ChatGPT for Health Data

Desislava Mihaylova founded the company for processing and analyzing Big Data, Sqilline, and a personal tragedy pushed her towards the healthcare sector. The company has successfully entered the markets of Romania, Serbia, and Croatia and is now confidently looking towards Western Europe.

You open ChatGPT, type your question, and within seconds, you already have an answer. Conducting research on an unfamiliar topic? Need to quickly gather highlights on a report? Or write a letter in a language you don’t master perfectly? No problem. In the last year or two, most of us have tried at least once some of the chatbots based on large language models.

Similarly, about a year ago, a doctor from a leading neurological hospital in Bulgaria used an AI algorithm, based on machine learning and large language models, to identify patients exhibiting specific symptoms indicative of a rare genetic disorder. The algorithm is configured to search for specific symptoms from thousands of medical documents from five major neurological hospitals over the past few years.

The platform quickly, almost as a joke, identifies one patient. The physician gets in touch immediately and after performing additional tests, it turns out that the patient does indeed suffer from a rare neurological disorder. The woman had been seeing several physicians for the last two and a half years without receiving a correct diagnosis explaining her symptoms – numbness and joint pain, altered gait, shortness of breath. She was primarily diagnosed with cardiological conditions, which did not quite explain what she was suffering from. In the meantime, the patient’s condition worsened, and she ends up using a wheelchair. After the confirmed correct diagnosis, she learns that there is also a genetic therapy that prevents the progression of this rare disease. Additionally, genetic tests have been conducted on all family members, showing that her daughter also suffers from the same rare disease, but her symptoms has not shown yet which means she can begin treatment at a very early stage.

This remarkable success story was made possible through the utilization of Danny Platform, the data analytics software developed by Sqilline.

“It’s so unique how technology is already helping patients in the real practice. For the first time in Bulgaria, with our analytic platform, we helped an entire family,” says Sqilline founder Desislava Mihaylova.

The platform was developed in 2018. Using artificial intelligence, it processes and structures data to provide real-time analysis and valuable insights, aiding specialists in decision-making. The company works with over 60 hospitals in Bulgaria, Romania, Serbia, and Croatia, and some of the largest pharmaceutical companies are among its clients. With offices in Germany and Romania, the company aims to expand into the UK, Switzerland, and the Netherlands in the coming years.

“We want to develop an algorithm for automated service delivery for physicians, researchers, to be used by registries, universities. We want to give them control over their data – to be able to automatically structure the data and interactively ask medical questions or search for specific information, and the software to give them a ready-made analysis. We aim to achieve this goal in the next three years. With this approach we will significantly accelerate the scientific research activity in Europe. Our algorithms currently process over 300 million patient records. This is a vast knowledge, feeding artificial intelligence to self-train and become better and better to benefit us, humans,” explains Mihaylova.

Like ChatGPT for health data, but the answers will be based entirely on medical health records and they will be in a protected environment.

“With ChatGPT, no one can afford to share sensitive information, while with Danny Platform physicians, hospitals, universities control their own data and they can draw insights from it.”

Desislava Mihaylova confidently walks among the metal cabinets in the server room. We’re on one of the floors of a large data center, where her company Sqilline is a client. Lights are blinking, cables intertwine, and small and large metal boxes, like guards, keep data on thousands of patients secure under the watchful eye of security cameras.

However, data and technology are not her first love. Mihaylova graduated in “Business Administration” and has a financial background. The opening to the technology sector comes while working in an oil company and participating in the implementation of business management software for the IT giant SAP.

It turns out to be love at first sight. Mihaylova immediately decides she wants to become a financial consultant on business software and she manages to achieve her goal.

Over the next few years, she accumulates knowledge and experience with the implementation of business software in various companies in Bulgaria and abroad, and in 2009, she decides to found her own company.

She registers the company “Sqilline Business Solutions.” At the moment, Desislava Mihaylova controls the majority of shares. Shares are also owned by Daniel Penchev and Mihail Zhekov—her colleagues in the company.

Initially, Sqilline develops as a consulting company and implements business software in businesses from various industries, mainly abroad. However, in 2017, her path changes abruptly. Mihaylova decides she wants to transition to a product company and at the same time focus specifically on the healthcare sector.

It all starts with a big project in the USA in 2015, in which Sqilline participates. The goal is to build an analytical platform for the American Society of Clinical Oncology (ASCO), which connects the data from about 40 oncology centers.

The push towards the healthcare sector is reinforced by a heavy personal loss. At the same time, she loses her mother to breast cancer at the age of only 55 after improper treatment. Mihaylova walks together with her mother the path of a patient with a serious diagnosis in the Bulgarian healthcare system, and the chaos she sees changes her future forever.

“The entire team and I have turned all accumulated knowledge and what I experienced around my mother, and decided to replicate in some form the project we were involved in in the States,” she says.

They receive support from a major pharmaceutical company at the idea stage, even before they have understood the state of the data in Bulgarian healthcare and the needs of the doctors.

They use the achievements from the ASCO project as a pilot program. They start working initially with data related to breast cancer and move on to other diseases.

“The beginning is the most difficult because you don’t know how these data are stored, in what format, how diverse they are, how you can connect them in the best way and interpret them,” explains Mihaylova.

They develop the algorithm with the help of doctors who guide them on the specifics of various diseases and what would be most useful to them in their work.

Thus, Danny Platform is born (named after Desislava Mihaylova’s mother). The platform processes heterogeneous data from electronic hospital systems in a given hospital or registry, for example, and provides real-time analysis. This allows for better decision-making regarding patient treatment or strategy based on real knowledge extracted from the data of the respective organization.

The platform provides comprehensive searching by various criteria and analyses based on various indicators—such as how a specific factor affects the outcome of patients, comparison of different types of treatment at a certain stage of the disease, and others.

Danny Platform assists in creating scientific and academic research papers, developing medications, and finding suitable patients for clinical trials.

The platform is also used by institutions such as the Ministry of Health and the National Council on Prices and Reimbursement of Medicinal Products. “With the assist of our platform, they can track data in real-time and observe what is happening with cancer patients, for example,” explains the company’s founder.

Sqilline’s newest product, on the other hand, allows tracking how well a patient’s journey aligns with the standard for treating a particular disease. In the standard, approved by the US Food and Drug Administration (FDA), all aspects of treatment and disease monitoring are indicated—from what tests need to be prescribed, through therapy for the respective condition, to subsequent observation. This algorithm is integrated into Danny Platform and shows in real-time how well a patient’s treatment adheres to this set standard. Several hospitals in Bulgaria are already testing the new functionality, and the product is supported by “Innovation Norway.”

Sqilline’s transition to a product company is almost complete, and it continues to grow and develop. According to Mihaylova, the company has recorded a 35% increase in revenues compared to 2023. Expectations are that in the next three years, the business will generate growth of over 200%.

The growth the company is currently experiencing comes after a slowdown during the COVID pandemic. “For our work, we needed access to doctors and hospitals, which we didn’t have at that time. We hadn’t even thought of engaging them,” she shares.

They had existing clients, thanks to whom they have passed through the difficult period, retaining their team, but new clients are missing. Expanding into new markets is also put on hold. “It was a very difficult time for us, but during it, we were preparing,” says Mihaylova.

In 2023, Sqilline entered the Romanian market, and since the beginning of this year, it has been on the market in Serbia and finalizing its entry into the Croatian market.

The company also has an office in Germany, opened during the pandemic, which will intensively start to develop during this year. A big goal for 2024 is to attract a strategic investor, as discussions in this direction are already underway. Additionally, stepping into more markets in Western Europe is also planned.

Desislava Mihaylova and her team do not know who the patients are behind the data processed by their platform. But in rare cases, the door opens, and they have the opportunity to change destinies thanks to their algorithm. Like in the case of the patient who received the diagnosis after two and a half years of suffering. “There is nothing more satisfying than creating a product that saves lives,” says Mihaylova with a smile.

Share this article:

More News & Highlights

Highlights

Health data as an infrastructure

Health data can be considered a crucial component of healthcare infrastructure, as it forms the basis for decision-making, research, policy development, and patient care within...

Read more...

News

Lung Cancer Real-world Evidence Secondary Use of Data: A Non-Interventional Study

Description of clinical study: Non-Interventional Study for Analysis of Molecular Diagnostics and Treatment Patterns in Metastatic Non-Small Cell Lung Cancer Patients (Newly Diagnosed and Relapses...

Read more...

Highlights

“What Prevents us from Reusing Medical Real-World Data in Clinical Research?”

A recent publication in Nature raised the question: “What prevents us from reusing medical real-world data in clinical research?”. Medical real-world data (RWD) stored in...

Read more...