How to see the use of big data to do disease diagnosis this matter_Data Analyst Exam

The cause of this is a friend's internship with a friend to play more than an hour's phone call BLABLA a variety of aspiring mobile medical, claiming that big data can change the status quo of medical treatment to lead mankind to a new era (error

And we have a dinner date results in my stay a person ate half a day resentment

And we have a dinner date results in my stay a person eat half a day resentment

What is the reason for this? p>

--------------------------- again the dividing line ---------------------------------------

I just want to give an example is Dr,Saint SYR.

He is a general practitioner from the United States, has been in Beijing He's a general practitioner from the U.S. who's been working in Beijing to educate people about PM2.5, air pollution, etc. (lighting candles and smoking indoors makes a lot of PM2.5!).

And he went through quite a struggle, thinking about whether to return to the United States. Things changed again in February of this year when his condition progressed and it was discovered that it was actually bacterial pneumonia! After antibiotics everything cleared up!

Tell this story because:

1, this is a very very interesting gossip

2, many people feel that the previous diagnosis is a misdiagnosis, including interviews with his media are so reported, but I really think it is not a misdiagnosis

3, the disease in different stages of the symptoms and signs of the body is really not necessarily typical, how to let the data diagnosis does not become How can the data diagnosis not become a misdiagnosis by the doctors?

Those multi-center retrospective evidence-based medicine experiments, but also to deal with a variety of data, the other side so many years can not do the data, why so many people so confident?

First of all, let's make the point: I think automated diagnostics is the future, but it's immature now, and there's a lot of room for development.

The first thing to point out is that big data diagnosis, not simply collecting data to draw statistical conclusions, but there is a certain artificial intelligence algorithms in which the role of inference. One of the simple and effective, and most consistent with the logic of human judgment algorithm is called Bayesian network, in enough data (this point is difficult is it) under the premise, can be completely more accurate than any individual person to make a judgment. In the case of insufficient data, there are also comparable to a number of individuals inference ability, at least in the diagnosis of rare diseases, its accuracy is much higher than the human. There are such diagnostic aids abroad, which are aimed at the rare disease diagnosis market. Domestic currently completely blank.

Let's take a look at the current clinical diagnosis and treatment.

1, modern medicine is evidence-based medicine (EBM). That is, clinical practice is all supported by basic science research and the results of large-scale clinical trials. These conclusions, are the result of massive data collection and analysis.

2. The clinical practice of modern medicine is still in an experience-based stage, and the conclusions of EBM do not directly and completely cover the specific situations that occur in actual patients. Based on the conclusions of the basic part of EBM, combined with practical experience, is still the most important way of clinical practice at this stage.

3. A large number of medical workers who are not well educated and lack the means of examination are still the main force of medical treatment, although the best educated doctors and the best examination means of the country's tertiary hospitals are still overcrowded.

4, the role of general practitioners is seriously underestimated, and a large number of specialists play the role of general practitioners, and spend a lot of energy in dealing with some of the "minor diseases".

To summarize: EBM guidance is limited, EBM+experienced medicine is the mainstream, poor conditions, low level, professional doctors are the mainstream.

What problems can be solved by big data:

1. Expand the scope of EBM. If data can be accurately captured, the share of EBM in healthcare will rise more quickly and the overall quality of care will improve.

2, personal experience is irrelevant, big data will make the personal experience with more into the whole human experience, misdiagnosis, missed diagnosis will be greatly reduced, thus improving the overall quality of health care.

3, hospital division of labor, the division of labor will be more clear: large hospitals to solve the responsible condition, hospitals to solve the general condition, small hospitals to solve chronic diseases in the vaccination health care. Because the doctor's diagnosis is no longer dependent on personal experience, so that the accuracy of common diseases and rare diseases can be guaranteed; only complex conditions, complex treatment means, the need to establish MDT (multidisciplinary team) patients, only need to deal with large hospitals and experts.

4. The workload of all doctors will be reduced to a certain extent, which will improve the quality of medical services for patients.

What big data can't do at this stage:

1. Accurate, automated data collection. There is no doubt that the same sample goes to different hospital laboratories at the same time to do laboratory tests, the results will be different, which is already a very high degree of automation (this problem can actually be solved by inter-laboratory calibration). Not to mention the wearable devices, there are just too few of them that can reach the clinical reference level. And the dimensionality of medical data is exceptionally high - how do you get big data to automate the processing of a patient's CT data? And medical history, physical examination and other descriptive information, more inseparable from the collection of clinical workers. In short, data collection, there is no way to leave the front-line clinical workers.

2, To cure sometimes, to relieve often, to comfort alway. --E. L. Trudeau. The diseases that can really be cured are really few and far between (in fact, most of them are not cured, but just the body itself). More often than not, all doctors are doing is easing the pain and soothing the soul. This part of the work, big data can help on the help is very limited, big data at most just to alleviate other aspects of the doctor's work, so as to exchange more energy to the humanities.

The real big data can rely on the doctor, it has to be the development of artificial intelligence can be beyond most of the human time. But that doesn't mean that big data has no value at all at this stage. The value of this part is actually very huge, it's just that it's very difficult to find people who have the ability to do it, and at the same time find people who will pay for it. Just like Google's research on driverless cars, the future will definitely be completely driverless most of the time, while the current driverless technology still has a huge technical value (for example, it can avoid a lot of car accidents on the highway).

The above is what I have shared with you about how to look at the use of big data to do disease diagnosis of this matter, more information can be concerned about the Global Green Ivy to share more dry goods