News Release

Computer analysis is not better than human clinical evaluation

Peer-Reviewed Publication

American College of Physicians

1. Computer analysis is not better than human clinical evaluation when it comes to measuring breast density
Since it is more reproducible, automated BI-RADS breast density may help to identify women in need of supplemental screening

Abstract: http://annals.org/aim/article/doi/10.7326/M17-3008
Editorial: http://annals.org/aim/article/doi/10.7326/M18-0941
URLs go live when the embargo lifts

Automated and clinical evaluations of breast density similarly predict risk for both screen-detected cancer and interval cancer (cancer detected between screening examinations). These findings suggest that either measure may be used to inform women of their breast density. Findings from a case-control study are published in Annals of Internal Medicine.

Thirty states require that women receive some level of notification of their breast density. The Breast Imaging Reporting and Data System (BI-RADS) breast density categories, estimated subjectively by radiologists, are the standard for reporting breast density. Concern has been raised about using clinical BI-RADS breast density measures, as density interpretations for an individual woman may vary across mammograms and by radiologists. Automated breast density measures assessed by computer algorithms could reduce variations and alleviate subjectivity.

Researchers from the University of California, San Francisco, used data from two case-control studies to compare the predictive ability of a radiologist's assessment with that of fully automated grading. Both types of assessments were compared on the basis of their ability to predict the risk for a future screen-detected breast cancer diagnosis and the risk for a future interval invasive cancer diagnosis. The researchers found that automated and clinical BI-RADS breast density measures had similar ability to predict cancer risk, regardless of timing of density measure from cancer diagnosis. They also found that both assessments more strongly predicted interval cancer than screen-detected cancer.

While either automated or clinical BI-RADS could be used to inform women of their breast density, the researchers suggest that automated evaluations may have a place in clinical practice. Since automated findings are more reproducible, they could be used to help identify women at high risk of an interval cancer who are most in need of supplemental screening.

Media contact: For an embargoed PDF, please contact Lauren Evans at laevans@acponline.org. To interview the lead author, Karla Kerlikowske, MD, please contact Elizabeth Fernandez at Elizabeth.Fernandez@ucsf.edu.


2. Quality of hospital care less likely a factor in readmissions occurring later than a week after discharge
Abstract: http://annals.org/aim/article/doi/10.7326/M17-1724
URLs go live when the embargo lifts

Hospital readmissions that take place within the first week after hospital discharge are more likely to be preventable than those taking place later after 8 to 30 days. Late readmissions were more likely to be amenable to interventions outside of hospital control. These finding suggest that it may be time to change the model for patient outcomes after hospital discharge to one that recognizes shared accountability for readmissions along the entire spectrum of care.

Under the Patient Protection and Affordable Care Act, hospitals that have excessive readmissions within 30 days after discharge are subject to financial penalties. While many experts believe that hospitals with more frequent readmissions provide lower-quality care, it is not known how the preventability of readmissions might change over time between when the patient is discharged and 30 days later.

Researchers from Beth Israel Deaconess Medical Center studied readmission records 810 adult patients seen at 10 academic medical centers in the United States. Using computer models to assess the preventability of readmissions, the researchers found a significant difference in rates of preventability between early and late periods with the 30 days after hospital discharge. They found that early readmissions were associated with double the odds of preventability compared with late readmissions, and adjusted preventability rates clearly decreased after 7 days post-discharge.

According to the researchers' assessments, readmissions occurring in the first 7 days were more likely to be amenable to interventions within the hospital and to be caused by factors for which the hospital is directly accountable. However, late readmissions were more likely caused by factors outside of hospital control, such as appropriate monitoring and managing of symptoms after discharge by the primary care team and end-of-life preferences. The researchers suggest that a 7-day readmissions window may be a more accurate measure of hospital quality.

Media contact: For an embargoed PDF, please contact Lauren Evans at laevans@acponline.org. To interview the lead author, Kelly L. Graham, MD, MPH, please contact Jacqueline Mitchell at jsmitche@bidmc.harvard.edu.


3. Stricter research standards needed to determine if anti-obesity public health initiatives work
Position Paper Abstract: http://annals.org/aim/article/doi/10.7326/M18-0309
Research Review Abstract: http://annals.org/aim/article/doi/10.7326/M18-0501
URLs go live when the embargo lifts

The National Institutes of Health Pathways to Prevention Workshop calls for more "natural experiments" to combat the significant public health threat of obesity and recommends ways to strengthen evidence ascertained from such experiments. Natural experiments include laws or policies (such taxes on sugar-sweetened beverages); programs implemented at a community level (such as park-based physical activity programs), and environmental changes (for example, construction of dedicated bicycle lanes). A systematic review analyzed nearly 300 studies of such programs to assess their methodological and analytic strengths and weaknesses. The findings are published in Annals of Internal Medicine.

An estimated 1.9 billion people worldwide are overweight or obese, a condition contributing to epidemics of heart disease, diabetes, high blood pressure and cancer. The financial costs of obesity are high, but so also are public health interventions. In an effort to get a better view of what works and doesn't work, it's important to study outcomes after an initiative has been in place. Many studies have analyzed obesity public health programs, but it is not well understood if such studies are effective at measuring the impact of such programs.

Researchers from Johns Hopkins University School of Medicine identified 294 studies, including 156 (53 percent) natural experiments, reporting the effects of anti-obesity programs and policies. For each study, two reviewers independently extracted data on the population characteristics, data sources, measures, analytical methods, and risks of bias in study design and outcomes. The researchers found that the studies of those experiments lacked strict standards for implementation and measurement. The risk for bias was also very high. The researchers say that better standards are needed, such as stronger study designs and more standardized reporting methods.

According to the authors, the field would benefit from a clearinghouse of high-quality, evidence-based obesity prevention and control studies, held to rigorous methodological standards, to provide best-practices or 'ready-to-scale" evidence for researchers, policymakers, and practitioners.

Media contact: For an embargoed PDF, please contact Lauren Evans at laevans@acponline.org. To interview the lead author, Wendy L. Bennett, MD, MPH, please contact Chanapa Tantibanchachai at chanapa@jhmi.edu.


Also new in this issue:

A Physician's Place in the #MeToo Movement
Anne L. Peters, MD
On Being a Doctor
Abstract: http://annals.org/aim/article/doi/10.7326/M18-0271

Machine Learning and Evidence-Based Medicine
Ian A. Scott, MBBS, MHA, MEd
Ideas and Opinions
Abstract: http://annals.org/aim/article/doi/10.7326/M18-0115

Iatrogenic Infertility After Curative Stem Cell Transplantation in Patients With Sickle Cell Disease
Adrienne D. Mishkin, MD, MPH; Markus Y. Mapara, MD, PhD; Ran Reshef, MD, MSc
Ideas and Opinions
Abstract: http://annals.org/aim/article/doi/10.7326/M18-0185

Remission of Nephrotic Syndrome After Therapy for Chronic Hepatitis C Virus Infection in a Patient With Systemic Lupus Erythematosus
Hiroyuki Nakamura, MD; Yuichiro Fujieda, MD, PhD; Shinsuke Yasuda, MD, PhD; Masato Nakai, MD, PhD; Tatsuya Atsumi, MD, PhD
Brief Research Report
Abstract: http://annals.org/aim/article/doi/10.7326/L17-0759

###


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.