Since its foundation, Riversoft has been active in many areas of growth, from tourism to technology and the marriage of both. Following our initial interview, we sit down with Riversoft’s founder, Alex Kuo, to discuss another topical investment area the company has in the pipeline. We ask him about the pitfalls associated with branching out into the field of medical data, along with the joys associated with it.
In this interview, Kuo shares his thoughts on the primary challenges the healthcare industry is currently facing, and how these challenges might be solved by the rapid adoption of innovative medicine services, such as MedLing.
Riversoft has now ventured into the healthcare sector. Please tell us about MedLing. How did the MedLing project come about? What were the business or industry trends that drove this innovation?
Riversoft was founded in 2008, dedicated to providing travel-related enterprise resource planning (ERP) with a SaaS approach to small and medium travel agencies. Also, we have been focused on developing our own natural language processing (NLP) technology platform since 2016 and market it in a SaaS model called Lingbot. During the COVID-19 pandemic period, we got an opportunity to develop a medical service chatbot for one of the largest biopharmaceutical companies in the world and developed the MedLing NLP engine for one of the largest hospitals in Taiwan. We are planning to develop a complete system of travel, travel insurance, and also travellers’ healthcare.
Can you briefly explain the technology that powers MedLing and how you utilise it?
As I mentioned previously, we have an NLP service called Lingbot. To adjust Lingbot to fit medical services, we have further defined an NLP engine, MedLing. With NLP technology, we are able to recognise medical terms in clinical notes, to get the correct semantics to exclude the negative medical terms and include the positive ones, etc.
MedLing uses Riversoft technology to improve the medical information system. Can you explain the problems MedLing aims to solve or the operational gaps that MedLing is trying to fill in the current system? How do you evaluate its effectiveness?
In order to answer your question, let me first introduce you to ICD-10. ICD-10 is the tenth revision of the international statistical classification of disease and related health problems. This is a medical classification list curated by the World Health Organisation (WHO). The ICD-10 code set contains codes for diseases, signs, symptoms, and external causes of injury or disease. This system has become a common protocol for disease classification. An ICD-10 coder is a certified administrator in a hospital who acts as a gatekeeper to identify patients’ medical classifications and match them to the appropriate insurance category. This labour-intensive work is expected to be improved by phasing in medical AI to enhance productivity. In a hospital, an ICD-10 coder needs an average of 20-40 minutes to classify a case.
We expect MedLing to be able to read patients’ discharge notes or history and match the results to their appropriate insurance category. This automatic system can also shorten the data processing time and enhance the productivity of medical staff.
We are continuing to fine-tune our Systematised Nomenclature of Medicine (SNOMED) ontology inference engine. The accuracy is about 0.85 right now. By the way, one of the largest hospitals in Taiwan is now using MedLing to assist the ICD-10 coder in order to improve their performance.
There are other companies also offering ICD-10 encoding services. What sets MedLing apart from other services?
First of all, we integrated Systematised Nomenclature of Medicine – Clinical Terms (SNOMED CT) ontology into the Riversoft approach. We extract keywords from medical records and map them to SNOMED CT ontology in order to find the disorder candidates. The parsed SNOMED CT concepts are then used to generate a precise training corpus for ICD-10 classification fine-tuned by means of a pretrained bidirectional encoder representations from transformers (BERT) model.
Second, our footprint is quite small, so that it is possible to do distributed computing.
What other fields can MedLing be utilised in? What are your long-term plans and goals for MedLing?
MedLing has begun to be used on medical record service for clinical decision support systems (CDSS).
We further define three possible business scenarios to expedite the commercialisation of the MedLing engine and its application.
Firstly, we have had discussions with a local insurance company about using the MedLing engine to assist them to speed up their medical claim service.
Secondly, we are developing a chat service for integration with wearable IoT, with cardiology diagnosis provided by professional medical institutions.
Thirdly, we plan to develop a platform for second-opinion services supported by the MedLing engine. This second-opinion service will be applied to meet the needs of travellers and tourists. We are conducting discussions on this with one of the large Japanese insurance services designed for foreigners travelling in that country.
What does the development and success of MedLing mean to Riversoft? After the pandemic, will the healthcare sector remain one of Riversoft’s development goals?
Actually, it is because of the pandemic that people started to notice the importance of e-medical services. So, yes, healthcare remains one of Riversoft’s important fields.
Our company began from travel. Years ago, we started a travel insurance business. We’ve always wanted to build up a complete system consisting of travel, travel insurance, and also travellers’ healthcare. So we were all pretty excited about the birth of MedLing.
Alex Kuo is the Chairman of Riversoft, a Taiwanese software company focusing on developing innovative solutions and management systems for the travel and tourism industry. Kuo has an MBA from George Washington University and is pursuing an EDBA from Université de Liège. He had over 20 years of career in the Acer group holding senior positions. Before moving to Canada, he worked in two publicly-listed tech-based companies in China for six years. Kuo has also been a venture angel supporting the growth of more than 10 companies globally.