Delays or errors in disease diagnosis increase healthcare budgets, hospitalisation and may even cost lives. European researchers have developed a digital assistant where, for every symptom entered, it returns a list of causes.

Digital Economy

Health

Each year, millions of patients suffer and even lose their lives due to misdiagnoses at various levels of healthcare provision. The continuously increasing number of diseases and related information can be overwhelming for clinicians who provide a wrong diagnosis 15 % of the time. Moreover, patients suffering from rare diseases may wait for years until a correct diagnosis is reached.

Using AI to support medical decision-making

The aim of the EU-funded Symptoma project is to improve diagnostic quality while reducing cost and time spent. According to the CEO and co founder of the Symptoma SME, Jama Nateqi: “Our vision is to offer every patient the right diagnosis and treatment they deserve, paving the way towards precision medicine.” The Symptoma team has developed the world’s most accurate digital health assistant. It consists of a chatbot that uses artificial intelligence (AI) to ask disease specific follow-up questions after considering the information entered by a user. Through this process, it may identify symptoms that are strong indicators of certain conditions, thereby reducing the number of possible diseases. After assessing all symptoms, the Symptoma search engine returns a list of potential medical causes sorted by probability. Unlike other online symptom checkers that often employ limited databases, Symptoma also includes information on rare diseases, overcoming skewed probability rates and inaccurate diagnoses. Its database takes into account more than 20 000 diseases and billions of connections to symptoms and risk factors. Medical journals, user data and electronic health records serve as sources for this information. Symptoma has attained the highest diagnostic accuracy rates according to internal, external, and peer-reviewed studies. Pfizer contracted a study comparing 107 solutions worldwide, concluding Symptoma’s to be the best. Even a competitor has confirmed its leading performance. Currently, Symptoma has 10 million users every month and overcomes language barriers as it operates in 36 languages.

Symptoma’s health assistant in COVID-19 management

“Our most significant achievement during the project was our impact on the pandemic,” says Nateqi. The Symptoma team trained its symptom checker for COVID-19 in January 2020, making it the first and only AI system to differentiate it from 20 000 other diseases, even those with similar symptoms. Results for over 1 000 cases validated a 96.32 % accuracy at assessing the COVID 19 risk in symptomatic patients. Since then, various governments, cities, hospitals, and companies have employed Symptoma’s technology to help fight the current pandemic. So far, the symptom checker has delivered more than 100 million risk assessments related to COVID 19. These have been combined with thousands of diagnostic test results, leading to new real-world evidence and a better understanding of the virus and its symptoms. “Symptoma is the first step towards a new paradigm of diagnostic decision-making based on objective rather than subjective reasoning,” concludes Nateqi. Partners envisage Symptoma’s solution as an accurate diagnostic support tool used by clinicians to aid rapid disease diagnosis taking into consideration the latest findings on rare diseases. At the same time, it is expected to contribute to a better understanding of biomarkers and clinical signs.

Keywords

Symptoma, symptom checker, symptoms, health assistant, digital health, AI, COVID-19, rare disease, precision medicine