Connow CEO Guo Hui: The application of large data in medical applications

Connow CEO Guo Hui in "2013 Zhongguancun Big Data Day" to do keynote speech, analysis of Connor Cloud, and large data in medical applications.

Guo Hui: Hello everyone, I'd like to talk to you today about the problem of using data to change health. Generally referred to health is to think of medical care, only when the disease will think of medical, there are diseases will go to the hospital examination for treatment. The medical diagnosis system is a black-and-white diagnostic model. When you are sick, you are not sick when you are over. But in life, most people exist in this state, people from health to unhealthy is a long-term process, this process is not today healthy tomorrow is not healthy, not sudden transformation, is the quantitative process.

Before speaking, I can give you a look at the public service ads, different people in the http://www.aliyun.com/zixun/aggregation/18178.html "> Life experience has a different state, you can also be a healthy life, Your life is not healthy, this situation is not a sudden formation, but accumulated over the day. Why not find out ahead of time and change in advance?

When it comes to using large data for health management, we first talk about where the data differences that we understand in large data and tradition are, in fact, any phenomenon, or any thing that is described with data, will be flawed. You can only describe it from a latitude or a few latitudes. The simplest example to say, we can see the children play the ball, colorful, if color-blind people say to pick out the red ball, he was puzzled.

The big data is adding new latitude, can find more causality, new technology enables us to find the area which cannot cover, lets us discover the new phenomenon analysis.

My blood pressure is usually 130, 140, low pressure will be 90, 100, also can see my BIM index will also exceed the range. This data will tell me the risks of certain diseases in the future, why I have not changed, do health management, in addition to our large data to find out the risk, how to change it. This is also the problem we have to discuss, do health management, in the final analysis is not to see technology, but to see how to use technology to bring the ability to fight against human nature. Man is lazy, man is physician.

We know that people are going to die through data analysis, and this conclusion doesn't help us, how we make big data produce the ability to drive behavior.

What we do is make the early warning by large data, here I particularly emphasize that our collection and traditional equipment acquisition is the difference is that we are more concerned about two consecutive words, because the traditional medical diagnosis is black and white mode, more than a range is unhealthy, not over the scope of the normal state. We are more concerned with the laws of data volatility, predicting future risks through the law of volatility.

The current situation of hypertension in China, the latest data should be 330 million, hypertensive people, why there is such a high proportion of the population. Many people have no way to know their physical condition in advance before going to the hospital to check. We think about how to make these people adjust when they start to show signs. These are chronic data, including high blood pressure, diabetes, and so on, and all chronic diseases are the human body's biological clock being destroyed.

We work with American laboratories, they study, chronic disease is because you destroy the biological clock, the meal time does not eat, the exercise does not exercise, the sleep does not sleep. If one wants to be healthy, he will open his legs and pipe his mouth. Let people put off their laziness to do things.

We have three parts of the core, the first analysis model of the core, we from single point data analysis to continuous analysis.

Second, single point collection and analysis, pay more attention to equipment data acquisition capability, and continuity.

The third part of the problem in the business model, we can gradually break away from the traditional mode of profit by selling hardware, users do not need data, the need is my body now, the future health there is no risk, how can I change, this change is effective, in fact, the user needs is to be able to effectively improve the service, and are willing to pay for this service, in the future we will expect the gradual conversion from hardware to become a profitable project.

It will affect you later. Data analysis and judgment, and we through to the continuous extraction fluctuation rule analysis, the single point anomaly can pass enough data, eliminates the influence which the single point data produces.

I am here to tell a case, this case is a friend of mine, is also an entrepreneur, 34 years old, have a family history of hypertension, he is very young have high blood pressure. The doctor says you are so young that you need to do a whole day's monitoring. Male high pressure normal fluctuation range, in the hospital to do ambulatory blood pressure monitoring, you exceed the normal percentage to determine whether to get high blood pressure, and then prescribe antihypertensive drugs. We can see several aspects of the information through this data, the first aspect, the traditional user, he was at night after 10 o'clock blood pressure drops, and my friend night blood pressure has not dropped. According to our data pattern, this type of kidney failure is 8 times times that of ordinary people.

If he's going to use exercise to relieve his blood pressure, at two o'clock in the afternoon, the high-pressure exercise is his own injury, through contrast and analysis, to find the traditional process can not find opportunities and methods.

In fact, in the United States, in Europe, in Japan, our collaborating laboratories are doing similar research. Can get better results, when the choice of treatment, which is to put your physical signs and time together to find a way to update.

We cooperated with it, they from the 60 's, over millions of human body databases and analytic models. We can see that this is when the human's this continuous blood pressure, we see the normal wave curve, when the curve appears different, the peak is not the same. For example, his risk of stroke, the risk of coronary heart disease, the risk of sudden death, see his risk is ordinary number of times. There must be signs before any disease, and early warning of disease before it is formed.

Here's a case where new things happen when the data adds new latitude. Astronauts, we have experiments in the analysis, put forward research topics, want to understand the rapid changes in the environment on the impact of human signs. In the country to collect the corresponding data, to see if he had local testers at that time, the results found two. A major earthquake in Wenchuan, in Chengdu, China doing biological research, the second is the 2011 Tokyo earthquake. Because there is laboratory cooperation in Tokyo.

In the three days before the quake, all the testers had a consistent change in blood pressure and were not restored to the traditional level until three o'clock the quake. What does that mean? People are reflective of the external environment. Man is a superior animal, but he cannot feel it. This timeline takes the earthquake day as the timeline.

This is the correspondence between our study of maternal blood pressure data and miscarriage in the United States. Let's look at the innovations on the device, the blood pressure data, the BIM data, the device because its data exists on a single computer. So for users, can only focus on the accuracy of data monitoring. After users know the data to learn what the meaning of this data, learn how to change this data.

So in recent years, smart hardware concept, in foreign countries come out a batch of new hardware. Make an address analyzer. This data helps traditional users to record and transfer problems and can give data to your authorized family physician. In the European and American ordinary family, the family doctor is more common, they transmit the data is very troublesome. With new hardware, these transmissions are convenient.

People know that helping you analyze data is more important to help you do regular testing. This kind of hardware, to the domestic situation, because the domestic consulting resources are relatively inadequate, in fact, the hardware can play a relatively weak role. What we are doing, first I will do three products, one is blood pressure management equipment. For middle-aged and elderly families. Athletic Watch. We will think more about the three kinds of equipment, how we regularly collect data, how to let users provide data, this is the most important thing we consider. If you just let the user get the data, there is no difference with the traditional hardware, the data transfer out, no monitoring analysis, you do not do anything in the traditional hardware to increase the data transmission module. Each of our design concepts discusses how to keep users on the go.

Many people have used Java and VB (sound), most users are difficult to more than three months, many times people find more and more meaningless. Every time the equipment changes, how to let users continue to use, how to make him more willing to use. At the same time we do in addition to the hardware part, we also do the corresponding content, we will ask the content manufacturing team. When the data comes in I tell you how to change, how to adjust, I hope that through the device data to motivate you to continue to reform. People in the tradition realize that I should go jogging today, play basketball tomorrow and not see the results. Through data and interaction, we are moving forward through data positive shifts to drive continuous improvement.

The part of the business model is actually looking at how we look at the hardware, we look at the hardware as an entry point to the data, and all the sensors in the future are going to keep the data flowing. is how long and how often the user is in the time.

In the tradition we see when data is generated primarily on computers, we are in the Internet era when mobile phones produce data in mobile internet age. In the Internet of things hardware value and positioning, unlike traditional hardware, traditional times hardware if you have this problem I come to you to solve, in the future hardware I want to capture your state, through the data analysis to know your needs, the things you need to directly push to you. I think we need to explore whether the hardware can make a profit from a single sale.

In our planning, we hope to be able to leave the field of hardware manufacturing after 5 years, because for us, the real core value lies in our analysis of continuous data, we hope many hardware vendors back-end data analysis and service vendors. And in the future, the whole industry chain is becoming clearer, from data collection, data storage and distribution. In this industry chain which one of the least competition, which will become a commercial upstream. I also hope that many people to discuss with me, I also hope to have more exchanges, so thank you.

(Responsible editor: The good of the Legacy)