.02), whereas emergency medicine (59.1%) comments predominantly focused on demonstrating self-possession and mastery (p = 0.06). Registered nurses (40.2%) provided multidirectional positive feedback more often than other clinician types in the hospital hierarchy (p < 0.001). CONCLUSIONS: Analysis of positive feedback from a mortality review survey provided meaningful insights into a health system's culture of teamwork and values related to civility and collegiality when providing end-of-life care. Systematic collection and sharing of positive feedback is feasible and has the potential to promote positive culture change and improve healthcare worker well-being.

Extraneuraxial Hemangioblastoma: An Unusual Soft Tissue Neoplasm that Mimics More Common Entities but Carries Distinct Clinical Implications.

Hemangioblastoma, one of the characteristic tumors associated with Von Hippel-Lindau (VHL) disease, most often presents in the central nervous system (CNS) but can uncommonly arise in extraneuraxial, or previously referred to as peripheral, locations. Without the clinical context of known VHL disease, hemangioblastoma may not enter the differential for a soft tissue mass outside the CNS. Here, we present two patients with diagnostically challenging extraneuraxial hemangioblastoma to highlight the importance of considering this entity within the differential diagnosis of soft tissue neoplasms containing clear cells and delicate vasculature. We review the relevant diagnostic features, including a suggested immunohistochemical panel, along with the potential associated clinical implications of making this diagnosis. It is recommended that affected patients be offered genetic counseling to assess for underlying VHL disease.

Self-Supervised Contextual Language Representation of Radiology Reports to Improve the Identification of Communication Urgency.

Machine learning methods have recently achieved high-performance in biomedical text analysis. However, a major bottleneck in the widespread application of these methods is obtaining the required large amounts of annotated training data, which is resource intensive and time consuming. Recent progress in self-supervised learning has shown promise in leveraging large text corpora without explicit annotations. In this work, we built a self-supervised contextual language representation model using BERT, a deep bidirectional transformer architecture, to identify radiology reports requiring prompt communication to the referring physicians. We pre-trained the BERT model on a large unlabeled corpus of radiology reports and used the resulting contextual representations in a final text classifier for communication urgency. Our model achieved a precision of 97.0%, recall of 93.3%, and F-measure of 95.1% on an independent test set in identifying radiology reports for prompt communication, and significantly outperformed the previous state-of-the-art model based on word2vec representations.

Understanding Urgency in Radiology Reporting: Identifying Associations Between Clinical Findings in Radiology Reports and Their Prompt Communication to Referring Physicians.

In this study, we aim to develop an automatic pipeline to identify clinical findings in the unstructured text of radiology reports that necessitate communications between radiologists and referring physicians. Our approach identified 20 distinct clinical concepts and highlighted statistically significant concepts with strong associations to cases that require prompt communication.

Assisting radiologists with reporting urgent findings to referring physicians: A machine learning approach to identify cases for prompt communication.

Radiologists are expected to expediently communicate critical and unexpected findings to referring clinicians to prevent delayed diagnosis and treatment of patients. However, competing demands such as heavy workload along with lack of administrative support resulted in communication failures that accounted for 7% of the malpractice payments made from 2004 to 2008 in the United States. To address this problem, we have developed a novel machine learning method that can automatically and accurately identify cases that require prompt communication to referring physicians based on analyzing the associated radiology reports. This semi-supervised learning approach requires a minimal amount of manual annotations and was trained on a large multi-institutional radiology report repository from three major external healthcare organizations. To test our approach, we created a corpus of 480 radiology reports from our own institution and double-annotated cases that required prompt communication by two radiologists. Our evaluation on the test corpus achieved an F-score of 74.5% and recall of 90.0% in identifying cases for prompt communication. The implementation of the proposed approach as part of an online decision support system can assist radiologists in identifying radiological cases for prompt communication to referring physicians to avoid or minimize potential harm to patients.

Use of Iterative Cycles in Quality Improvement Projects in Imaging: A Systematic Review.

PURPOSE: Studies suggest that quality improvement (QI) projects in health care lack scientific rigor, but the actual frequency of use of proven scientific QI methodology is unknown. The purposes of this study are to (1) conduct a systematic review of QI projects in radiology journals on the frequency of use of iterative cycles, a marker of proven QI methodology, and (2) assess association of the use of iterative cycles with characteristics of these projects. MATERIALS AND METHODS: We searched English-language radiology journals on MEDLINE between 2008 and 2015 for published QI studies. Three reviewers appraised studies and extracted data. Use of iterative cycles was identified, and results were summarized qualitatively. χ2 Analysis evaluated associations of iterative cycles with other data elements. RESULTS: Of 3,134 potentially eligible citations, 44 studies met inclusion criteria. Only 46% of these used iterative cycles to refine intervention. Use of iterative cycles were associated with projects designed to improve process, QI expert support, reporting of unintended effect of intervention, and explicitly stated use of iterative cycles. General lack of scientific rigor was represented by failure to report baseline data (9%), describe unintended effects (66%), and discuss limitations (36%). CONCLUSIONS: Our systematic review found fewer than half of the QI projects in radiology journals used iterative cycles to refine intervention, a scientific strategy central to many proven improvement methodologies. Use of iterative approach was associated with projects designed to improve processes, QI expert support, report of unintended effect, and explicitly stated use of iterative cycles.

Deep neural networks for automatic detection of osteoporotic vertebral fractures on CT scans.

Osteoporotic vertebral fractures (OVFs) are prevalent in older adults and are associated with substantial personal suffering and socio-economic burden. Early diagnosis and treatment of OVFs are critical to prevent further fractures and morbidity. However, OVFs are often under-diagnosed and under-reported in computed tomography (CT) exams as they can be asymptomatic at an early stage. In this paper, we present and evaluate an automatic system that can detect incidental OVFs in chest, abdomen, and pelvis CT examinations at the level of practicing radiologists. Our OVF detection system leverages a deep convolutional neural network (CNN) to extract radiological features from each slice in a CT scan. These extracted features are processed through a feature aggregation module to make the final diagnosis for the full CT scan. In this work, we explored different methods for this feature aggregation, including the use of a long short-term memory (LSTM) network. We trained and evaluated our system on 1432 CT scans, comprised of 10,546 two-dimensional (2D) images in sagittal view. Our system achieved an accuracy of 89.2% and an F1 score of 90.8% based on our evaluation on a held-out test set of 129 CT scans, which were established as reference standards through standard semiquantitative and quantitative methods. The results of our system matched the performance of practicing radiologists on this test set in real-world clinical circumstances. We expect the proposed system will assist and improve OVF diagnosis in clinical settings by pre-screening routine CT examinations and flagging suspicious cases prior to review by radiologists.

Detection of femoroplasty on pre- and post-arthroscopic comparison radiographs following treatment of femoroacetabular impingement syndrome: multi-reader accuracy and agreement study.

OBJECTIVE: To assess diagnostic accuracy and agreement among radiologists in detecting femoroplasty on pre- and post-arthroscopic comparison frog lateral and anteroposterior (AP) pelvic radiographs after treatment of femoroacetabular impingement (FAI) syndrome. MATERIALS AND METHODS: In this retrospective, cross-sectional study, 86 patients underwent hip arthroscopy (52 with and 34 without femoroplasty) for treatment of FAI syndrome. Three radiologists blinded to clinical data and chronological order of the pre- and post-arthroscopic comparison radiographs independently examined AP pelvis and frog lateral radiographs to detect femoroplasty changes. Statistical analysis outputs included diagnostic accuracy parameters and inter- and intra-observer agreement. RESULTS: Identification of femoroplasty in the frog lateral projection has mean sensitivity 70%, specificity 82%, inter-observer agreement κ 0.74-0.76 and intra-observer agreement κ 0.72-0.85. Using the AP pelvis projection to detect femoroplasty has mean sensitivity 32%, specificity 71%, inter-observer agreement κ 0.47-0.65, and intra-observer agreement κ, 0.56-0.84. CONCLUSIONS: Radiologists are only moderately sensitive, though more specific, in femoroplasty detection in the frog lateral projection. The AP pelvis projection yields lower sensitivity and specificity. Both projections have moderate inter- and intra-observer agreement.

No More Waits and Delays: Streamlining Workflow to Decrease Patient Time of Stay for Image-guided Musculoskeletal Procedures.

Long wait times limit our ability to provide the right care at the right time and are commonly products of inefficient workflow. In 2013, the demand for musculoskeletal (MSK) procedures increased beyond our department's ability to provide efficient and timely service. We initiated a quality improvement (QI) project to increase efficiency and decrease patient time of stay. Our project team included three MSK radiologists, one senior resident, one technologist, one administrative assistant/scheduler, and the lead technologist. We adopted and followed the Lean Six Sigma DMAIC (define, measure, analyze, improve, and control) approach. The team used tools such as voice of the customer (VOC), along with affinity and SIPOC (supplier, input, process, output, customer) diagrams, to understand the current process, identify our customers, and develop a project charter in the define stage. During the measure stage, the team collected data, created a detailed process map, and identified wastes with the value stream mapping technique. Within the analyze phase, a fishbone diagram helped the team to identify critical root causes for long wait times. Scatter plots revealed relationships among time variables. Team brainstorming sessions generated improvement ideas, and selected ideas were piloted via plan, do, study, act (PDSA) cycles. The control phase continued to enable the team to monitor progress using box plots and scheduled reviews. Our project successfully decreased patient time of stay. The highly structured and logical Lean Six Sigma approach was easy to follow and provided a clear course of action with positive results. (©)RSNA, 2016.

Quality initiatives: statistical control charts: simplifying the analysis of data for quality improvement.

Quality improvement (QI) projects are an integral part of today's radiology practice, helping identify opportunities for improving outcomes by refining work processes. QI projects are typically driven by outcome measures, but the data can be difficult to interpret: The numbers tend to fluctuate even before a process is altered, and after a QI intervention takes place, it may be even more difficult to determine the cause of such vacillations. Control chart analysis helps the QI project team identify variations that should be targeted for intervention and avoid tampering in processes in which variation is random or harmless. Statistical control charts make it possible to distinguish among random variation or noise in the data, outlying tendencies that should be targeted for future intervention, and changes that signify the success of previous intervention. The data on control charts are plotted over time and integrated with various graphic devices that represent statistical reasoning (eg, control limits) to allow visualization of the intensity and overall effect-negative or positive-of variability. Even when variability has no substantial negative effect, appropriate intervention based on the results of control chart analysis can help increase the efficiency of a process by optimizing the central tendency of the outcome measure. Different types of control charts may be used to analyze the same outcome dataset: For example, paired charts of individual values (x) and the moving range (mR) allow robust and reliable analyses of most types of data from radiology QI projects. Many spreadsheet programs and templates are available for use in creating x-mR charts and other types of control charts.

Increased radiodensity of the proximal pole of the scaphoid: a common finding in computed tomography imaging of the wrist.

OBJECTIVES: To determine whether the proximal pole (PP) of the scaphoid is denser than the distal pole (DP) in a diverse population of patients undergoing computed tomography (CT) imaging of the wrist. METHODS: Sixty-nine patients who had CT of the wrist were retrospective reviewed. We measured CT density of the medullary bone of PP and DP of the scaphoid and calculated PP/DP ratio. To evaluate the variability of PP/DP ratio, we compared the ratios of these 3 groups. These patients were separated into 3 groups based on their diagnoses: group 1, healed scaphoids treated by casting; group 2, scaphoid nonunions; group 3, intact scaphoids. RESULTS: Proximal pole was denser than DP in most subjects. Proximal pole/distal pole ratio was similar among groups (P < 0.05). Fracture healing, sex, and age did not affect this ratio. However, we did not study scaphoids with avascular necrosis. CONCLUSIONS: Proximal pole was denser than DP in most of our subjects, including those with intact scaphoids, healing scaphoid fractures, and nonunions.

Magnetic resonance elastography of the plantar fat pads: Preliminary study in diabetic patients and asymptomatic volunteers.

OBJECTIVE: To test the feasibility of applying the magnetic resonance elastography (MRE) technique to map the elastic modulus of the plantar fat pads in diabetic and nondiabetic subjects. METHODS: A prototype MRE imaging apparatus was used to produce quantitative maps of the heel fat pad in a pilot study of 12 volunteers and 4 patients with diabetes with neuropathy. Anatomic images corresponding to MRE maps allowed precise selection of regions of interest in the fat. RESULTS: Magnetic resonance elastograms of the heel fat pads were successfully created; mean measurements in the volunteers and the diabetic patients were 4.85 and 5.26 kPa, respectively. CONCLUSION: It is feasible to perform MRE on the plantar fat pads and to produce elasticity maps. The trend toward stiffer fat pads, as demonstrated in patients with diabetes, suggests that the fat pads were qualitatively different. Magnetic resonance elastography offers great potential to investigate the mechanical properties of soft tissues in vivo noninvasively.

Forensic implications of the foot and ankle.

To clinicians, forensic medicine and the foot and ankle complex may appear to have little connection. However, if decomposed, burned, fragmented, or skeletonized human remains are recovered, the foot and ankle can provide information leading to identification and incident analysis. Furthermore, analysis of data collected at accidents can lead to safety modifications. This article reviews common elements of forensic medicine applied to the foot and ankle in 2 areas: establishing positive identification by using radiographic comparisons and analyzing aviation incidents. An understanding of these cases will enhance clinicians' awareness of the potential forensic value of the foot and ankle.