As early as July, the U.S. Food and Drug Administration (FDA) issued three regulations that are important for future medical innovation .
The three specifications are:
Legal norms for low-risk public health products;
In order to support the regulatory decision-making for medical devices, provide evidence-based legal norms;
Adaptability design specifications for medical device access clinical trials.
In response, Healthblog commented on the FDA's efforts to keep up with the tide and pointed out that although the FDA has made efforts, it also needs to be regulated.
Gifsec
The promulgation of the norm can be described as the desired result. This proves from the perspective of the category of medical technology innovation in the information age. The FDA is also making every effort to ensure that the former is unfairly treated. However, these specifications also make FDA unable to respond to new things in a timely manner.
The current law does not explicitly state that FDA is based on the power of healthcare applications in artificial intelligence and machine learning. Although these regulations have reduced the administrative burden on entrepreneurs in the medical industry, in the digital era, this does not accurately stipulate the authority of the FDA as the law.
Fortunately, the new legislative work has also been carried out in an orderly manner. The US House of Representatives has passed the 21st Century Treatment Act, and the House of Lords will also reach an unanimous endorsement of the bill. The 114th Congress will soon pass a reasonable bill arrangement.
The first new specification emphasizes "Public Health Products," including "audio recordings, video games, software programs, and other products that retail stores can buy." However, the good news is that the FDA will not list these products as medical consultations, provided that these products are for the general public and there are no safety risks to consumers.
A similar product may claim that it can "help reduce the risk of a disease" or "can help alleviate some of the symptoms of a chronic disease." For a software product, a plausible slogan should be "It helps to grasp some breathing and relaxation techniques that help reduce your migraine symptoms." The value of this product derives from the information it delivers to consumers, rather than directly acting on the consumer's body.
The role of information goes far beyond this. The second new specification emphasizes the emphasis on "actual evidence-based" in research. "Evidence-based" is derived from experimental data from clinical trials. Although we do not commonly use "authentic evidence-based" methods to obtain permission for new equipment, we can use it to increase the use of equipment that has received FDA approval. Where does the actual evidence-based data come from? "These data usually come from the electronic systems that provide medical services. These medical devices (including home-type devices) collect various medical data and track the patient's treatment status."
The third new specification emphasizes the FDA's adaptive design for clinical trials of medical device access. With regard to the adaptability of clinical experimental design, it is advisable to make appropriate modifications based on experimental data without losing the integrity and validity of the experiment.
If not done properly, adaptive design may cause potential problems for the patient; if performed smoothly, adaptive design can shorten clinical trials and reduce costs. The third specification describes how to plan a reasonable clinical experiment adaptability design. If designers follow this specification, then their new equipment will be more in the morning.
This large amount of data information interacts not only with patients, doctors, equipment, drug dealers, and supervisors. These data also exist in the cloud for diagnosis and treatment decisions.
Cooperpartners
Let's take some examples. According to data from the National Human Genome Research Institute, in 2001, sequencing of the human genome cost US$100 million. By 2015, it will cost less than 1,500 US dollars. However, not all gene sequencing requires very high accuracy, and not all genomes make sense. If limited to the sequencing of protein-coding genes, the cost can be further reduced to $1,000. This has enabled visionary medical institutions to gradually join the ranks of gene sequencing. Illumina, which is engaged in the research of gene sequencing technology, has achieved an annual two-percent increase in revenue and is expected to continue.
For another example, President Obama has invested one million dollars in the establishment of personal health data for the Precision Medical Program. If things go well, patients' public service awareness will be greatly enhanced. At the same time, a huge patient population will provide scientists with unprecedented real-world data.
In the artificial intelligence and medical field, the most vivid examples are IBM's Watson and Deep Mind. Watson successfully treated a leukemia patient this month; the cooperation of Deep Mind and NHS will also be applied in the diagnosis field.
Deepmind
This is just the beginning. As Forbes's donor Todd Hixon stated, the wave of machine learning is coming. In the past five years, many large companies have acquired some start-up companies. Even so, terms such as "artificial intelligence" and "machine learning" do not appear in FDA regulations.
In order to build a better 21st century, the President and the Congress have the ability and responsibility to repair loopholes in the law and regulate the authority of the FDA. Only with the FDA standardization, patients, doctors, scientists, and entrepreneurs can be more assured and bold to develop new medical technologies and put them into practical use. Â
Recommended reading:
Watson diagnosed rare leukemia for the first time. What's difficult to diagnose in 10 minutes?
After deepening Go, Deepmind wanted to “see through†human eyes