More data is being generated than can be analysed – but enough of it is being studied to spot patterns in behaviour and genes to speed up drug discovery and development.
Patients’ illnesses could soon be diagnosed by AI – The Guardian – 12-Sep-2017
Computers could start diagnosing patients’ illnesses within the next few years as artificial intelligence increasingly ousts doctors from their traditional roles.
NHS belives that machines may soon be able to read X-rays and analyse samples of diseased tissue, such as lumps that can indicate the presence of breast cancer.
AI-powered application to assess visual aging biomarkers in lab animals – LEAF – 29-Aug-2017
Lifespan.io is launching a crowdfunding campaign to support MouseAGE.
Software will assess visual biomarkers of aging in laboratory animals.
Project should increase the pace of research on aging.
While also reducing animal suffering in experiments.
IBM and MIT power research into microbiome – Health Data – 24-Aug-2017
Will study the connection between body bacteria and autoimmune diseases.
Using supercomputing power crowdsourced via IBM’s World Community Grid.
Virtual experiments will map the 3 million bacterial genes found in the human microbiome and predict the structure of their associated proteins.
Volunteers can provide compute power by downloading a secure software program to run virtual experiments.
9 Computational Drug Discovery Startups Using AI – Nanalyze – 25-Apr-2017
Including BenevolentAI which already has 24 drug candidates.
98% of big data has only been created in the last several years.
Cloud computing becoming cheaper and cheaper.
Also helped by emergence of deep learning algorithms.
AI Versus Aging – LEAF – 18-Apr-2017
Interview with Dr. Alex Zhavoronkov, CEO of Insilico Medicine
Company awarded the ‘Most Promising Company’ title at Personalised Medicine World Conference in 2015.
In 5 years we want to build a comprehensive system to model and monitor the human health status and rapidly correct any deviations.
Main breakthroughs expected in the area of multi-modal biomarkers of aging.
“Applying AI to aging is the only way to bring it under the comprehensive medical control.”
New machine-learning algorithms may revolutionize drug discovery – Kurzweil AI – 8-Feb-2017
University of Toronto Scarborough algorithms can generate 3D structures of nanoscale protein molecules.
Proteins can’t be seen directly without using sophisticated techniques like electron cryomicroscopy (cryo-EM).
3D structure pieced together from high-resolution 2D images.
Existing techniques often generate incorrect structures unless an expert user provides an accurate guess.
Health insurer calls analysed for signs of disease in your voice – New Scientist – 6-Feb-2017
US start-up Canary Speech using deep-learning to spot neurological conditions in voices.
Training algorithms with hundreds of millions of phone calls labelled with speaker’s medical history.
Vocal cues could distinguish someone with or without a particular condition.
Stricter UK data protection laws may prevent a similar project there.
Aims to detect vocal indicators of Alzheimer’s disease within two months, then depression, stress and dyslexia.
AI that knows you’re sick before you do – ZDNet – 01-Feb-2017
IBM’s research labs are combining the company’s existing machine learning and artificial intelligence systems with newer kit.
Hyperimaging systems visible light and other parts of the electromagnetic spectrum.
Hyperimaging could help people with allergies to scan their food for traces of substances that trigger their condition.
Designing lab-on-a-chip with nanostructures to analyse samples of body fluid for potential diseases.
Creating a prototype that can help mental health professionals diagnose patients just from the content of their speech.
Deep learning algorithm beats dermatologists in diagnosis of skin cancer – Kurzweil AI – 25-Jan-2017
Stanford Artificial Intelligence Laboratory used a deep convolutional neural network (CNN) algorithm.
Performed as well or better than 21 board-certified dermatologists at diagnosing skin cancer.
Built on work by Google to differentiate cats from dogs.
Algorithm was trained with nearly 130,000 images representing over 2,000 diseases.
Artificial intelligence predicts when heart will fail – BBC – 16-Jan-17
Software learned to analyse blood tests and scans of beating hearts
Studied 256 patients with pulmonary hypertension.
Measured the movement of 30,000 different points in the organ’s structure.
Combined data with eight years of patient health records.
Correctly predicted patient one year survival about 80% of the time – better than 60% for doctors.
How big data is improving breast cancer prediction rates – Health Data Management – 5-Dec-2016
Influence score (I-score) improved the prediction rate in breast cancer data from 70 percent to 92 percent.
New method for differentiating between “noisy” and predictive variables in big data.
I-score fares especially well in high-dimensional data with many complex interactions between variables such as genetic data.
Could be valuable for analyzing other big healthcare data such as electronic health records and wearable sensors/trackers.
Big data promise exponential change in healthcare – FT – 29-Nov-2016
GlaxoSmithKline employs online technology and a data algorithm developed by F1’s elite McLaren Applied Technologies team.
Potential lies in its ability to fine tune research and clinical trials.
Helped by continued falls in data storage costs and improving computing power.
Handling the personal details of millions of people creates huge data quality, privacy and security problems.
Data gathered 10 years ago by a brain scan is infinitely less detailed than what you would get today
Big data analytics to control Hospital Acquired Infections – ETCIO – 24-Oct-2016
Asia’s largest healthcare group Apollo Hospitals says HAI or nosocomial infection is one of the biggest challenges faced by hospitals.
Most of the patients who come to the hospital for treatment, have a weak immune system.
Analytics platform used for analysis and communication of disease and infection surveillance information to both clinical and non-clinical teams.
Decision engine shown to the doctors in the initial stages to ensure that the algorithm and the processes were constantly fine-tuned.
Information, such as probability of an infection affecting a patient, is presented and doctors’ clinical judgement is greatly improved.
Herr Watson will see you now – BBC – 18-Oct-2016
IBM’s AI system will attempt to solve some complex medical cases in Germany.
Watson will be based at the Undiagnosed and Rare Diseases Centre at the University Hospital in Marburg.
Patients may have very long medical histories and have seen up to 40 different physicians.
Watson will analyse the patients’ medical files to offer a series of ranked diagnoses.
How Machine Learning, Big Data And AI Are Changing Healthcare Forever – Forbes – 23-Sep-2016
IDC predicts that 30 percent of providers will use cognitive analytics with patient data by 2018.
Computers and deep learning algorithms are getting more and more adept at recognizing patterns.
Pathway Genomics is developing a simple blood test to determine if early detection or prediction of certain cancers is possible.
CareEdit tool helps create the best course of treatment for different types of cancers.
Microsoft’s Project Hanover aiming to ‘solve’ cancer – ZDNet – 20-Sep-2016
Seeking to make precision cancer therapy available to all.
Currently, volume of research required to assess each cancer limits personalisation to the most serious cases.
US government’s PubMed service publishes two new papers each minute, or more than one million a year.
Microsoft’s knowledge extraction available through an Azure-hosted biomedical search service called Literome.
Individual projects include:
– machine learning to personalize the drug combinations used to treat acute myeloid leukemia
– helping radiologists track disease progression
– Bio Model Analyzer (BMA) helps investigate what happens to cells when disease strikes
– Biological Computation Group is programming cells in the human body rather than silicon on chips
Medical claims big data and the social determinants of health – State of Reform – 25-Aug-2016
Healthcare industry uses Big Data to identifying populations most likely to utilize its services.
Organizations presume that the absence of medical claims is the best indication of whether a person is happy and healthy.
Need to look at the Social Determinants of Health (SDOH).
Daily stressors can often manifest themselves into real physical and emotional health problems.
People who report concerns about life necessities were 5x more likely to report poor health.
Dr. Ralph Perfetto of the Eliza Corporation talks about “data fear” – knowing something that you cannot or will not act on.
AI Saves Woman’s Life By Identifying Her Disease – Futurism – 5-Aug-2016
IBM’s artificial intelligence system, Watson, compared the woman’s genetic information to 20 million clinical oncology studies.
Previously undetected illness determined to be an exceedingly rare form of leukemia.
Positive identification allowed doctors to develop a treatment for the woman, ultimately saving her life.
COMMENT: Watson identified the condition in 10 minutes – the same amount of time you get with a normal GP appointment
Cancer-patient big data can save lives if shared globally – Kurzweil AI – 23-May-2016
Hospitals, laboratories and research facilities hold huge amounts of cancer patient data.
GA4GH intends to provide a common framework for the responsible, voluntary and secure sharing of patients’ clinical and genomic data.
Could match patients from different parts of the world with suitable clinical trials and develop personalized treatments for each individual’s cancer.
Research paper in Nature Medicine highlights logistical, technical and ethical challenges.
Global Alliance for Genomics and Health (GA4GH) is involves more than 400 organizations in over 40 countries.
Machine learning rivals human skills in cancer detection – Kurzweil AI – 22-Apr-2016
Samsung Medison has updated its RS80A ultrasound imaging system for breast-lesion analysis.
Uses big data collected from breast-exam cases and recommends whether the sample is benign or malignant.
University-Purdue University found open-source machine learning tools are faster, cheaper and better than humans in detecting cancer from pathology reports.
Shaun Grannis: “A human’s time is better spent helping other humans”
How Big Data Is Transforming Medicine – Forbes – 16-Feb-2016
Healthcare professionals treat us according to evidence-based medicine.
EBM explores which treatments work best for which illnesses and in which groups of patient.
Mining practice-based clinical data (actual patient records) will provide more information on who has what condition and what treatments are working.
How Big Data Is Quietly Fighting Diseases and Illnesses – Dataconomy – 18-Jan-2016
Big Data has already offered up amazing proof of its applications
Ebola – predicting the disease’s geographical spread and sending relief organizations.
Sepsis – Amara Health Analytics created a predictive model based on data gathered from bedside monitors.
Polio – Vaccine-Preventable Outbreaks chart lets users explore data based on year or disease type.
Using Big Data to Make Wiser Medical Decisions – Harvard Business Review – 14-Dec-2015
The bigness of data is not its absolute size, but the task of transforming it into wisdom.
Beth Israel Deaconess Medical Center (BIDMC) creates applications that lead to wise clinical decisions:
BIDMC@Home – blood pressure data BP data gathered telemetrically at home – yields clinical and financial benefits.
I2B2 – open-source tool – query large databases across institutions to determine most effective medications and surgery.
Screening Sheets – continuous data analysis of electronic health record and alerts clinicians when to take action. Will soon incorporate big data from the patient’s genome.
Big data’s big issues – PHG Foundation – 23-Nov-2015
Curation and integration
Sharing and access
Engagement and trust
Which Vendors Lead the Healthcare Big Data Analytics Market? – Health IT Analytics – 19-Nov-2015
In contrast to the EHR marketplace, where providers are more or less limited to a set list of certified products, healthcare organizations have a little more leeway when it comes to choosing big data analytics technologies to support more advanced health IT capabilities.
Epic Systems leading both markets.
Big Data Is Driving Personalized Medicine Revolution – Information Week – 12-Nov-2015
Drugs can be expensive, difficult to research, hard to get approved.
Most drugs don’t work on large parts of the population.
6 ways big data is helping:
Precision Medicine Initiative Cohort Program
National Cancer Institute-Molecular Analysis For Therapy Choice (NCI-Match) Trial
Wisdom Study – are mammograms the best way to detect breast cancer
Tanner Project – looking for N’s-of-1
Stand Up To Cancer And Melanoma Research Alliance Dream Team
Big data, better health – New Scientist – 4-Nov-2015
Current electronic health records will soon be dwarfed by data generated by genetic and blood-marker tests, in addition to advanced scans and activity trackers.
Main challenge is getting unstructured health data into a meaningful form so that it can be analysed.
5 per cent of people with chronic diseases account for around 50% of healthcare spending.
Trial equipped 135 patients with telehealth monitoring technology at home – enabling a mobile intensive care team to visit patients’ homes and intervene at the first sign of trouble.
Predictive, algorithm-fuelled approach reduced hospitalisations by 45 per cent, care costs by 27 per cent and acute response costs by 32 per cent.
Big data’s early detection of life-threatening infection – Computer World – 20-Oct-2015
Penn Medicine’s massively parallel computer cluster uses a huge volume of data to build prototypes of new care pathways.
Pathways then tested with patients and results feed back into data set.
Clinicians can predict patients at risk of sepsis, a life-threatening complication, 24 hours earlier than before.
AAN calls for social and behavioral data in EHRs – Health IT Analytics – 14-Sep-2015
American Academy of Nursing says electronic health records are incomplete and potentially insufficient if missing social and behavioral data.
Less than half of accountable care organizations ACOs have access to behavioral health records.
Problems range from data standardization, inconsistent workflows and patient engagement.
Hospital computer predicts death with 96% accuracy – BBC News – 14-Sep-2015
Super computer at Beth Israel Deaconess Medical Center, Boston, MA, collects and analyzes patient data every 3 minutes. Uses artificial intelligence to compare medical history and current medication with data from 250,000 patients over 30 years.
Enables AI to recognise rare diseases quickly.
Also predicts death in next 30 days with 96% accuracy. Raises ethical questions on whether those patients will be provided with the same level of resources.
Improvement in diabetic patients from big data – Modern Healthcare – 22-Aug-2015
11% improvement in a comprehensive measure of diabetes management.
Mercy Health identifies patients who have not improved despite clinical intervention.
Explorys software allows customers to analyze de-identified data from their own and other systems to benchmark their performance and search for insights about other treatment models.
70m pharmacy customers targeted by IBM Watson and CVS – Washington Post – 30-Jul-2015
IBM Watson artificial intelligence system to identify physiological indicators and red-flag behaviors of chronic conditions such as heart disease and obesity. 1 in 5 Americans could consult with Watson in a pharmacy kiosk – alerting your doctor of any contraindications.
Predictive analytics model for readmissions case study – HealthITAnalytics – 20-Jul-2015
Advocate Health Care, a twelve-hospital health system based in Illinois, wanted to reduce unplanned readmissions which it considered a failure in transitions of care. Wanting to identify patients at highest risk for readmissions whilst still in the hospital Advocate Health Care implemented data-driven accountable care.
Major challenge was interoperability of a mix of EHR technologies used across its sites which needed to be collated to enable big data analytics of clinical data. Now their predictive analytics model for readmissions is updated every two hours during the patient’s stay.
Big data analytics aims to stop hospital killer – Networks Asia – 20-Jul-2015
Hundreds of thousands of people die each year in the US from sepsis – a condition where the immune system over responds to an infection triggering widespread, and often fatal, inflammation.
It’s difficult to diagnose as symptoms are similar to those of feverm and often occur after a patient has been discharged from hospital.
Hitachi Consulting is working with Vital Connect’s HealthPatch (a disposable wireless biosensor) and analytics specialist ClearStory Data to monitor symptoms in real-time. A solution that could save costs as well as lives.
Machine learning to design personalised care plans – OnWindows – 14-Jul-2015
Dartmouth-Hitchcock Health System is piloting a healthcare solution called ImagineCare built on Microsoft’s new Cortana Analytics Suite.
Information from sensors and devices (for example, blood-pressure, pulse oximeter, activity trackers) is monitored and nurse alerted if personal thresholds exceeded. Initial phase to start in October with 6,000 patients.
70 year study tracks lifestyle and disease – BBC News – 13-Jul-2015
5000 people selected for study in 1946, over 2000 still taking part. Project able to track affects of upbringing and lifestyle on disease and longevity.
Big data and health predictions: new study on correlations among medical problems – MIT News – 7-Jul-2015
MIT researchers analysed anonymous data from over 500,000 patients. Able to compare individual patient’s medical history with other people to predict diseases.
How DNA sequencing is transforming the hunt for new drugs – Reuters – 13-May-2015
Pharmaceuticals companies analysings hundreds of thousands of DNA samples to identify new gene targets. Previously took decades to find and study enough people with rare traits.