“Sandbox” for diagnoses – Newspaper Kommersant No. 241 (7442) of 12/27/2022
[ad_1]
Experts from ANO Tsifrovaya ekonomika suggested introducing the institution of data intermediaries to find a balance between protecting personal data and providing scientists and developers of artificial intelligence systems with access to medical information (about diagnoses, clinical trials, visits to medical institutions, etc.). Now most of the medical data belongs to the category of personal and cannot be processed without the consent of the subject, even in an anonymized form. The use of the “regulatory sandbox” mode in this area, according to ANO experts, will provide developers with access to data sets in a secure loop and at the same time protect patients.
The creation of solutions based on artificial intelligence technology in healthcare will require the settlement of many legal issues related to access to medical data and the procedure for their processing, experts of the ANO Tsifrovaya ekonomika fix it in a report on this topic. Now such information belongs to the category of particularly sensitive data and cannot be processed even in an anonymized form. To solve the problem, experts suggest using an experimental legal regime (EPR, “regulatory sandbox”). It will allow the introduction of a secure yet free exchange of medical data for research purposes through the institution of data mediation. The main functions of data intermediaries are the collection, processing and depersonalization of data, as well as their exchange. Developers in this case will have access to data sets in a secure circuit, and the interests of patients will be protected.
“Between the two extremes – “data should be available to everyone” and “data should not be available to anyone” – there are compromise options for the treatment of anonymized data through data intermediaries,” the report says. Now researchers and developers cannot receive medical data from the Unified State Health Information System (EGISZ) for the formation of analytics and the development of systems based on AI. ANO experts talk about the need to lift the ban on the personal data operator (PD) from using the collected data for other purposes with the consent of the PD subject. The development is hindered by the need to obtain written consent to the processing of personal data of patients (information about the state of health, results of clinical trials, diagnoses, etc.) – even impersonal ones.
It should be noted that the Ministry of Economy has already developed a draft amendment to the government decree of 2020 on innovations, in respect of which the EPR regime can be established. The Office proposed to cover them and “processing of data, including personal data, and activities related to such processing.” As the department explained to Kommersant, this “will allow businesses to test solutions, including those involving the commercial use of big data technologies” (see Kommersant dated November 22).
The All-Russian Union of Patients calls excessive administrative barriers the main factor preventing the introduction of software using AI. “It is believed that the implementation of such solutions is associated with a high level of danger. Increased barriers to registration are set even if the software is used to diagnose or treat a disease with a low risk to public health and generates reference information that does not indicate the need for immediate action,” said Yury Zhulev, co-chairman of the union. According to him, this slows down the development of AI-based solutions in medicine, including in medical decision support systems, despite the fact that they are aimed only at providing background information to a specialist who makes his own decision.
[ad_2]
Source link
