The first obstacle to the establishment of these warehouses mentioned by hospital teams is the cultural blocking that may exist around the sharing and reuse of health data.
"It was not obvious just a few years ago to put in place governance at the 39 hospitals of AP-HP and set the rules for sharing and accessing the data collected. as part of the care ", testified Elisa Salamanca, director of the department web innovation and data of the Public Assistance-hospitals of Paris (AP-HP).
The head of the data clinic at the University Hospital of Nantes, Pierre-Antoine Gourraud, linked these obstacles to the question of "the confidence to create vis-à-vis patients, caregivers and colleagues". "One can very quickly fall into a perception of the warehouse that would be a matter of copying practices," he said, pointing to the importance of having "above all" a "strong governance".
Still on the human side, he said that the success and use of these new IT infrastructures require "a lot of pedagogy and repetition". He has already presented three times to the Medical Establishment Commission (CME) warehouse, in addition to being asked by the medical services to talk to staff. "The best way to arouse interest is to show concrete examples," he said.
Another "key to success" cited by Nicolas Delaporte and Prosper Burq, respectively director of IT services (DSI) and big data project manager of Toulouse University Hospital, is "the recruitment of a specialized team" for the warehouse, as well as that "the cooperation between the various clinical and non-clinical services" of the establishment.
At the University Hospital of Lille, Drs Grégoire Ficheur and Vincent Sobanski, co-directors of the Include project, also lamented "the continuing fantasies about health data". "The sensitivity of access to health data can be a brake, it raises concerns on the part of patients and CHU teams who may fear that their activity is scrutinized."
To get around the problem, they insisted on the need to integrate the EDS "in the project of the establishment" and to work together various profiles. "In the context of an EDS, no one can claim to have a transversal knowledge of the subject and it is necessary to form a mixed team that meets all the needs," they explained.
To help increase usage, Elisa Salamanca explained that she wanted to "rely on some medical personalities who could highlight the interest of EDS in their specialty" at AP-HP.
A point of view shared by Professor Pierre Champsaur, radiologist, chairman of the commission "information system and DPI" at the Public Assistance-Marseille hospitals (AP-HM). In charge of the development of the Marseille EDS, he pointed out the "need to convince the interest of the project internally", calling for "working in close collaboration with a multidisciplinary team (doctors, engineers, developers, computer scientists, etc.)." .
He also stressed the interest of integrating "progressively and within the framework of specific projects" industrialists, academics, researchers, and doctoral students specialized in artificial intelligence (AI).
If it is accepted by all the health institutions interviewed by TICsanté that expert teams need to be formed, the new profiles of data scientistsDevelopers or managers capable of supervising technical teams also push CHUs to thoroughly review their recruitment processes.
Few people currently have the skills to manipulate these massive volumes of data and they are, in addition, expensive profiles, prized by the lucrative private. In this context, UHCs face a lack of attractiveness of the public sector.
"The implementation of an EDS is expensive, it involves significant investments and recruitment of expensive profiles like the data scientists. To recruit for these positions, we must count significant salaries and we face competition from banks or Gafam (Google, Apple, Facebook, Amazon and Microsoft, ed) who pay better, "lamented Drs Sobanski and Ficheur. University Hospital Lille.
The same tone at the University Hospital of Nancy where Jean-Christophe Calvo -chef of the territorial department of digital transformation and biomedical engineering of the institution- regretted "the recruitment difficulties" faced by recruiters in the public sector.
"It takes experts to qualify the data, but they are rare profiles, often expensive and they are not necessarily interested in the public," he noted.
While the issue of recruiting "expert profiles" is central to the technical challenges of the DHS, the main challenge is the continuous addition of hospital information system data in the warehouse, which is characterized by their high degree of sophistication. heterogeneity.
"As the DPI (computerized patient file) is unique to the AP-HP, as some software, including historical information systems, are widely dispersed, which requires significant work of standardization and data qualification", a Elisa Salamanca, taking the example of the 23 existing databases for chemotherapy.
A sentiment shared by Jean-Christophe Calvo, in Nancy. "The success of a warehouse depends on the quality of the data it contains and its relevance because it will also support the development of artificial intelligence (AI) and we must take this turn to ensure the success of the project", he explained.
In Nantes, "the regulation is much more complicated", according to Pierre-Antoine Gourraud. He referred to the filing of an application for authorization to the National Commission for Computing and Liberties (Cnil) to initiate data processing, but also the redesign of the circuit of information of the patient on the use which is information in accordance with the European General Data Protection Regulation (GDPR).
In the East and in the West, where the institutions have initiated joint reflections, the teams interviewed by TICsanté have put forward the interest of these regional and inter-regional dynamics to answer the various technical and regulatory challenges. posed by these new infrastructures.