When technology meets healthcare, everyone benefits.
With innovation as one of its pillars, Oncoclínicas sought out Dataside to enhance its excellence through intelligent technology, in partnership with Microsoft.
Learn more about the first use case of Azure Text Analytics for Health in Portuguese, as well as the web portal and mobile app developed to store clinical and medical procedures using Azure Cognitive Services.
Azure Text Analytics for Health
According to Microsoft, Azure Text Analytics for Health is a tool that extracts insights from unstructured medical data.
What is unstructured data?
Unstructured data refers to information that does not follow a specific format or organization. It is not organized in columns, tables, or fields, and lacks a predefined pattern. Unlike structured data, which is organized in formats like databases, spreadsheets, or management systems, unstructured data is more difficult to analyze and interpret automatically because it lacks a clear structure.
Analyzing such data can be complex and requires advanced natural language processing (NLP), machine learning, and other unstructured data analysis techniques to extract meaningful insights.
Bringing it to the healthcare context
Imagine a clinic responsible for many patients, each in a different stage of treatment. Now, think from the patient's perspective—many patients likely have a drawer or folder full of envelopes collected over the years. Some test results might even get lost along the way.
As a result, diagnoses for starting any treatment can be delayed. This directly affects people’s health because it delays the start of treatment, and we know that in some cases, a patient's recovery depends heavily on how quickly treatment begins.
This is where Microsoft’s tool becomes important. As mentioned earlier, Azure Text Analytics for Health extracts insights from unstructured data but specifically from medical data, identifying terms related to the medical field.
"Text Analytics for Health uses NLP techniques to locate and label valuable information in unstructured clinical documents, such as medical notes, discharge summaries, clinical documents, and electronic health records. It also links medical ontologies and domain-specific coding systems, identifies meaningful connections between concepts mentioned in the text, and addresses negation in medical text," explains Microsoft.
Another example that highlights the intelligent application of this tool is the ability to analyze disease outbreaks. For example, suppose there were 10 cases of "y" virus last week, and this week there are 20. It’s possible to quickly identify, predict, and control the increase or decrease in disease cases.
Oncoclínicas Case
The growth of the Oncoclínicas Group stems from a culture of innovation. Combining a team of excellent professionals with investment in technology is the strategy that delivers multiple benefits to all involved.
Oncoclínicas is a pioneer in oncology service management in Brazil. Serving two main groups—patients and their families, and healthcare professionals—the group is a reference in oncology, radiotherapy, hematology, and bone marrow transplantation.
As the largest oncology group in Latin America, Oncoclínicas operates in 13 Brazilian states and the Federal District, with 129 units across the country. The group provides over 1.4 million consultations annually— quite a large amount of data to analyze!
The primary challenge when the group approached Dataside was processing medical exams at this large scale. Centralizing the data in an efficient system would bring benefits to healthcare professionals and patients by speeding up treatments and providing critical insights for treatment studies. Dataside not only achieved this but also went beyond.
"Microsoft and Dataside were strong partners. It was an innovative project, and we knew we would face challenges, but we overcame each one and are very pleased with the result." - Marcio Guimarães, Executive Manager of Data and Analytics at Oncoclínicas.
Solution offered by Dataside
In addition to digitizing the data, Dataside enabled understanding through NLP (Natural Language Processing) and OCR (Optical Character Recognition), processing the text in PDF documents (the exams) and structuring the data.
This allows for quick access to each patient’s history. Instead of merely opening a PDF and reading the document, the information is already structured in a database.
For example, patient "A" took test "B" in region "C" with doctor "X," and the result was positive or negative. The information is efficiently understood.
AI was applied in the project, resulting in two platforms: a mobile app, which allows users to take a photo of a document to upload it into the system, and a web platform for both consultation and uploading pre-digitized PDF documents.
One of the main benefits of the project is the improved quality and speed of consultations. With this intelligence, doctors can arrive at conclusions faster through more efficient questioning, without compromising consultation quality. Another benefit is the reduction in document analysis costs.
Focused on solving the client’s problem, Dataside developed the world's first project to use Microsoft's Health API in Portuguese.
"Unstructured text remains a challenge in the healthcare industry—made even more complex by language barriers. Unlocking biomedical NLP technology in multiple languages is a huge step forward in bridging gaps in health equity caused by language barriers. At Dataside, along with Microsoft, we aim to continue creating technology to help close this gap," explains Microsoft in the article "Expanding AI technology for unstructured biomedical text beyond English."
"The first benefit is being able to access important information in a structured way, which can now be used for statistical, analytical, and other purposes. The second is how much easier it is to structure the information." - Marcio Guimarães, Executive Manager of Data and Analytics at Oncoclínicas.
תגובות