Self-Compression Works | No Surprises in California | AI Follow-up Success

“Physicians in anesthesiology, radiology, and orthopedic practices reported unprecedented decreases in payers’ offered rates and less interest in contracting since AB-72 was passed into law.”

RAND on the impact of California’s 2017 surprise billing law.


Imaging Wire Sponsors

  • Carestream – Focused on delivering innovation that is life changing – for patients, customers, employees, communities and other stakeholders
  • Focused Ultrasound Foundation – Accelerating the development and adoption of focused ultrasound
  • Medmo – Helping underinsured Americans save on medical scans by connecting them to imaging providers with unfilled schedule time
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter
  • Pocus Systems – A new Point of Care Ultrasound startup, combining a team of POCUS veterans with next-generation genuine AI technology to disrupt the industry
  • Qure.ai – Making healthcare more accessible by applying deep learning to radiology imaging

The Imaging Wire


More Self-Compression Evidence

A Mass General Hospital team found that when women use a patient-assisted compression (PAC) device to self-control mammography compression, it results in similar image quality as technologist-controlled mammograms (TC) and improves their screening experience.

  • Study – The researchers looked at 685 women who underwent synthetic 2D/tomosynthesis mammography screening, comparing the PAC and TC methods using survey responses and clinical metrics.
  • Results – The 148-woman survey (50 who used PAC) found that 73% of the women who used the PAC device preferred it over TC, while 83% didn’t have anxiety during PAC-based screenings. The study found no difference between PAC and TC for image quality, compression thickness, and average glandular dose. Women using the PAC device also applied “significantly” more compression force than in TC-based screenings.
  • More Evidence – This study comes about six months after French researchers found that women who use self-compression applied more force, but reported less pain and their scans had equal image quality.
  • Significance – Given the fact that compression pain is a major reason that women avoid mammograms, this study suggests that self-compression “may be a relatively simple way to increase their adherence to screening mammography.”



No Surprises in California

For those wondering why there’s so much industry attention on federal surprise billing legislation, a new RAND study on the impact of California’s 2017 surprise billing law confirmed that out-of-network (OON) payment standards can indeed reduce providers’ negotiating power. The law also did what it was supposed to do and limited California patients’ surprise bills.

  • California Assembly Bill 72 (AB-72) – The two year-old law limited how much providers could charge patients for out-of-network services, requiring payers to reimburse out-of-network providers at either their local average contract rate or at 125% of Medicare’s fee-for-service rate (whichever is higher).
  • The Study – Interviews of 28 stakeholders (industry groups, practice executives, benefits companies) revealed that AB-72 significantly impacted negotiations between hospital-based physicians and payers.
  • The Fallout – By setting a maximum rate, the California law “appears to be reducing physicians’ leverage to negotiate higher in-network payments, and in turn is speeding the consolidation of physician groups as they seek to regain lost leverage.” In the meantime, the law is resulting in “unprecedented decreases” in rates to anesthesiology, radiology, and orthopedic practices.

This is exactly what the folks on the provider side of the federal legislation debate have been arguing will happen on a national scale if congress chooses to base OON reimbursements on a median benchmark (not arbitration).



AI Follow-up Success

A team from Philips, Harvard Medical, and U of Washington developed an algorithm to help ensure radiology report follow-ups take place by automatically identifying the completion of a recommended follow-up imaging study.

  • Study – Using 559 imaging reports that featured follow-up recommendations and 8,691 subsequent reports, the researchers had three radiologists identify appropriate follow-up exams to create a ground truth dataset. They then trained the ML and NLP-based algorithm with recommendation attributes, report text, and metadata to determine the most likely follow-up exam based on a preceding recommendation.
  • Results – The algorithm revealed that 172 of the 559 original cases did not complete the recommended follow-ups, achieving an F-score of 0.807 compared to the radiologists’ 0.853 to 0.868 F-score range assessing the same cases.
  • Application – Calling the algorithm results “acceptable,” the team suggested that a similar algorithm could be “integrated into a follow-up management application” that proactively sends reminders to referrers, PCPs, and patients to ensure follow-ups.

Noting that members of this same research team previously found that only about half of all follow-ups actually take place, a solution like this certainly seems reasonable.


The Wire

  • A Boston-based research team found that emergency physicians use imaging in inconsistent ways, calling for new measures to help improve ED imaging order quality and reduce waste. The study looked at all ED visits at a level 1 trauma center for a year (n=56,793 patients, 51 physicians), finding that 49.5% of the patients underwent imaging during their visits (38.2% low-cost imaging, 21.9% high-cost imaging). After controlling for various factors, the study found that these significant variations in low-cost and high-cost imaging were attributable to individual physician decision-making, as 47% of the physicians ordered more high-cost imaging than average and 49% ordered more low-cost images than average.
  • While we’re at it, research from the University of Washington (n = 149 litigated exams from 18 states) found that 46% of radiology malpractice actions stem from emergency department patients, compared to 38% from outpatient and 17% from inpatient. The study suggested that ED’s higher malpractice risk may be because ED scans are read by a variety of specialists, a lack of subspecialty training, and emergency radiology’s fast reading pace and unusual hours.
  • A German study published in Science of the Total Environment revealed that gadolinium was present in municipal tap water in six major German cities and made its way into tap water-based soft drinks in the regions’ fast food chains, serving as the first evidence of GBCAs entering the human food chain.
  • IHS Markit reports that global CT equipment revenues grew by 7.5% in 2018 to $4 billion, driven by upgrades to higher-slice equipment (>128-slice grew by 10%) and demand for ROI-driving workflow and usability features. GE Healthcare and Siemens Healthineers maintained the top two spots in the CT market (now combine for 60% of market), followed by Canon and Philips. The Chinese CT market saw the greatest growth and the European market was strong, while the U.S. market was flat, and the Middle East and Japan both saw declines.
  • IHS Markit reported that the global MRI equipment market grew by 3% in 2018 to $4.3 billion, as increased competition and greater use of bulk purchasing drove down market pricing, causing manufacturers to embrace a partnership-based approach and adopt efficiency and ROI-focused features. Siemens Healthineers and GE Healthcare maintained their spots at the top of the MRI market, and although IHS didn’t specify Canon/Philips/Hitachi’s share, it did highlight the increased competition from United Imaging’s U.S. entrance. The Chinese market led all regions with double digit growth due to new healthcare reform policies, while U.S. revenues fell by 1.4% (the U.S. is expected to grow again in 2019).
  • The Royal Australian and New Zealand College of Radiologists (RANZCR) released its Ethical Principles for Artificial Intelligence in Medicine, providing doctors, developers, and healthcare organizations with guidelines for the research and deployment of medical ML systems and AI tools. RANZCR’s first major milestone since launching this initiative in early 2019 is based on nine principles (safety, privacy, avoiding bias, transparency/explainability, applying human values, decision making on diagnosis and treatment, teamwork, responsibility) as well as broader ethical frameworks.
  • AMGA Consulting’s 2019 Medical Group Compensation and Productivity Survey (n= over 117,000 providers) found that although physician compensation increased by a median of 2.92% in 2018, diagnostic radiology compensation fell by 1% to $482,599 (only 26% of specialties saw compensation fall). Productivity (measured in work relative value units – wRVUs) increased by 0.29% across all physicians and average compensation per wRVU increased by 3.64%, while radiology saw a much greater 5.9% increase in wRVUs along with a 2.2% drop in compensation per wRVU.
  • A UW-Madison professor developed a new image reconstruction method (for CT, PET, SPECT and MRI) that could improve image quality by separating sections of a scan that contain artifacts and then recombining the sections using a set of image references. The method examines data inconsistency metrics (DIMs) to identify and select data sets with low inconsistency and then reconstructs sub-images with “minimal artifact contamination.”
  • A new study published in AJR validated an ML model that uses CT image texture analysis to evaluate immunohistochemical (IHC) characteristics in patients with suspected thyroid nodules, applying these characteristics to support thyroid nodule diagnosis. The study looked at 103 patients who underwent CT before thyroidectomy surgery (using 2/3rds of the patients and 86 radiomic features for training), accurately predicting the presence of cytokeratin 19 (84.4% training, 80.0% validation), galectin 3 (82.5% training, 85.0% validation), and thyroperoxidase (81.4% training, 84.2% validation).
  • Baltimore’s F.M. Kirby Imaging Center at Kennedy Krieger Institute will install a Magnetic Insight Magnetic Particle Imaging (MPI) scanner. Johns Hopkins researchers will use the new imaging system for a range of precision medicine studies including stem cell tracking, immunotherapy, tumor hyperthermia, treatment of diabetes, gene therapy, and drug delivery.

The Resource Wire

– This is sponsored content.

  • This Carestream case study compares images of foot trauma captured using the OnSight 3D Extremity System to images captured on 2D X-rays.
  • This Healthcare IT News article details how the University of Rochester Medical Center’s Backstop program leverages Nuance’s mPower Clinical Analytics and PowerScribe Follow-up Manager solutions to track and complete more follow-up imaging recommendations.
  • By partnering with Medmo, imaging centers can keep their schedules full, their equipment busy, and increase revenue. Here’s where to get started.

Leave a Reply

Basic HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS

Get every issue of The Imaging Wire

Delivered right to your inbox. We will never sell or share your information.