AI in the fight against cancer: How Parloa is guiding the UKSH toward the future of patient care

The University Medical Center Schleswig-Holstein (UKSH) is one of the largest university hospitals in Europe. It has cutting-edge medicine and research at its locations in Kiel and Lübeck and is home to the Cancer Center (CCC), which treats cancer patients from all over Germany. The UKSH already uses artificial intelligence in various areas, including its telephone systems, and is now the first hospital in Europe to use artificial intelligence in clinical trial support and patient communication.

Vision

About 1.8 million people in Germany are living with cancer. Only about 5% of them participate in clinical trials. Despite trials offering access to modern treatment options, not every patient is medically eligible, geographic distance can be restricting and some patients feel hesitation towards the novelty of them. One of the most limiting factors to moving trials forward, however, is the sheer quantity of information teams need to sift through in order to coordinate the studies. Study protocols run to hundreds of pages, inclusion and exclusion criteria are highly complex, and medical teams hardly have the capacity to review every study amidst their day-to-day work. There’s also the personal side to manage, as the topics being discussed around the survey are emotionally overwhelming, and language translation can present more obstacles.

The UKSH sought a pilot with Parloa to accomplish their vision: Doctors should be able to focus on providing medical care while AI agents take on the coordination and administrative tasks that currently burden them,

“Together with Parloa, we want to show how AI can support the medical team – not replace it. Our goal is to find new ways to make innovative therapies accessible to more patients and to provide them with better support throughout complex treatments.” 

Prof. Dr. med. Anne Letsch, Director of the Oncology Center at the Kiel Campus, Member of the Board of the University Cancer Center SH (UCCSH)

How AI agents support the medical team 

The ongoing pilot project tests the use of Parloa’s AI agents in three primary areas:

1. Smart clinical trial search

Parloa’s AI agents search a curated database of oncology clinical trials and identify potentially suitable studies based on the information provided, achieving a match accuracy of up to 80%. While the agents accelerate the process of finding clinical trials, medical evaluation and decisions remain the responsibility of the human team. In this use case, the key technical consideration continues to be how to bridge the gap between the medical jargon of the study database and the everyday language of patients. 

2. Telephone follow-up interviews in the context of clinical trials

The AI agents call patients who are already participating in a clinical trial and conduct structured interviews in accordance with the study protocol. They reliably collect standardized data, relieving the human study team of repetitive data collection tasks. Important to note: The agents do not conduct informed consent discussions or provide medical assessments, they just collect the data. 

3. Round-the-clock administrative relief

AI agents handle routine administrative communication around appointment scheduling, medication reminders, and medical record inquiries 24/7. In doing so, medical staff aren’t overburdened and can prioritize direct patient care.

Data protection and clinical compliance

For hospitals, data protection is a fundamental prerequisite to any technology usage. The UKSH has correspondingly high standards, which Parloa fully meets:

• Data storage within the clinical infrastructure: Patient data never leaves the clinical IT infrastructure. Processing takes place within the UKSH’s existing system landscape, without any transfer to external clouds or third-party systems.

• Access control and segregation of duties: Access to sensitive patient data is strictly limited to authorized user roles. AI agents operate only with the data points that have been approved for the respective task.

• Compliance with regulatory requirements: The platform is designed to meet the requirements of the GDPR as well as hospital-specific data protection guidelines. As part of the pilot project, data protection impact assessments and clinical approval processes are being conducted in collaboration with the UKSH data protection team.

• No training with patient data: Conversations and patient information are not used to improve AI models outside the agreed-upon framework.

Why Parloa?

The UKSH, in collaboration with PwC, evaluated several voice AI providers. Ultimately, they selected Parloa for a few key reasons: 

• Market-leading voice AI: Parloa’s voice-first platform performs in multilingual conversations and is reliable in emotionally sensitive situations. 

• Real-time processing of complex clinical trial data: The platform not only automatically analyzes extensive clinical trial protocols and criteria but also presents the findings to the medical team in a structured format.

• Proactive patient engagement: AI agents can make outbound calls, communicate information clearly, and answer questions.

• Privacy and compliance by design: The architecture is designed from the ground up to meet clinical privacy requirements.

“We want to use AI to ease the burden on the medical team and make it easier for more patients to access the right treatment. With Parloa, we’ve found the right partner for this.” 

Prof. Dr. med. Anne Letsch, Director of the Oncology Center at the Kiel Campus, Board Member of the University Cancer Center SH (UCCSH)

PwC is supporting the project with strategic implementation and integration into existing hospital and clinical trial processes.

What's next

The pilot project at the UKSH is the first step to testing how AI agents can assist the clinical team with coordination, communication, and administrative tasks. 

Once proven successful, the voice AI agents will expand to other medical specialties. What is being tested in oncology today could serve as a model for the broader integration of AI into clinical care. Throughout the innovation, the focus will always remain the same: the AI agents support organizational processes, while all medical decisions remain with the clinical team.