close
AI

Fast AI meets slow medical treatment: sinking, stretching, diving into the deep sea

94c81020-d494-11e9-ae89-aaaa00151f43.jpg

Editor’s Note: This article comes from WeChat public account “CV Intelligence” (ID: CVAI2019), author Yu Yangyang, editor Zhang Lijuan, 36氪 authorized to publish.The wind of AI medical care has been blowing for three years.After the vents, everyone mentioned that they are getting together images, landing hard, financing is getting less and the profit is still not in place.A more difficult problem is that B-side gold is not easy to dig, and medical business is harder to do.“It’s really hard,” Cheng Guohua, the founder of Jianpei Technology, exclaimed to CV’s intelligence.At the beginning of the venture, Cheng Guohua took the server and set foot on the train from Hangzhou to Xinjiang. He stayed with the local doctor for a month, and barely obtained the trust of the first hospital customer.Many AI medical company founders and investors also confessed to CV: AI medical is still in a very early stage of development, and domestic AI companies have experienced two products from storytelling to gradual product arrival.Three years; technology is a slow process for the upgrading of hospitals, and it takes no more than ten years.Behind the “fastness” of the AI ​​medical company, it is a “slow practice” compared to the traditional medical industry.To do AI medical care, you have to open the door of a hospital in one family. It is not easy to open a hospital door because of many reasons such as information islands.A safe and healthy person told CV that “a hospital usually has more than 200 different systems, and some hospitals will even reach 1,000. There are so many systems in a hospital. What do you do?” Long-term attention to the medical fieldSun Qi, the founding managing partner of Dow Investment, also told CV that the medical market does not have the high degree of openness of the Internet. The transformation of the Internet into traditional industries can be said to be devastating, but in contrast, the medical transformation is relativelyslow.The story of speed and slow combination is obviously not so good.The innovator’s game AI Medical Wind started three years ago.2016 is a watershed. Due to capital concerns, Imagine Technology, Senyi Intelligent, and Shenrui Medical, a group of AI medical startups that have been admitted to the market have sprung up after this year.Sun Qi recalled to CV intellectuals that the first time I saw Chaucer was in the Spring Festival of 2017, which coincided with the biggest snow in Beijing in ten years.At that time, Chaucer was the vice president of Siemens China Medical Division Greater China, and two other co-founders of Shen Rui, Lei Ming and Li Yiming, chatted directly with Sun Qi in the hotel lobby for an hour and a half. Several people reached a consensus.As a Taoist investment that has invested in a number of AI medical projects including Shen Rui Medical, Landing Medicine, and Brain Doctor, Sun Qi is optimistic about the strength of the Shen Rui team, but has also worried that Shen Rui’s admission time is too late.Hesitated before he made up his mind to invest.”There are some tracks. If you say that you are a year and a half late and two rounds later, you basically have no chance to catch up. Unless the people in front took the time to detour, they were always thinking about it.But in the end, it has not given up, and the industry competition pattern is still far from being formed.” Synchronized with the feng shui, everyone is gradually deepening their understanding of the industry.Jianpei Technology Cheng Guohua also said to CV Intelligence that the process of starting his own business is to insist that he can’t find a customer from 12 years of business, and he has done related activities for cultivating the market for 15 years until the first half of the 16th year of AlphaGo’s Go game.It has been cultivated.Ma Handong, vice president and co-founder of Senyi Intelligent, told CV that AI has a lot to do in the medical industry, and the application is not only to judge on image assist, but for example, the pioneer of the University of Science and Technology, which may account for the current AI medical NLP market.Most of them, but in the AI ​​medical market, the University of Science and Technology News may not even account for 1%.What followed was that in addition to the surge in rivals on the entrepreneurial track, startups like Shen Rui also faced competition from traditional medical giants and Internet giants.Compared with startup companies, giants often have strong financial strength or rich experience and resources in the medical field.CV Intelligence interviewed a number of industry insiders and learned that in the current AI medical field, the main players can be divided into three categories.The first category is the AI ​​startup company represented by Yitu Technology, Pushu Technology, Shenrui Medical, Jianpei Technology, Senyi Intelligent, etc. The first category is Internet companies represented by Tencent Yingying, Ali Health, and Ping An Health.Another category is the traditional medical device provider represented by Siemens.Based on huge data and traffic, Ping An, who was founded in 2014, has set a time earlier than most AI medical companies that have entered the market, and has built a relatively complete Internet medical product system.Techniques such as consultations are part of the icing on the cake.A person close to health and safety told CV that Ping An had nearly 2.1 billion insurance customers to drain it. The amount of data was diverted through Ping An Insurance. Even the safe and healthy accounting management was controlled by the group.Mingzhe is the real trader behind him.Since Ping An Health has data and traffic, Siemens has equipment and systems. The giants have no advantage and then make a icing on the AI ​​auxiliary diagnostic module. Where is the opportunity for the startup?In this regard, many AI medical field entrepreneurs and investors told CV intellectuals that hardware manufacturers are not their main rivals.Most hardware equipment manufacturers do not choose to do their own AI modules. Because they are not core strategies, they will not invest heavily. On the contrary, many hardware manufacturers will even outsource these services to AI companies.Before joining the AI ​​medical business, Cheng Guohua also did internet education.“It’s definitely better to bring your own resources,” he said. Being free from the constraints of past experience and resources is precisely the biggest advantage of a startup. “After breaking into a new field, we will be brave and there will be some more innovations.The idea is not subject to experience and rebellion.” “A lot of new fields, your original practitioner, he may not be able to go out of the way, because he has a lot of experience, I think this is not good, that can’t.” Cheng GuohuaIt is believed that traditional companies do AI medical care, and the risks are more obvious. The startups look more in the future, and they are more fierce. “I think these subtle differences sometimes lead to a big difference in strength.Whether the team has strong and strong willpower, some innovative ideas will start to do.” Most of the respondents agree with Cheng Guohua, and the giants may be trapped in the consideration of risk and input-output ratio, and evenThe outsourcing of AI medical-related business to startups is still a game of innovators and brave people.B-side business is difficult to do, medical treatment is more difficult to do in 2012, Cheng Guohua founded the AI ​​medical company Jianpei Technology with medical imaging technology as the core at the West Lake.Unlike most startups that have entered the market after 2016 and are standing on the cusp, Cheng Guohua has entered the field of AI medical care for seven years.”There is nothing that can be learned from it. We need a lot of willpower. We have some ideas to start with, and we are not limited by some past experience. This is different from other companies.” The starter is often better than the latter.Taking advantage of the competition is true at any entrepreneurial track, but starting also means taking more detours, stepping on more pits, and even paving the way for the latecomers with the pits they stepped on.Cheng Guohua told CV that because it was too early to enter the field, there was no experience to learn from. This became the biggest difficulty in the entrepreneurial process, but it also made more possibilities.In the first year of entrepreneurship, Cheng Guohua led the team to make a set of algorithmic models to aid diagnosis. The most difficult problem was the failure to find customers.”How can we not find a customer? We have to ask for people everywhere. We can’t find a country nearby. We can always find it. Later, the first customer we found was in Xinjiang.” Cheng Guohua recalled finding the first hospital customer.experience.Cheng Guohua’s first client, the Chabuchael People’s Hospital in Xinjiang, is a grassroots hospital on the national border, located in the Chabuchar Autonomous County of Ili, a city bordering Kazakhstan.Cheng Guohua, on the other hand, took the server and set foot on the train from Hangzhou to Urumqi.”Luggage is too heavy, I can’t sit on the plane. I am alone on the server and I took the train. I sat in the Urumqi for a few days. It was also to save money. I didn’t bring a lot of people. I didn’t have any income at the beginning. If I can let the teamAfter living for a month, there is hope for survival.” After arriving at the hospital, Cheng Guohua personally debugged and held a remote meeting with the team in Hangzhou.Afraid to disturb the doctor’s work, he slept in the temporary bed set up by the hospital during the day and debugged at night.For more than a month, Cheng Guohua lived in the hospital, had dinner with the doctor and chatted, and no doctor in the entire department did not know.“Really, like an engineer, the customer doesn’t know that I am the CEO of a company. I know that I am an engineer and call me Chenggong.” In addition to promoting the company’s products, doctors need to help design a small program. What equipment does the hospital have?To repair, they also called Cheng Guohua to help.”Write a schedule program, write a management tool for spot checks, etc., and do a lot of tools for the imaging department to really solve their pain points. It can sell money, and the environment of my AI auxiliary diagnosis system is also set up.The iterative development of the product has begun to shun.” Cheng Guohua told CV that the B-side business is difficult to do, and the medical field is more laborious. It took a lot of effort to gently tear open a floor in the hospital.The AI ​​medical business is difficult to do, and the same situation occurs on the C side.On the line for three years, the number of users of Ping An Good Doctor is still less than 30,000.”Because medical treatment is a low-frequency behavior, the disease does not happen often, and it is difficult for users to form a habit of using our APP. This problem has not been solved online until now. We can only go offline. We are now trying one by one.Push.” ​​Ping An health people told CV intellectuals, because the line has been unable to run through, they even began to try to increase the number of users by offline “push”.The person told CV that the safe doctor has not yet achieved profitability, but he believes that this window will not be too long, maybe three years, maybe five years.Xue Suzhen, the vice president of technology marketing, said that in the process of landing, there are indeed some factors that affect the landing, including doctors’ awareness of AI, information security issues and so on.However, he believes that these problems are not the fundamental problems that hinder the widespread application of AI. Through in-depth communication and communication with the hospital, they can eventually be solved.“Different doctors have different perceptions and acceptances of AI. Some doctors may think that AI will replace doctors to generate distrust and even rejection. When AI imperceptibly relieves doctors’ daily high-pressure work and improves efficiencyIt is also very good to report quality control, doctors’ recognition of partners will be greatly changed.” Imagine also developed information security and data security solutions, “patient data transmitted to our AI system is desensitized to ensureThe patient’s privacy is not leaked; the server is also deployed in the hospital’s intranet, preventing external attacks through multiple security policies such as firewalls.” “In short, such problems can be solved through research and development, industry development and doctors’ awareness.For the landing, Imagine a set of views in the process of exploration for several years.Xue Suzhen told CV intellectuals that the current product development thinking is “one horizontal and one vertical”. “Longitudinal” refers to medical quality control, patient follow-up, prognosis and health management in the whole process of patient and disease diagnosis and treatment.”Horizontal” is a horizontal extension of different diseases such as malignant tumors, cardiovascular diseases, cerebrovascular diseases, chronic diseases and the like.For the planning of the profit stage and the profit mode, Xue Suzhen told CV that it is not convenient to disclose.Senyi Intelligent has taken a more pleasing road.Ma Handong said that all AI imaging companies are now looking forward to the official certificate, but before this, Senyi has already started to provide products based on AI core to the hospital, and is taking the way of software sales, which is currently in the preliminary exploration stage..If the market is big, business is difficult to do, and if you don’t talk about profit, you should do it first.Sun Qi said bluntly to CV intellectuals that the hospital’s To B market is more dispersed, and the hospitals have their own autonomy. Not all hospitals rely on a certain platform, but they have their own self-selectivity.In addition, the maturity of AI medical products takes time, and the market advancement also needs the market. Only when it becomes later will become the To C market, and the speed will accelerate.”Jingdong has been working for 15 years, and gradually established its own barriers. The technology must completely change the industry. The process is very long. After all, this is not the same as O2O, which will bring about the continuous evolution of the entire industry chain, but relatively,In this long-term track, you don’t need to worry about the lack of enthusiasm, you can understand the eagerness of profit, but it is still a long-term development to form several top-level industry enterprises.” The opportunity is sinking, AI medical into the township Beijing Longfu Hospital, self-serviceThe machine and the self-service registration machine fill the outpatient hall in a row.Even so, the flow of people who come to pick up the film is still endless. People occasionally even line up in front of a row of machines, the scene is lively, comparable to the airport to pick up tickets.The same situation occurred in the well-known top three hospitals in Beijing such as Concord, Beijing Medical Third Hospital, and Beijing Medical Sixth Hospital. A staff member of the North Hospital and the Third Hospital introduced to CV Intelligence that “the CT appointment queue takes half a month to one month.Time, but the film is very fast, for example, the respiratory department may only take a day or two. If you want the doctor to read the film, you have two choices. One is to reserve a fixed reading time, and the other is to go directly after the film is taken.Looking for a doctor, the film may not come out, but the image will be transmitted to the computer. The top three hospitals are still full of people every day, but self-service machines equipped with self-service tablets and reports are helpful to greatly shorten the overall situation.The medical treatment process reduces unnecessary waiting time.Not only the hall, but also the various departments of the hospital have been filled with the smart devices and systems of various AI companies.In the past two years, various departments of the North Hospital of the Third Hospital have been filled with more and more intelligent devices from AI Medical, such as Ping An Medical Insurance, Tencent Yingying, and Shen Rui Medical. Doctors use the AI-assisted diagnostic system provided by these companies.To help read the film.The next morning after getting the film, Liu Wei came to the hospital.”I have more than 50 people waiting in front of me. There are too many people, and I have wasted all my time.” Liu Wei complained to CV intellectuals. “Now the doctor uses auxiliary diagnostic equipment and reads a film and operates it on the computer.”I thought it was faster, but the time for us to wait for the result was even longer.” Most of the patients who visited the clinic told CV that there is no obvious feeling that the artificial intelligence technology has brought about the change in the whole medical treatment process. On the contrary, at the doctor.In the reading session, the time required for the film is delayed because the computer-aided reading is longer.But they also expressed their optimism about technology to help improve the medical experience. In Liu Wei’s view, intelligence is the trend of this era. A series of processes that need to be operated offline have been simplified and transferred to the Internet.There is a lot of trouble left.The problem is, AI medical company has a million hospitals, which one to choose?Senyi Smart did not hesitate to choose the top three hospitals.”Up to now, our customers are basically large-scale top three hospitals. Except for some strategic-level hospitals, all of them are done in the form of sales. Generally speaking, we do not do grassroots hospitals. At least dozens of top three are currently pushed down.The hospital’s effect is not bad.” Senyi intelligent customers are basically large-scale top three hospitals.There is still a small gap between the top three hospitals and the grassroots hospitals.Ma Handong said: “In the top three hospitals, the products are polished and seamlessly along. This is actually not so simple. The data of the top three hospitals is applicable in the primary hospitals. How to cooperate with the township hospitals is far more difficult than imagination.”But more entrepreneurs and investors tell CV that “the real opportunity for AI medical care is not in the top three, but at the grassroots level.” In the view of investor Sun Qi, AI medical must begin at the grassroots level at the beginning, “at the grassroots level.”The market volume of government procurement and two cancer screening is still very large.” “Early research must be done in cooperation with the top three hospitals. You must take the most reliable and best data and let the best doctors come.Give you data annotations, so that the products you take out can be covered and touched by the grassroots hospitals.” Sun Qi told CV that in the process of landing, the top three and grassroots hospitals showed different roles because of the disparity in strength.In most cases, the top three hospitals are responsible for research, and the grassroots hospitals are advancing.For a period of time, Jianpei has been mainly engaged in tertiary hospitals. However, after a period of time in large hospitals, Cheng Guohua found that the dependence of large hospitals on AI technology is not so strong. First of all, the staff of the hospital itself is adequate, notThe lack of doctor resources, followed by the comprehensive performance requirements of AI products is also relatively high.”These cases add up. It is actually more difficult to be a big hospital. We will turn our attention to primary care.” Cheng Guohua told CV that “the big hospitals are still working together, but some attempts at commercialization areThe grassroots found the application.” “The most basic level, for example, township hospitals, community service centers, X-rays out of the lack of doctors to write reports, often this report will be remotely let the people’s hospitals in this county come out.With the aid of AI, this gap can be well compensated.We use the big data analysis of images to produce a report that not only reduces the burden of a lifetime, but also solves the imbalance of medical resources in primary medical institutions. This is a good application.”The business of a single primary hospital is definitely relatively small, so we must connect it.””By connecting the grassroots level with the big hospitals, Cheng Guohua often gets a list of one area.” We basically do it one by one, the artificial intelligence assists the diagnostic diagnosis cloud, and also docks the company’s image.Cloud, which means that grassroots medical institutions scattered around the country can be connected first through the image cloud, and then connected to the diagnostic cloud for remote diagnosis, forming a complete, systematic medical association.body.”The pattern of the grassroots area has even allowed Jianpei to achieve some meager profit.” We started to have income gradually around 2015. We can earn money slowly in 16 or 17 years, achieving breakeven and even starting to haveSome are profitable.”Ping An Good Doctor also put his eyes on the grassroots level. “The community health center is a safe one-minute clinic. The service fee for this hardware device is only 53,000 yuan for three years. The cost of one day is less than 10 yuan.Greatly reduce the medical expenses of the grassroots.”You are not facing a doctor. A monitor is enough. A minute of clinics will open up the data of the testing equipment in the community clinic. You can see your medical data through the remote device doctor, and then give you some rational advice.”On the other hand, this is also aimed at community doctors. In the short term, he has no way to reach the top three levels, but with the aid of data, he can do something.In the view of people in terms of health and safety, AI medical floor-to-floor, in addition to making money, is also a good thing for the benefit of the country and the people. This coincides with Chen Kuan, the founder of the technology. “Many people are curious.”Why, in many industries, it is assumed that technology has chosen a technology-intensive industry with strict thresholds for medical supervision.Medical care is an industry that is closely related to each of us. Any advancement in the medical industry is of great significance to people’s livelihood.”Thinking technology founder and CEO Chen Kuan told CV intellectuals that he wants to use AI technology to embrace the medical industry and better support the lives of ordinary people. However, the founders of AI Medical want to use technology to reduce the difficulty of medical treatment for the people, butThe time required for change may be longer than they had expected. It is difficult for many entrepreneurs and investors to start with the fact that due to restrictions in systems, policies, information islands, etc., AI medical process is slow and requiresThe time is even more than ten years, but the market needs to be strong and painful. Although the vents have been up for three years, the industry is still in its infancy, patient, and suffocating, so that it can wait for the sun.It’s difficult at the beginning. You can get through this channel at first. It’s possible to find a product that may be slower. When there are three models, it’s really fast. It’s faster, but the premise is that you have toI believe there will be these changes, right?As an investor, Sun Qi believes that in time, AI medical can be rolled up like a snowball. “As with e-commerce, how can you sell a car and sell a house as soon as you come up?”All transformations are made from shallow to deep, and AI is the same.”The capacity of the market is still very large, but it is said that the process of landing is relatively slow.Cheng Guohua believes that the difficulty of landing does not hinder the existence of capacity, just needs and pain points in this market. To what extent can AI transform the medical industry, no entrepreneurs and investors can give a positive statement, but there areOne point is the same: go to the hospital first, first get things done. After all, everything is difficult at the beginning. A less friendly signal is that the enthusiasm of the capital market for AI medical care has begun to fade this year. CVSource hits the data displayIn the AI ​​medical field, the investment event began to decline in 2017, and it began to decline in 2018. In the first half of this year, it was even less than 50, and it gradually fell back to the level that was equal to the heat three years ago.Without a period of time, AI Healthcare needs to explore a sustainable road to commercialization. Image source: Pexels.