Editor's guide: With the continuous development of modern medicine, driven by new technologies such as "Internet +" and artificial intelligence, artificial intelligence medical treatment has emerged and benefited the vast number of doctors and patients. But at the same time, artificial intelligence medical care is also facing the dilemma of commercialization.
In order to allow readers to quickly understand the artificial intelligence medical industry, there is a concept of artificial intelligence medical treatment. So at the beginning of the article, I compiled a panoramic map of AI in the medical field. The content of the panoramic map includes application scenarios, application values, related companies, policy analysis, technical status, and so on.
The current version is 1.0. If there are any omissions or better suggestions, you can scan the QR code of the panoramic map to contact the editor-in-chief, and work together to improve it.
1. Value
Most of the current applications of AI in the medical field are in the medical imaging industry, and the most intuitive value it brings is a tenfold or even a hundredfold improvement in efficiency and accuracy.
Take the medical imaging of lung nodules as an example. The greatest value is that when the workload of doctors is heavy, it is difficult to find lung nodules below 3 mm. With the aid of AI, at least 80% can be saved. Repetitive labor, the original doctors may spend a lot of time in screening, but now only the doctor is required to check once, and it can effectively reduce the misdiagnosis rate.
AI medical care will account for one-fifth of the market in the artificial intelligence industry. A large part of it is because medical care has a wide range of market needs and diversified business trends, and has a diversified business space.
First of all, the data source is more comprehensive, because it is connected to hospital desensitization data, and the user group is relatively clear. General imaging doctors can start quickly, but for image analysis, they generally only provide expert assistance, and will not directly give medical results. The final decision-making action still needs the doctor to operate.
In the medical industry, there is also a particularly great value, that is, its data has strong scalability. Basically, deep learning can be done well, and then it can be extended to other application scenarios. In fact, this can cleverly solve the late stage of start-ups. The problem of business peaking.
In essence, artificial intelligence is the use of technology to solve medical problems. The effective clinical medical data is structured, and then deep learning is carried out. The corresponding scheme model is found in the same medical case and the standard is given. There are also many successful cases of this application.
In fact, there are many AI medical scenarios in our daily lives, such as a hospital guiding robot, which can conduct diagnosis and triage according to your painful part, such as disease analysis in search engines, and solution path matching and search. , Q&A, and customer service are all examples of artificial intelligence applications.
There is no data-based process that cannot be solved by artificial intelligence. It is just that artificial intelligence uses high costs to reflect the value of medicine.
Regarding AI medical care, we talked about the application scenarios from the technological breakthrough, and then from the application scenarios to the user value, and then to the current landing and commercialization. In the first two years, everyone may be confused, but now most companies have compared Knowing the needs of customers, including the logic and general direction of the entire industry, the next question is how to accelerate the pace of commercialization.
Hematopoiesis has always been a big problem for artificial intelligence.
2. Dilemma and Breakthrough
In fact, whether it is knowledge graphs, image recognition, etc., which have realized the development from the previous manual assistance to more rigorous intelligence, artificial intelligence has basically taken shape in the medical field.
Taking artificial intelligence + hospital management (CDSS) as an example, CDSS is the use of AI technology to compare and link multiple laboratory test reports and examination reports when doctors treat patients, and finally give the doctor a diagnosis and treatment suggestion. This is essentially a Multivariate analysis process.
In fact, Tencent and Ali have done a lot of work in it, and they have also landed in many tertiary hospitals. However, judging from the current performance and evaluation of the hospital market, CDSS does have some problems.
It is not difficult to find, from medical imaging to hospital management to medical research. Regardless of "word of mouth", the problem of "time cost" that they jointly reflect is more worthy of attention.
In fact, AI medical treatment has always encountered the problem of "time cost", that is, the time cost of cultivating cross-industry collaborative cognition. In fact, the biggest cost comes from people’s cognition time, including the cognition between customers and enterprises, the cognition between developers and medical experts, the cognition between capital and the market, and the commonality between them requires a big deal. Run in for a while.
The biggest contradiction at present is that capital has lost patience with AI medical care. How to solve this problem has become the key to the development of the industry.
We all know that in the past few years, to some extent, it is an advancement in medical technology, but there is no way to build products or business models with only advanced academics. The core of the commercialization of capital requirements is the landing of the scene. In order to allow the AI medical "soft landing" and at the same time to explain the pace of landing clearly to the capital, segmented development is the best method.
It is precisely because artificial intelligence maintains the development of the entire industry by continuously optimizing algorithms and expanding scenarios. As the label of "burning money" of artificial intelligence becomes more and more obvious, and it is not yet mature to adapt to the market, AI medical care for capital blood transfusion Said that the next AI medical entrepreneurs will bear more pressure.
After all, the current technological maturity is not enough, and the role of pulling the medical industry that can only be brought about by the next labor will be weakened. Especially, in 2B mode.
In fact, if AI medical companies and the next entrepreneurs who want to enter AI medical can adapt to the staged landing model as soon as possible, for capital, this staged landing will become a problem. Looking forward, the development prospects that can be achieved are extremely impressive.
Commercializing the intermediate results in the research and development process and expanding the scene for a "soft landing" is not a long-term solution.
In the short term, we only need to provide good service to medical customers, such as helping hospitals to automate the medical process and greatly improving the efficiency of hospital management, such as helping private hospitals to refine their operations, or opening up existing technical interfaces to make more Many companies or institutions come to feed back the effectiveness of the technology. This is what the segmented development of AI medical care must do.
It is not that a medical image is made, and all diseases are screened in advance, which is a success.
For 2B commercialization, the steps are divided into four steps. Technical verification, single-business model experiment, single-business market verification, and copy mode are the safest methods.
From an objective point of view, whether it is commercialization or reproducibility as a key or bright spot, it is a understandable operation. However, if the market performance after landing is relatively flat, the user's evaluation may be directly transferred to the immature technology, which will consume word-of-mouth. This can actually be seen as a consumption of AI medical reputation.
Therefore, for how to commercialize AI medical care in the future, how to speed up the establishment of an effective data closed loop while accelerating segmented development has become the biggest key. Similarly, controlling the boundaries of the business, the pace of time, and the location of the scene are also the core of the commercialization of artificial intelligence.