The context
Since the 2004 reform, activity-based payment (T2A) has been the main way French healthcare facilities are funded. Medical coding values each stay according to the resources used to care for the patient, so that the hospital is paid as fairly as possible. Producing complete coding on time, as close as possible to the care actually delivered, is therefore a strategic priority for any hospital.
Yet coding is a complex and time-consuming task. It is done manually by the Medical Information Department (DIM), and results can vary. Medical information technicians (TIM) handle large numbers of stays of many different kinds, and complex records sometimes run to more than 50 pages. Even an expert team with the best processes will naturally miss some codes or run short of time on some stays. That is why an optimization audit makes sense.
The solution
The DCSS Hospital Group called on Parallel for a new kind of optimization service, carried out under a Lamda* on records from 2024.
Parallel provided the hospital with a consultant equipped with its proprietary generative AI solution built for PMSI coding. Parallel’s AI:
- identifies the records most worth optimizing;
- reads the entire patient record and understands the stay as a whole;
- suggests diagnoses that were missed during routine coding, or more precise diagnoses that better reflect the hospital’s activity;
- justifies each diagnosis with the relevant rule from the PMSI guide and an excerpt from the patient record.
Parallel’s consultant then reviewed and approved the AI’s suggestions, before final validation by Dr Pierre Hertz, DIM physician at the DCSS Hospital Group, and his team. Thanks to this double review by AI and consultant, Parallel was able to process a large number of records with a high level of relevance.
Results
- €1,900 average revaluation proposed per record
- 15% optimization rate on targeted records
- 75% of revaluations accepted by the DCSS Hospital Group’s DIM team
« Working with Parallel on this type of optimization project showed us the benefits of a generative AI solution in our hospital, with no workload for our IT team and no financial risk. »
Dr Hertz, Head of the DIM, DCSS Hospital Group
See how Parallel’s generative AI can secure and increase your coding revenue across your entire hospital.
*A Lamda (from LAMDA, Logiciel d’Aide à la Mise à jour des Données d’Activité, the software used to update activity data) is a procedure that lets healthcare facilities make retroactive changes to activity data they have already sent to the ARS (Regional Health Agency).
