share

BUSINESS

3min read

Machine Learning in Healthcare—Interview with Alan Warren

Machine learning and IoT was a big theme at Web Summit 2018. To get some first-hand insights, we met with the CTO of Oscar Health Alan Warren and talked about their product and the role machine learning plays in healthcare and other industries.

Oscar Health is a technology-driven health insurance company that “with more coverage and less hassle” strives to give their members the most accessible healthcare and best experience in the US. Learn more about their awesome product here.

How does machine learning help Oscar Health users to access the best possible healthcare?

There’s a huge data quality issue in the healthcare industry. Typically 40% of all claims coming into an insurer have some error in either the number of information or the content of information. Some human has to actually go look at it and crack it before it continues processing. We built our models and machinery in such a way that when a claim doesn’t process it runs through our model, which corrects all the information and repeats the process again. We’re also using AI around voice and language recognition, content from phone calls, instant messages, keyword detection, labeling categorization etc.

Abnormal is very nuanced and machine learning models are very good at picking up nuance in large volumes of data.
Alan WarrenCTO at Oscar Health

Is user experience as important as the technology behind the product?

This is a really interesting question. In the tech world that I come from, you build a product, some app or system, you give it to people and it runs by itself. Then you watch how it works and look at your users’ feedback so you can make a better version of the product. But the machine is pure machine—it runs all by itself.

The insurance world is not like that. We are looking at claims, modifying information and updating state, we have care teams that are looking at member’s information etc… So the insurance world has this human-machine mix, where a machine is running under the hood but with all these people wrapped around it. The product that we actually expose to our end users is a machine learning app that looks like the ones they use every day; that does step-tracking, helps looking up doctors, communicating with them and so forth. That’s the digital part of our product. On the other side it also includes concierge teams, a personalized group of care guides and nurses assigned to each member to help guide their health care journey, plus the whole user-doctor interaction. Our challenge is to make all those elements play together nicely as a coherent product.

Source: Oscar Health press kit

Apart from the use of machine learning in healthcare industry, what other areas will be significantly influenced by this technology?

There’s huge work to be done around large volumes of data. When you have a huge amount of data, extracting the needles from a haystack is very hard to do for a human. Machine learning models are very good at looking at large health records and singling out those that look suspicious to find out what’s different about them. Another example is that machine learning systems work well in security, by looking at network traffic and being able to monitor the things that are no longer behaving normally. ‘Abnormal’ is very nuanced and machine learning models are very good at picking up nuance in large volumes of data.

Oscar Health in 2030—how do you envision it? What’s the dream?

My dream is we’re running the best insurance company in the world and the United States and that people are using our technology to provide health insurance machinery in the world. The dream is that our technology and data-centric approach is used to provide a better and less expensive healthcare experience for everyone, around the globe.

share


UlaCommunication Specialist

POLIDEA NEWSLETTER

Sign in and expect sharp insights, recommendations, ebooks and fascinating project stories delivered to your inbox

The controller of the personal data that you are about to provide in the above form will be Polidea sp. z o.o. with its registered office in Warsaw at ul. Przeskok 2, 00-032 Warsaw, KRS number: 0000330954, tel.: 0048 795 536 436, email: hello@polidea.com (“Polidea”). We will process your personal data based on our legitimate interest and/or your consent. Providing your personal data is not obligatory, but necessary for Polidea to respond to you in relation to your question and/or request. If you gave us consent to call you on the telephone, you may revoke the consent at any time by contacting Polidea via telephone or email. You can find detailed information about the processing of your personal data in relation to the above contact form, including your rights relating to the processing, HERE.

Data controller:

The controller of your personal data is Polidea sp. z o.o. with its registered office in Warsaw at ul. Przeskok 2, 00-032 Warsaw, KRS number: 0000330954, tel.: [0048795536436], email: [hello@polidea.com] (“Polidea”)

Purpose and legal bases for processing:

 

Used abbreviations:

GDPR – Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016
on the protection of natural persons with regard to the processing of personal data and on the free movement
of such data, and repealing Directive 95/46/EC (General Data Protection Regulation)

ARES – Polish Act on Rendering Electronic Services dated 18 July 2002

TL – Polish Telecommunications Law dated 16 July 2004

1)        sending to the given email address a newsletter including information on Polidea’s new projects, products, services, organised events and/or general insights from the mobile app business world |art. 6.1 a) GDPR, art. 10.2 ARES and art. 172.1 TL (upon your consent)

Personal data:name, email address

2)       statistical, analytical and reporting purposes |art. 6. 1 f) GDPR (based on legitimate interests pursued by Polidea, consisting in analysing the way our services are used and adjusting them to our clients’ needs, as well as developing new services)

Personal data:name, email address

Withdrawal of consent:

You may withdraw your consent to process your personal data at any time.

Withdrawal of the consent is possible solely in the scope of processing performed based on the consent. Polidea is authorised to process your personal data after you withdraw your consent if it has another legal basis for the processing, for the purposes covered by that legal basis.

Categories of recipients:

Your personal data may be shared with:

1)       authorised employees and/or contractors of Polidea

2)       persons or entities providing particular services to Polidea (accounting, legal, IT, marketing and advertising services) – in the scope required for those persons or entities to provide those services to Polidea

 

Retention period:

1)       For the purpose of sending newsletter to the given email address – for as long as the relevant consent is not withdrawn

2)       For statistical, analytical and reporting purposes – for as long as the relevant consent is not withdrawn

Your rights:

 

Used abbreviation:

GDPR – Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016
on the protection of natural persons with regard to the processing of personal data and on the free movement
of such data, and repealing Directive 95/46/EC (General Data Protection Regulation)

According to GDPR, you have the following rights relating to the processing of your personal data, exercised by contacting Polidea via [e-mail, phone].

1)       to access to your personal data (art. 15 GDPR) by requesting sharing and/or sending a copy of all your personal data processed by Polidea

2)       to request rectification of inaccurate personal data
(art. 16 GDPR) by indicating the data requiring rectification

3)       to request erasure of your persona data (art. 17 GDPR); Polidea has the rights to refuse erasing the personal data in specific circumstances provided by law

4)       to request restriction of processing of your personal data (art. 18 GDPR) by indicating the data which should be restricted

5)       to move your personal data (art. 20 GDPR) by requesting preparation and transfer by Polidea of the personal data that you provided to Polidea to you or another controller in a structured, commonly used machine-readable format

6)       to object to processing your personal data conducted based on art. 6.1 e) or f) GDPR, on grounds relating to your particular situation (art. 21 GDPR)

7)       to lodge a complaint with a supervisory authority,
in particular in the EU member state of your habitual residence, place of work or place of the alleged infringement if you consider that the processing
of personal data relating to you infringes the GDPR
(art. 77.1 GDPR)

No obligation to provide data:

Providing your personal data is not obligatory, but necessary for Polidea to provide you the newsletter service

Refusal to provide the above data will result in inability to receive the newsletter service.

Profiling

In the process of providing the newsletter service, we make decisions in an automated way, including profiling, based on the data you provide.

 

“Profiling” means automated processing of personal data consisting of the use of your personal data to evaluate certain personal aspects relating to you, in particular to analyze or predict aspects concerning your personal preferences and interests.

 

The automated decisions are taken based on the analysis of clicked and viewed content. They affect the targeting of specific newsletter content to selected users registered to receive the newsletter service, based on the anticipated interests of the recipient.