Artificial Intelligence

Automating some processes can deliver diagnoses more quickly, increasing the chance of better outcomes for patients while reducing clinicians' workloads, says Kevin Shah, head of Enterprise New Business at Fujifilm Europe.

Consultation proposes that primary care services, such as Babylon’s GP at Hand, need to focus on areas with doctor shortages to reduce health inequalities.  

The money will help Atlanta University Center Consortium Data Science Initiative improve education in healthcare analytics at historically black colleges and universities.

Simer Sodhi, director of data management & analytics at Westchester Medical Center, says the organization integrated data from the EHR using a machine learning algorithm to connect patients to the right sources of care.

Google, the University of Chicago Medical Center and University of Chicago are listed as defendants in a class action suit that alleges they failed to properly de-identify sensitive patient medical data. WHY IT MATTERS The complaint, filed by Matt Dinerstein in the U.S. District Court for the Northern District Of Illinois, claims UChicago "promised in its patient admission forms that it would not disclose patients' records to third parties, like Google, for commercial purposes." Instead, the university "did not notify its patients, let alone obtain their express consent, before turning over their confidential medical records to Google for its own commercial gain," the document states. The suit alleges that "Google and the university claimed the medical records were de-identified. But that’s incredibly misleading. The records the University provided Google included detailed datestamps and copious free-text notes." Google's expertise in data mining and artificial intelligence, Dinerstein charges, means it is "uniquely able to determine the identity of almost every medical record the university released." In addition to seeking monetary compensation, the suit calls for an injunction requiring the University of Chicago to comply with all HIPAA de-identification regulations, enjoining the organization from disclosing identifiable patient medical records to third parties without first obtaining consent. It also calls for an injunction prohibiting Google from using patient records obtained from U of C and an order requiring Google to delete all patient records received from the university. Since electronic health records contain patients' highly sensitive and detailed medical records, including records revealing not only a person's height, weight and vital signs, but whether they suffer from certain diseases or have undergone a medical procedure, the University's release of EHR data would be in violation of HIPAA, Dinerstein's suit alleges.. "The personal medical information obtained by Google is the most sensitive and intimate information in an individual's life, and its unauthorized disclosure is far more damaging to an individual's privacy," the lawsuit states. THE LARGER TREND The use of de-identified patient data has been common practice for years, but hasn't been without its scrutiny.  Back in 2010, the Office of the National Coordinator for Health IT launched a study on how to manage the privacy risks of using health information that had been stripped of personal identifiers. Even as researchers and technology developers noted that such de-identified data is a must-have for population health and other purposes, ONC said it was seeking a "consensus on what risk we can tolerate for identification and then what level of removal, what kinds of removal of information, are required to get to that level of risk," then National Coordinator Dr. David Blumenthal told Congress. More recently, de-identified data has become essential to the training and development of new artificial intelligence algorithms that are impacting every corner of healthcare, including AI technologies such as Google's DeepMind, which the class action suit names as one way the company could more easily "find connections between various data points" and compromise privacy, even with de-identified data. In 2017, Healthcare IT News reported on a DeepMind initiative, Verifiable Data Audit, that was exploring a blockchain-like service that "could give mathematical assurance about what is happening with each individual piece of personal data, without possibility of falsification or omission." The goal was to give providers and patients real-time insight into where and how data is being used. "For example, an organization holding health data can’t simply decide to start carrying out research on patient records being used to provide care, or repurpose a research dataset for some other unapproved use," according to DeepMind. ON THE RECORD "We believe our healthcare research could help save lives in the future, which is why we take privacy seriously and follow all relevant rules and regulations in our handling of health data," said Google in a statement responding to the lawsuit. "In particular, we take compliance with HIPAA seriously, including in the receipt and use of the limited data set provided by the University of Chicago." Nathan Eddy is a healthcare and technology freelancer based in Berlin. Email the writer: nathaneddy@gmail.com Twitter: @dropdeaded209

Infolytx's natural language processing technology is helping with clinical trial cohorts as AI gains real traction across healthcare, explains CEO Badrul Husain.

The CEO says cash raised by the stock offering will help broaden the analytics and data capabilities it can offer its payer and provider customers – and drive innovations in AI, blockchain, robotics and more.

The most promising use case for blockchain in healthcare is the elimination of "redundant work, rework and reconciliation," says Tej Anand, professor of practice at the University of Texas' McCombs School of Business.

Buzzwords like Artificial Intelligence (AI) and machine learning are commonly heard at conferences and industry events and they often conjure up images of robots or killing machines from the Terminator. However, panelists from the Innofest Unbound conference in Singapore all felt that technologies such as AI should not replace humans as it is commonly imagined – rather, they should augment the work of clinicians and hopefully, even enhance the patients’ interactions with their doctors. Human-centric AI A medical doctor by training and also the founder of MEDGIC, a startup which utilises AI to detect skin conditions, Dr Reid Lim feels that the use of AI should always involve doctors and not replace them. “Healthcare systems are becoming unsustainable and we need AI to help automate some things and to help alleviate the burden on doctors. AI is not new and it seems strange that some people are only beginning to grasp the use AI.” “A lot of radiologists are already using Computer Aided Diagnosis (CAD) for mammography and it has been happening for some time. So the idea is for us as a tech startup to pursue what we call human-centric AI. We try to make AI as explainable as possible and we always want humans to be involved in the whole process,” he added. Dr Philip Wong, a practicing cardiologist and founder of WEB Biotechnology, concurred that as doctors, there is a sense of compassion and empathy to want to help patients regardless of their health condition, and that is something AI cannot do at the current moment. “Anything thing (or tech) that is adopted by the hospitals, we always have qualified specialists or ‘men/women in the loop’. For instance in radiology, even though a lot of analysis is done by AI, in the end the person who signs off is the radiologist or specialist radiologists who have to sign off the report.” “We’re not afraid of losing our jobs as healthcare providers, and in this respect, what I think most people or most healthcare providers don't understand is the benefit that you get from the first problem I pointed out, right? The problem is we have a whole huge plethora of data, which we're trying to analyse but we haven't got a sense of it. So you pop up the electronic health records, right, I see two computer screens with approximately 1000 fields. I've looked at this, analysed all this and this is where I think AI can really help the doctor, the patient as well,” said Dr Wong. Meaningful and novel applications of tech Mr Chua Chee Yong, Emerging Services & Capabilities Group, iHIS, the IT agency for the Health Ministry in Singapore, shared that there are possible concerns about specialists being ‘replaced’ by AI even after years of training and practice. However, he stressed that it is about how to meaningfully apply technologies so that it makes a difference. “We all know today that AI is not quite explainable, not quite there yet. What do we do? We can choose to ignore it and only use it when AI reaches a point that it is explainable. Or we can choose to apply it in a safe way that allows us to improve our productivity.” “The example in this case is retinal image scanning. We use a simple AI to do a simple classification of ‘normal’ vs ‘abnormal’. Guess what, 70% of the normal cases got eliminated by AI and the leftover 30% are read by the human being – productivity suddenly increases and the person just focuses on the very complex 30% abnormal cases. The accuracy also improves – you see how we apply it? Despite the constraints and limitations of AI, if we apply it meaningfully, we can harness value out of it,” he explained. Looking into the future According to Dr Wong, what he sees in the future is that data is becoming more ‘fluid’, for example from the Spyder wireless ECG monitor that he invented. For example, even if someone is wearing the ECG monitor in the US, the ECG data can be acquired by the doctor back in Singapore almost instantaneously.  Another future development that is very exciting for Dr Wong is that the smartphone will become the ‘replacement’ of healthcare in many respects, becoming the apparatus and interface to make an appointment to see a doctor, collect health information and do much more.  He concluded with a wish as a doctor for a ‘digital health persona’ for patients, to get a lot more health information from them and try to solve their problems. With the digital health persona, health information is being collected continuously, even outside the hospital and the information can even be front-loaded to the doctor before the patient sees him/her. This gives the doctor a broader picture of the patients’ health. 

Pre-treatment scans were input into a deep-learning model, which analyzed the scans to create an image signature that predicts treatment outcomes.