DPR, a smart tool to facilitate the work of primary care social service professionals
Institut Municipal de Serveis Socials de l’Ajuntament de Barcelona (Barcelona City Council Municipal Institute of Social Services)
Artificial Intelligence tool to support primary care professionals for social services. Analyzes cases and advises the professional workers to respond to the attended people. The primary care worker who serves a visit, after transcribing on the computer his notes about the most important person in the interview, pressing a button obtains in a few seconds from the DPR system (Requires, Problems, Resources in catalan) a recommendation on the response to give or the action to be undertaken.
DPR processes the annotations of the social worker, classifies the problem and the person’s request, and suggests to the professional the answer to give, according to the resources and services the institution can offer or advise. This “intelligent” system allows social professionals to respond more quickly, more safely and objectively to the user, and users to receive a clear diagnosis with high accuracy reliability.
This tool that automatizes the process of typifying the demands received by social service centers has been trained using machine learning techniques from the data of 300,000 interviews conducted by the social services of Barcelona City Council in previous years.
The DPR system is thus a tool of “collective intelligence”, which does not replace the decisions of social professionals, but provides them with guidance based on the collective experience of previous years and leaves the final decision on the response to each case or situation in their hands.
It is a pioneer tool in Spain which, in addition to supporting decision-making and providing guidance to professionals, makes it possible to improve the planning of resources and services of the administration.
Barcelona City Council
The creation of this tool arose in the Barcelona City Council with the aim of reducing the hours of primary care professionals for social services intended for management, and thus increasing the time devoted to social care. Through artificial intelligence, the machine read anonymized conversations and comments from the hundreds of social workers in Barcelona in previous years, and because many interviews are repetitive and problems resemble, through machine learning techniques it trained to classify situations, demands and possible responses, until it became able to suggest in each case the best response from what it had learned.
Level of implementation
Barcelona City Council first conducted a pilot test of the DPR system at social service centres in the neighborhoods of the Verneda, Sant Gervasi and Casc Antic. And in view of the good results achieved, the realization of the city’s other social services centers is peaceful. The goal is to streamline the first 50,000 interviews that municipal social services perform each year on the most diverse issues.
Currently, work is also being done on new versions of this tool that allow social professionals to write interview reports through voice and automatically become text, as well as using the DPR to help detect hidden problems, such as situations of male violence.