Breaking News
February 17, 2019 - Sinks by Toilets in ICU Patient Rooms Harbor Harmful Bacteria
February 17, 2019 - Cancer cells’ plasticity makes them harder to stop
February 17, 2019 - Young cannabis users have increased risk of depression and suicidal behavior
February 17, 2019 - Tasmanian Devils Likely to Survive Cancer Scourge
February 17, 2019 - Neoadjuvant PD-1 blockade seems effective in glioblastoma
February 17, 2019 - Personal, social factors play role in enabling sustainable return to work after ill health
February 17, 2019 - Mouse studies show ‘inhibition’ theory of autism wrong
February 17, 2019 - Study shows how neuroactive steroids inhibit activity of pro-inflammatory proteins
February 17, 2019 - Use of liver grafts from older donors decreased despite better outcomes in recipients
February 17, 2019 - MUSC researchers discover new mechanism for a class of anti-cancer drugs
February 17, 2019 - HPV misconceptions are causing women to miss smear tests
February 17, 2019 - Sanofi and Regeneron Offer Praluent (alirocumab) at a New Reduced U.S. List Price
February 17, 2019 - Researchers say auditory testing can identify children for autism screening
February 17, 2019 - New method analyzes how single biological cells react to stressful situations
February 17, 2019 - WVU gynecologic oncologist investigates novel treatment for cervical and vaginal cancers
February 17, 2019 - ADHD diagnoses poorly documented
February 17, 2019 - Majority of gender minority youth do not identify with traditional sexual identity labels
February 17, 2019 - AbbVie, Teneobio enter into strategic transaction to develop potential treatment for multiple myeloma
February 17, 2019 - Lower Birth Weight May Up Risk for Psychiatric Disorders
February 17, 2019 - Scientists identify reversible molecular defect underlying rheumatoid arthritis
February 17, 2019 - Moffitt researchers shed light on how CAR T cells function mechanistically
February 16, 2019 - Female Anatomy May Play Big Role in Sperm’s Success
February 16, 2019 - BMI may mediate inverse link between fiber intake, knee OA
February 16, 2019 - Movement impairments in autism can be reversed through behavioral training
February 16, 2019 - Studies address racial disparities in postpartum period and cardiovascular health
February 16, 2019 - Scientists implicate hidden genes in the severity of autism symptoms
February 16, 2019 - Decreased deep sleep linked to early signs of Alzheimer’s disease
February 16, 2019 - Neuroscientists show how the brain responds to texture
February 16, 2019 - Gilead Announces Topline Data From Phase 3 STELLAR-4 Study of Selonsertib in Compensated Cirrhosis (F4) Due to Nonalcoholic Steatohepatitis (NASH)
February 16, 2019 - What Can I Do About Sweating? (for Teens)
February 16, 2019 - Companies navigate dementia conversations with older workers
February 16, 2019 - Newly developed stem cell technologies show promise for treating PD patients
February 16, 2019 - Collaborative material research could advance self-assembling nanomaterials
February 16, 2019 - Researchers take major step in creating technology that mimics the human brain
February 16, 2019 - Erasing memories associated with cocaine use reduces drug seeking behavior
February 16, 2019 - Artificial intelligence can accurately predict prognosis of ovarian cancer patients
February 16, 2019 - Racial disparities in cancer deaths on the decline for America
February 16, 2019 - FDA authorizes new interoperable insulin pump for children, adults with diabetes
February 16, 2019 - Coexisting Medical Conditions, Smoking Explain PTSD-CVD Link
February 16, 2019 - Skin Cancer Screening: MedlinePlus Lab Test Information
February 16, 2019 - ‘Happiness’ exercises can boost mood in those recovering from substance use disorder
February 16, 2019 - Cell manipulation could soon halt or reverse aging
February 16, 2019 - Pumped Breast Milk Falls Short of Breastfed Version
February 16, 2019 - Men’s porn habits could fuel partners’ eating disorders, study suggests
February 16, 2019 - Rapid progression of age-related diseases may result from formation of vicious cycles
February 16, 2019 - Immune checkpoint molecule protects against future development of cancer
February 16, 2019 - New method produces hydrogels that have properties similar to cells’ environment
February 16, 2019 - $4.1 million funding for heart research on Valentine’s Day
February 16, 2019 - General anesthesia in early infancy unlikely to have lasting effects on developing brains
February 16, 2019 - New breakthroughs for muscular dystrophy research
February 16, 2019 - First Opinion: Embryo editing for higher IQ is a fantasy. Embryo profiling for it is almost here
February 16, 2019 - Vapers develop cancer-related gene deregulation as cigarette smokers
February 16, 2019 - Bringing Antimicrobial Susceptibility Testing (AST) to the Community
February 16, 2019 - Decolonization protocol after hospital discharge can prevent dangerous infections
February 16, 2019 - Children with ASD more likely to face maltreatment, study finds
February 16, 2019 - Study finds genetic vulnerability to use of menthol cigarettes
February 16, 2019 - Promising drug developed to rejuvenate muscle cells
February 16, 2019 - H-RT should be the standard of care for men with low risk prostate cancer, study shows
February 16, 2019 - New technique using patients’ own modified cells could help treat Crohn’s disease
February 16, 2019 - Therapeutic endoscopy has an expanding role in the treatment of IBD
February 16, 2019 - Blood clot discovery could lead to development of better treatments for blood diseases
February 16, 2019 - Intervention can increase exclusive breastfeeding rates
February 16, 2019 - New project explores how gaming technologies can help cancer patients communicate better
February 16, 2019 - Catalyst Biosciences Presents Updated Data from Its Phase 2/3 Trial of Subcutaneous Marzeptacog Alfa (Activated) in Individuals with Hemophilia A or B with Inhibitors at the 12th Annual EAHAD Congress
February 16, 2019 - Rerouting nerves during amputation reduces phantom limb pain before it starts
February 16, 2019 - A Hormone Produced When We Exercise Might Help Fight Alzheimer’s
February 16, 2019 - Millions of British people breathe toxic air travelling to GPs
February 16, 2019 - Conformance of genetic characteristics found to be crucial for longer preservation of kidney graft
February 16, 2019 - Researchers use optogenetic tool to control, visualize receptor signals in neural cells
February 16, 2019 - New reversible antiplatelet therapy could reduce risk of blood clots, prevent cancer metastasis
February 16, 2019 - Testosterone is not the only hormone needed for penis development
February 16, 2019 - FDA Advisory Committee Recommends Approval of Spravato (esketamine) Nasal Spray for Adults with Treatment-Resistant Depression
February 15, 2019 - Heart surgery technology developed at Baptist Health debuts after years of secrecy
February 15, 2019 - Prescription Opioids Double Risk of Triggering Fatal Car Crash
February 15, 2019 - New study helps doctors better understand high blood pressure in pregnant women
February 15, 2019 - Beta wave control in Parkinson’s diseased brain could be a potential therapy
February 15, 2019 - Media representations of love may justify gender-based violence in young people
February 15, 2019 - Yoga May Help With Rheumatoid Arthritis Symptoms, Severity
February 15, 2019 - Obstructive sleep apnea linked to inflammation, organ dysfunction
February 15, 2019 - Master your mind: A challenge from WELL for Life
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