The National Health Insurance Fund (Cnam) launched a new site on June 20: Data pathologies. It provides the general public, professionals and public decision-makers with data on the management of French pathologies, analyzed each year between 2015 and 2020 in the form of medical expenditure mapping.
The technological solution was provided by Opendatasoft, a French company specializing in data visualization and sharing.
Make data readable
A major producer of health data, this is not the first time that the Cnam has started using open data. Each year, it publishes a report in which “it carries out a number of analyzes and makes recommendations to improve care management“, details Damien Vergé, director of strategy, studies and statistics of the Cnam, at L’Usine Digitale. All of this information is based on data published on its site. “On the other hand, the content of the data was not very well valued. A number were published as PDFs, which is not very interactive“, he adds.
Hence the idea of creating a dedicated platform to allow everyone to access health insurance data and understand them simply. It offers an approach by pathology (psychotropic treatment, diabetes, cancers, maternity, etc.) and by territory for the 68 million policyholders in France. A tab also contains “source data“that can be downloaded and reused in the form of an API, notes Damien Vergé.
Algorithms to identify pathologies
The work carried out by the Cnam teams therefore aims to “make the healthcare bills speak”, the anonymized information of which goes back to the National Health Data System (SNDS). Initially, pathologies are identified using algorithms taking into account long-term conditions, diagnoses coded during hospital stays for the pricing of the stay, specific drugs for the treatment of certain pathologies and sometimes other information.
From this work, 57 categories were established, for example to identify in the context of cancer, 4 types (breast, prostate, lung and colon), each also differentiated according to whether it is in the active phase of treatment or under surveillance. These groups of pathologies are grouped into 17 broad categories. In a second step, expenditure is allocated to pathologies, except for expenditure in hospital establishments which is allocated directly by taking into account the reasons for stays.
A pool of knowledge for start-ups
Although they can obviously access it, this site does not primarily target start-ups specializing in health, says Damien Vergé. “They can use it as framework data according to their fields, for example the monitoring of certain chronic pathologies such as diabetes, he adds. It is more of an aid than a tool to help them progress in their projects.“