About GH Diaconesses Croix Saint-Simon

A leading healthcare provider in Eastern Paris, the Diaconesses Croix Saint-Simon Hospital Group (GH DCSS) employs 1,200 professionals and treats nearly 300,000 patients each year. With a comprehensive care offering, the hospital consistently ranks among France’s top institutions – particularly in surgery, oncology, women’s health, palliative care, and geriatric medicine.

Committed to both high-quality care and digital innovation, GH DCSS is actively embracing artificial intelligence to help achieve clinical and operational excellence.

Opportunity

Since France’s 2004 reform introducing activity-based costing (T2A), accurate and comprehensive medical coding has become critical to hospital financing. Properly coded hospital stays ensure fair reimbursement based on the actual resources used to care for patients.

But medical coding is both complex and time-consuming. It’s typically handled manually by the hospital’s Medical Information Department (DIM). Medical information technicians (TIMs) manage large volumes of diverse medical stays, some of which span over 50 pages. Even the most experienced teams – with well-established processes – can miss codes or run out of time, making optimization audits a valuable step toward ensuring completeness and accuracy.

Solution

To tackle these challenges, GH DCSS partnered with Parallel for a next-generation coding optimization project.

Parallel deployed both a consultant and its proprietary generative AI platform, purpose-built for PMSI (Programme de Médicalisation des Systèmes d’Information) medical coding. This solution:

• Identifies high-impact stays likely to be reviewed and optimized;

• Reads the entire documentation of a medical stay and understands it globally;

• Suggests missed or more precise diagnoses that better reflect the care provided;

• Justifies each suggestion with a reference from the official PMSI guidelines, along with a supporting excerpt from the patient’s file.

All AI-generated suggestions were reviewed by the Parallel consultant and then validated by GH DCSS’ CMIO. This dual review process – AI and expert – enabled a high-volume, high-accuracy optimization.

Impact

Key results from the project

• €1,900 – average uplift per optimized stay

• 15% – optimization rate on targeted records

• 75% – of proposed corrections accepted by the DIM

  
  

"Parallel doesn’t replace medical experts — it enhances their work by offering well-justified, compliant coding suggestions."

Dr. Quentin Jarrion, Chief Medical Officer at Parallel

Thanks to this collaboration, GH DCSS was able to secure significant revenue gains in just a few weeks, while identifying strategic opportunities for ongoing support from generative AI in the coding process.

"Partnering with Parallel for this optimization project gave us the opportunity to test a generative AI solution within our hospital environment — with no added burden on our IT team and no financial risk."

Dr. Pierre Hertz, Chief Medical Information Officer, GH DCSS

Looking Ahead: Reducing Administrative Burden Through Generative AI

Beyond the immediate financial impact, this project illustrates the broader potential of generative AI to transform administrative processes in healthcare.

Parallel’s long-term ambition is to become a strategic partner for healthcare organizations, helping them transition to more efficient, data-driven care by reducing administrative workloads through the power of generative AI.