Trying to keep up with the news has proven to be a challenging task. For this reason, we’ve decided to expand our Roundup to a twice-weekly offering. In today’s Wednesday edition: a privacy breach at a genealogy website; GitHub is purchased by Microsoft; biohackers raise alarm (and prompt spirited defenses as well); oncology makes a lot of news; and much more.
- STAT News reports that a popular genealogy site has suffered a security breach that exposed email addresses and hashed passwords of approximately 92 million users.
- A new article by Borsky and colleagues in Health Affairs addresses the notable lack of access to high-priority preventive services in the United States.
- A recent New England Journal article examining the excess mortality following Hurricane Maria (and featured in last Friday’s roundup) has generated controversy – as has the reporting surrounding it. The Washington Post provides a detailed look.
- A pioneering attempt to treat a patient with metastatic breast cancer with immunotherapy has yielded a striking success, as reported in Nature Medicine. Oncologists are now trying to temper excitement with caution while figuring out what comes next.
Researchers from Rice University and Duke University are applying hashing to improve the accuracy of casualty estimates from the Syrian civil war.
- Microsoft has purchased the development platform GitHub, triggering a sizeable migration of projects to competitor GitLab.
- At Statistical Thinking, Frank Harrell takes a deep dive into the topic of heterogeneity of treatment effect and its implications for patient care in the era of precision medicine.
- Health data, privacy, and ownership issues coalesce in a thought-provoking editorial in BMJ by Robin Baddeley. Meanwhile, author Baratunde Thurston tackles recent revelations about digital privacy and the use of personal data in a New Tech Manifesto on Medium.
- An editorial in BMJ looks at the potential offered by “digital patient encounters.”
- The transition from paper to electronic health records was relatively swift, but promised benefits and efficiencies have been slower to arrive. A new viewpoint article in JAMA examines this “productivity paradox” and offers a case for optimism.
- The 2018 Information Is Beautiful Awards are now accepting entries.
- Do-it-yourself “biohackers” equipped with cheap, powerful gene editing technologies are raising alarm in some quarters, but others assert that the movement is being unfairly maligned.
- A recent observational study published in Human Vaccines & Immunotherapeutics finds that how healthcare providers talk about vaccination can have a substantial effect on parents’ decisions to vaccinate (H/T @BrendanNyhan).
- Trials announces a call for papers for a thematic series on “big data for randomized trials.”
- An article by Johnson et al. in the Journal of the American College of Cardiology provides a thorough introduction to the application of AI and machine learning to the field of cardiology.
- A stumbling block for many deep learning applications is the costly and time-consuming process of annotating unlabeled data. In a preprint available from the BioRxiv repository, Fries and colleagues present a method for training a deep learning algorithm that circumvents this problem when applied to unlabeled cardiac MRIs.
- A new review article in the New England Journal of Medicine aims to help readers navigate the perils of multiple comparisons in clinical trials.
Some Highlights from the 2018 American Society of Clinical Oncology (ASCO) Conference
- The TAILORx trial findings have important implications for the use of adjuvant chemotherapy for breast cancer.
- The PRODIGE trial of the multi-agent mFOLFIRINOX combination therapy vs. standard gemcitabine therapy for patients with nonmetastatic pancreatic ductal adenocarcinoma reported a startling 20-month survival advantage for mFOLFIRINOX (ASCO press release).
- Next-generation sequencing offers new insights into the connection between cancer and Lynch syndrome (ASCO press release).
- The retrospective IMPACT study found that targeted cancer treatment matched to the tumor’s genetic profile yielded better outcomes than unmatched therapy (ASCO press release).