Fujitsu has developed an artificial intelligence text mining application that automatically translates medical plain text into medical codes.
At the Fujitsu Innovation Gathering held in London, Fujitsu European Laboratories announced new artificial intelligence technology, which automatically translates plain text into medical code, making it easier to process and thus saves up to 90 percent of the time. The new automatic medical coding solution, which greatly simplifies compliance with medical classifications, reduces the registration process, which normally takes an average of 15 minutes, to less than a minute. Unlike previous generation technologies, Fujitsu's Artificial Intelligence text mining technology combines semantic information and Natural Language Processing (NLP) and Deep Learning technologies to analyze medical notes and extract valuable data.
Fujitsu European Laboratories have been working on various clinical projects for nearly four years with innovation partners in the health sector, including Madrid's leading San Carlos Clinical Hospital. Medical Coordinator In her speech, Julio Mayol explained the importance of birlikte co-development yaklaşım approach from a medical point of view; “We are constantly looking for new ways to improve clinical decision making. Our work with Fujitsu European Laboratories is a key development to improve productivity. Most of the health care systems available today do not fully meet the requirements of the doctor-patient relationship. The truth is, the difficulty in using existing systems is causing wear to healthcare professionals. With innovations such as Fujitsu's new Artificial Intelligence text mining technology, we can achieve concrete improvements in the clinical decision-making process. ”
Dr. Adel Rouz, Chairman of the Executive Committee of Fujitsu European Laboratories, said in a statement; Strate Our strategy to act with business partners, such as the San Carlos Clinic Hospital, gave us an important insight into the challenges facing the healthcare sector, especially in medical decision-making. We were able to put forward many important innovations that make a difference in the business processes of medical professionals. This solution helps to increase the reliability of clinical data, as well as to digitize hospitals, health insurance companies and government agencies. We believe that our technology can be easily adapted to solve similar problems in many areas such as insurance, law and compliance. ”
Doğru yapılandırılmış bilgi, tıbbi konularda karar verme aşamasında ve sağlık hizmeti sunumunu iyileştirmede önemli bir rol oynuyor. Bununla birlikte, klinik sahada çalışan uzmanların hastalara ayırabildiği zamanın gittikçe azaldığı günümüzde Sağlık Takip sistemine giriş sırasında kaybedilen zaman daha da anlamlı hale geliyor. Sağlık uzmanlarına, hasta raporunu yazarken düz metin kullanımı gibi daha esnek veri giriş yöntemleri sağlanarak, hasta başına düşen maliyet azaltılabilir ve aynı zamanda uzmanların daha etkin hasta verisi kaydetmeleri sağlanır. Fujitsu Laboratuvarlarının bu çözümü, Sağlık Kayıt sisteminin gerektirdiği yapılandırılmış bilgileri sağlık uzmanının düz anlatım metninden otomatik olarak alır, derin analiz yaklaşımını kullanarak doğru kodlamaya çevirir. Sistem, birçok mevcut kodlama sisteminin kullandığı karmaşık dilbilimi kurallarına kıyasla ek esneklik sağlıyor. Sonuçta sistem, Uluslararası Hastalık İstatistikleri ve İlgili Sağlık Sorunları Sınıflandırması (ICD) kodlarının içinden çok daha geniş bir kesit alanı çıkararak eşleştirme havuzu kullandığı için yüksek doğruluk derecesi elde edilebiliyor.
Fujitsu’nun Yapay Zeka Metin Madenciliği yaklaşımı iki temel bileşenden oluşuyor:
Building the knowledge base: a pool of information has been designed to reveal medical classifications that are semantically enriched by external sources. Ontologies and word phrases are used for semantic enrichment.
Recognition and assignment: Medical terms using the possibility of deep learning and analysis include the definition of weighted score ranking formulas to calculate the recognition process followed by the potential coding of clinical grades.
Fujitsu's Artificial Intelligence technology was evaluated on two sets of English data sets, including 200 private anonymous clinical notes and 5000 abstracts extracted from the MIMIC-III1 database. Another advantage of this technology developed by Fujitsu is that it can be classified as another medical or easily adapted to other languages and does not require pre-annotated data sets.
(1) MIMIC-III is a public clinical database and a current reference / gold standard for comparing such problems in the healthcare community. MIMIC-III contains unidentified health-related data from Beth Deaconess Medical Center's Critical Care Units between 2001 and 2012. This includes demographic data, vital sign measurements, laboratory tests, clinical notes, and so on. information included.
