Healthcare Meets Big Data

By Gali Weinreb
Globes, Tel Aviv, Israel

WWR Article Summary (tl;dr) How to use big data to improve patient care? That is just one of several issues tackled at this years “Reboot Forum 2017.” Below is a list of some of the promising medical startups.

Globes, Tel Aviv, Israel

How can the healthcare system be rebooted?

Last month, ten projects — commercial startups, initiatives from within the healthcare system, and emerging ideas of private individuals, participated in the Leading Healthcare Initiative of Reboot Forum 2017, a multidisciplinary forum comprising diverse representatives from the healthcare system, who teamed up to brainstorm on ways to improve the system and ensure that it will continue to be economically sustainable.

It was clear from the finalists that data is the sought-after good in the healthcare system. Half of the ventures making presentations were commercial ventures that use big data analytics capabilities to improve patient care. Two other ventures are also engaged in data, but not with big data algorithms; and three ventures were classic medical device companies.

The judges’ discussion was lively; less because of disagreement over the quality of the ventures, and more because of the dilemma over what is the best way to kick-start change in the system — whether by supporting new ventures for which every shekel is important, or by supporting more mature ventures which have a lesser need of cash but set a standard of how the different parties want to see the healthcare system function in the future. The three winners will be announced at the joint Reboot Forum and “Globes” 2017 Sustainable Healthcare Conference on 13 June.

MedAware Ltd. — preventing prescription errors
MedAware uses the capabilities of learning systems to prevent prescription errors. In the presentation to the judges, Dr. Gidi Stein explained that the technology was developed after he learned of the case of a 9-year old boy who died after the doctor prescribed, instead of the routine drug that the child should have received, the next drug in the list of the HMO computer — a powerful blood thinner for adults. The boy fell off his bicycle and died of a brain hemorrhage, “A child was killed because of a typo,” says Stein.

Stein says that current alert systems try to predict doctors’ errors and warn them, but that many errors are still missed.

MedAware’s approach is to warn of an unusual prescription. “In effect, the system learns from the doctor’s previous conduct about the reasonable range of medications prescribed for such a patient. Deviations from this reasonable range trigger an alert. It is important that we only sound an alert infrequently. Current systems sometimes send to many alerts, most of which are false alarms, resulting in doctors ignoring them. We consciously miss some errors, which is the price we pay to create a system with few false alarms, so that when a doctor gets an alert, he knows that something extraordinary has happened.”

A recent study tested the system’s ability about previously treated patients. It detected errors which were not detected by current systems, including one regrettable case of a death caused by an uncaught error.

“There are bodies lying on the road, which this company picks up,” praised judge Dr. Ran Balicer

MedAware is already operating in the US. At the event, Stein said that if the company wins the prize, it would use the money to install the system in Israel. “It is surprising to discover that it is harder to install such a system in Israel than in the US,” he said, “because of the health funds’ long sales cycle.”

A Drop for Research — aiding the next discovery
What would you say to the next proposal: sign a general consent form, after which every time you take a blood, urine, feces, or other test, you share part of the sample with a biological sample bank, which will save the data about you, although it will still be under the responsibility of your health fund. Academic and industry researchers will be able to submit questionnaires to this database, which will make possible the discovery of a great deal of new information.

Would you give consent? Are you worried about privacy? Ask yourselves what do you have to gain personally? Prof. Varda Shalev, director of the Big Data Research Institute at Maccabi Health Services, who initiated the current venture, explains, “We realized that our customers have a strong will to make change. That is why a person who gives such a sample can see exactly what happened with his sample: which studies he has participated in, and what happened in these studies.”

There is also a less altruistic incentive: every sample will be sent for basic genetic testing, the results of which will be sent to the customer. In this way, every time something is learned from the sample which might be relevant to that health of the patient who donated it, he will be notified.

A survey by Maccabi among its customers found that 50% said that they were willing to participate in the venture.

Asked about why Maccabi had entered the competition and requesting the fairly negligible prize money, compared with Maccabi’s budget, Shalev said that it was mainly seeking recognition and support of the Reboot Forum in order to get the venture through the regulatory obstacles.

Datos — preventing test errors
Many companies are developing algorithms to analyze medical data, but for good quality results of this analysis, the data input must be good quality. The measurement of medical data can frequently be erroneous or biased. A patient did not correctly place the blood pressure monitor on his arm, or a child decided to have fun monitoring his blood pressure or that of the family cat, or the measurement was taken in a very hot or cold room, affecting the sensors, and so on and so forth.

Datos detects unlikely measurements and clears them to obtain a correct snapshot of the patient’s health. The task is complicated, because the unusual reading is the one we are seeking as it indicates a change in the patient’s condition. Is it possible to separate erroneous data from correct but unusual data? Datos believes that it has the tools to do this.

In addition to clearing errors, Datos’s system also enables the integration of data from different instruments and provides an easy-to-use interface to read and analyze the data.

Valera Health Ltd. — detecting mental deterioration
Like other companies founded in recent years, Valera Health is seeking to detect deterioration in a patient’s condition by analyzing his day-to-day behavior. Valera Health, which specializes in mental health, does this without any external equipment, using only data from a mobile phone.

Today, a mental health patient will usually encounter the healthcare system when he is in the middle of some kind of episode.

Alternatively, he will go to prescheduled meeting with a therapist, once every few months. During the intervals between meetings, he is definitely liable to face a serious deterioration in his condition. A patient’s decline into a psychotic, manic, or depressed episode causes great suffering that is liable to result in suicide, and it is not certain that the patient can recover. If there is early intervention, it may be possible to halt or mitigate the episode through appropriate medication of psychological treatment.

Valera Health’s product tracks symptoms such as the speed of the patient’s gestures, whether he leaves the house, his sleep patterns, frequency of his conversations, and the strength of his voice when talking. These metrics are integrated with the patient’s characteristic habits, and when he deviates from them is a suspicious manner, the system opens a chatbot for him, which asks him how he is. If the patient says that his condition has worsened and it is not a mistake, there will be human intervention.

Related News

Leave a Reply

Your email address will not be published. Required fields are marked *