Breaking News
December 16, 2018 - Parents spend more time talking with kids about mechanics of using mobile devices
December 16, 2018 - Biohaven Announces Positive Results from Ongoing Rimegepant Long-Term Safety Study
December 16, 2018 - Arterial stiffness may predict dementia risk
December 16, 2018 - Study explores link between work stress and increased cancer risk
December 16, 2018 - Sex work criminalization linked to incidences of violence finds study
December 16, 2018 - Johns Hopkins researchers discover swarming behavior in fish-dwelling parasite
December 16, 2018 - Schistosomiasis prevention and treatment could help control HIV
December 16, 2018 - Early postpartum opioids linked with persistent usage
December 16, 2018 - Johns Hopkins researchers identify molecular causes of necrotizing enterocolitis in preemies
December 16, 2018 - Advanced illumination expands capabilities of light-sheet microscopy
December 16, 2018 - Alzheimer’s could possibly be spread via contaminated neurosurgery
December 16, 2018 - Unraveling the complexity of cancer biology can prompt new avenues for drug development
December 16, 2018 - Inflammatory Bowel Disease, Prostate Cancer Linked
December 16, 2018 - Cannabis youth prevention strategy should target mental wellbeing
December 15, 2018 - Recent developments and challenges in hMAT inhibitors
December 15, 2018 - Sewage bacteria found lurking in Hudson River sediments
December 15, 2018 - CDC selects UMass Amherst biostatistician model that helps predict influenza outbreaks
December 15, 2018 - Researchers reveal brain mechanism that drives itch-evoked scratching behavior
December 15, 2018 - New computer model helps predict course of the disease in prostate cancer patients
December 15, 2018 - Obesity to Blame for Almost 1 in 25 Cancers Worldwide
December 15, 2018 - How the brain tells you to scratch that itch
December 15, 2018 - New findings could help develop new immunotherapies against cancer
December 15, 2018 - World’s largest AI-powered medical research network launched by OWKIN
December 15, 2018 - Young people suffering chronic pain battle isolation and stigma as they struggle to forge their identities
December 15, 2018 - Lifespan extension at low temperatures depends on individual’s genes, study shows
December 15, 2018 - New ingestible capsule can be controlled using Bluetooth wireless technology
December 15, 2018 - Researchers uncover microRNAs involved in the control of social behavior
December 15, 2018 - Research offers hope for patients with serious bone marrow cancer
December 15, 2018 - Link between poverty and obesity is only about 30 years old, study shows
December 15, 2018 - Mass spectrometry throws light on old case of intentional heavy metal poisoning
December 15, 2018 - BeyondSpring Announces Phase 3 Study 105 of its Lead Asset Plinabulin for Chemotherapy-Induced Neutropenia Meets Primary Endpoint at Interim Analysis
December 15, 2018 - Study finds that in treating obesity, one size does not fit all
December 15, 2018 - Tenacity and flexibility help maintain psychological well-being, mobility in older people
December 15, 2018 - Study reveals role of brain mechanism in memory recall
December 15, 2018 - High levels of oxygen encourage the brain to remain in deep, restorative sleep
December 15, 2018 - Experimental HIV vaccine strategy works in non-human primates, research shows
December 15, 2018 - Genetically modified pigs could limit replication of classical swine fever virus, study shows
December 15, 2018 - FDA Approves Herzuma (trastuzumab-pkrb), a Biosimilar to Herceptin
December 15, 2018 - Cost and weight-loss potential matter most to bariatric surgery patients
December 15, 2018 - Cancer Research UK and AstraZeneca open new Functional Genomics Centre
December 15, 2018 - New research lays out potential path for treatment of Huntington’s disease
December 15, 2018 - Prestigious R&D 100 Award presented to Leica Microsystems
December 15, 2018 - Study shows septin proteins detect and kill gut pathogen, Shigella
December 15, 2018 - Study sheds new light on disease-spreading mosquitoes
December 15, 2018 - 2017 Saw Slowing in National Health Care Spending
December 15, 2018 - Monitoring movement reflects efficacy of mandibular splint
December 15, 2018 - Study supports BMI as useful tool for assessing obesity and health
December 15, 2018 - Self-guided, internet-based therapy platforms effectively reduce depression
December 15, 2018 - Organically farmed food has bigger climate impact than conventional food production
December 15, 2018 - Faster, cheaper test has potential to enhance prostate cancer evaluation
December 15, 2018 - Researchers study abnormal blood glucose levels of patients after hospital discharge
December 15, 2018 - Swedish scientists explore direct association of dementia and ischemic stroke deaths
December 15, 2018 - Study finds 117% increase in number of dementia sufferers in 26 years
December 15, 2018 - Eczema Can Drive People to Thoughts of Suicide: Study
December 15, 2018 - Link between neonatal vitamin D deficiency and schizophrenia confirmed
December 15, 2018 - Nurse denied life insurance because she carries naloxone
December 15, 2018 - Ritalin drug affects organization of pathways that build brain networks used in attention, learning
December 15, 2018 - Research pinpoints two proteins involved in creation of stem cells
December 15, 2018 - Gut bacteria may modify effectiveness of anti-diabetes drugs
December 15, 2018 - A new type of ‘painless’ adhesive for biomedical applications
December 15, 2018 - Early physical therapy associated with reduction in opioid use
December 15, 2018 - Breast cancer protection from pregnancy begins many decades later, study finds
December 15, 2018 - How often pregnant women follow food avoidance strategy to prevent allergy in offspring?
December 15, 2018 - Using machine learning to predict risk of developing life-threatening infections
December 15, 2018 - How imaginary friends could boost children’s development
December 15, 2018 - Folate deficiency creates more damaging chromosomal abnormalities than previously known
December 15, 2018 - Study provides new insights into molecular mechanisms underlying role of amyloid in Alzheimer’s disease
December 15, 2018 - For the asking, a check is in the mail to help pay for costly drugs
December 15, 2018 - UA scientists uncover biological processes leading to rare brain disorder in babies
December 15, 2018 - The largest database on industrial poisons
December 15, 2018 - ESMO Immuno-Oncology Congress showcases novel technologies set to benefit many cancer patients
December 15, 2018 - Ovid Therapeutics Announces Plans to Move into a Phase 3 Trial in Pediatric Patients Based on End-of-Phase 2 Meeting for OV101 in Angelman Syndrome
December 15, 2018 - Left ventricular noncompaction – Genetics Home Reference
December 15, 2018 - Children’s sleep not significantly affected by screen time, new study finds
December 15, 2018 - When should dementia patients stop driving? A new guidance for clinicians
December 15, 2018 - Researchers use INTEGRA’s VIAFLO 96/384 to streamline the experimental workflow
December 15, 2018 - Researchers discover protein involved in nematode stress response
December 15, 2018 - Cancer patients have greater risk of developing shingles, study shows
December 14, 2018 - UAlberta scientists identify biomarkers for detecting Alzheimer’s disease in saliva samples
December 14, 2018 - Study uncovers link between tube travel and spread of flu-like illnesses
How should an algorithm generate recommendations for patient care?

How should an algorithm generate recommendations for patient care?

image_pdfDownload PDFimage_print

Say you’re a doctor. You’d like guidance on how to treat a particular patient, and you have the opportunity to query a group of physicians about what they’d do next.

Who do you include in that group? All doctors who have seen similar patients? A smaller number of doctors who are considered the best?

These questions are at the root of recent Stanford research.

The study, published in the Journal of Biomedical Informatics, examines a key aspect of medical artificial intelligence: If machines are to provide advice for patient care, who should those machines be learning from?

Jonathan Chen, MD, PhD, tackled this question as the latest step in his quest to build OrderRex, a tool that will mine data from electronic health records to inform medical decisions. In a prototype, an algorithm that works like Amazon’s recommendation feature supplied information to doctors on how their peers managed similar patient cases — i.e., it tells users that other doctors in this situation ordered this medication or this test.

Now, Chen and his colleagues are working through logistics of how the computer decides what to recommend. Research assistant Jason K. Wang is the study’s lead author. Chen, the senior author, told me,

The distinguishing question was, if you had to learn medical practices, should you just try to get everybody’s data and learn from everybody? Or should you try to find the ‘good’ doctors? Then that evokes some really deep follow-up questions, such as, what is a preferred doctor or a good expert? What does that mean? How would we define that?

He asked multiple doctors a simplified version of these questions — if you had to grade each doctor by a number, how would you come by that number? — and tumbled into the usual thicket surrounding physician evaluations. Which outcomes should be considered? What if you see sicker patients? What about hospital readmission, length of stay, and patient satisfaction?

Ultimately, Chen and his colleagues settled on the measure of 30-day patient mortality rate: actual number of deaths compared to expected deaths, as reliably calculated by a computer using three years of electronic health record data.

“We tried a couple of things, but found they were correlated enough that it basically didn’t matter. The doctors who rose to the top of the list versus the bottom — it was a pretty stable list,” Chen said. “Plus, you can’t really game mortality: it’s the most patient-centered outcome there is.”

With “expert” doctors identified as those with low patient mortality rates, the researchers turned to defining “optimal” care to compare with recommendations that the algorithm would produce. They decided to try two different options. One was care that followed standards from clinical practice guidelines based on literature available through such venues as the National Guideline Clearinghouse. The second was care reflecting patterns that typically led to better-than-expected patient outcomes — calculated by the computer from data in the electronic patient records.

With this established, they tested the algorithm to see how its recommendations for care for a condition such as pneumonia compared to both standard clinical practice guidelines and to computer-generated above-average patient care. First, they fed the algorithm information from patients seen by expert doctors. Then, separately, the algorithm looked at patients seen by all doctors.

The result? Not much difference between recommendations derived from care in the expert group versus all of the doctors.

Essentially, Chen said, outlier behavior from specific doctors was neutralized. And without significant differences in the results, the findings argue for building machine-learning models around data from more physicians rather than a curated bunch. The logic is similar to calculating an average: with more numbers in the equation, each number carries less weight and the average is less influenced by individual numbers — including anomalies.

Chen summed it up:

As a human, it’s not feasible to interview thousands of people to help learn medical practice, so I’m going to seek out a few key mentors and experts. What large datasets and computational algorithms enable us to do is to essentially learn from ‘everybody,’ rather than being constrained to learn from a small number of “experts.”

With this question resolved, Chen and his colleagues have now built a prototype user interface incorporating the suggestion system, and they plan to bring in doctors to test it with simulated patient cases.

“The next big challenge to closing the loop on a learning health care system,” he said, “is… where we don’t just learn interesting things from clinical data, but we design, study, and evaluate how to deliver that information back to clinicians and patients.”

Photo by Markus Spiske

Tagged with:

About author

Related Articles