By Sujey Edward, CTO
While it may be difficult to believe, the leading diseases causing death globally are largely preventable and/or treatable. Yet these diseases continue to kill at rapid rates. Experts in emerging technologies are working hard to reverse that trend by innovating side by side with healthcare experts in the federal health IT sector. As a result, agencies are starting to make strides by adopting automation, embracing plans that include artificial intelligence (AI) while plotting a future in which the public is served more efficiently. This growing collaboration is key to yielding better, faster outcomes that save lives. But more is needed.
As a CTO with a passion for innovation and lifesaving technology, I encourage our teams to challenge the status quo and come up with ideas to solve REAL problems our Federal Government customers face in health and beyond. Here are five AI-enabled innovations Octo is proud to have developed for health agencies, tools that also can be customized for organizations across the Federal Government:
COVID-DPR – Most illnesses have patterns that need to be discovered, and that cannot happen without collaboration. The COVID-19 pandemic highlighted the lack of real time, worldwide collaboration critical to a vaccine and the return to a more normal way of life. Globally, clinical institutions, researchers, and medical experts needed a way to enhance data availability and leverage AI. The COVID Digital Pathology Repository (COVID-DPR) was created to enable international collaboration by providing a centralized, cloud-based repository for sharing and annotating digital whole slide images of lung, liver, kidney, and heart tissues from patients infected with COVID-19, as well as the closely related coronaviruses associated with SARs and MERs. Octo built the interface to enable HALO, an AI powered medical research tool, to be accessible to the public and other health agencies to assist NIH in gathering as much data as possible. This tool helps researchers and practitioners quickly identify patterns to facilitate collaboration around actual findings from all over the world. While the tool is currently being used to address COVID-19, it can also examine other illnesses we are still learning about.
LAVAAR – Understanding medical diagnosis and treatment options can be complicated for anyone, which is why Veteran’s Administration patients and staff needed an intuitive application to visualize the human body. Octo’s answer was the Light Application for Veteran’s Administration Augmented Reality (LAVAAR), which bridges the communication gap between medical personnel and veteran patients. Using augmented reality, LAVAAR leverages APIs to ingest patient images from digital medical record systems. Practitioners can explain medical imagery (CT, MRI, X-ray, PET, Ultrasound, and others) using LAVAAR’s photorealistic, real time, 3D game engine representation of the patient’s body. LAVAAR can also be used as a visual aid for medical educators. The learning mode presents a true to life 3D model with all major anatomical systems. This tool can be customized and used in any medical setting and can be applied in Defense, Intelligence, and other sectors.
NLP NER – When it comes to preserving lives, speed and accuracy are essential. But data abstraction can be a time and labor-intensive undertaking. Octo has proven that medical chart abstraction can be automated using AI. We leveraged a combination of natural language processing (NLP), specifically named entity recognition (NER), to extract data from electronic health records and a neural network to automatically answer algorithm driven questions. That data can be used to report a variety of findings, such as adverse event rates for hospital-acquired conditions. This tool demonstrates abstraction can be fully or close to fully automated via AI and be useful in virtually any sector.
G-Search – Grants are critical to the advancement of our health practices and the overall health of our society. Information on research funding is important to investigators, policy analysts, advocacy organizations and, of course, the funding agencies themselves, but it is not always easy to sort through. G-Search is Octo’s modernized system that uses publicly available grant data and open-source software to statistically analyze grant research information. Using machine learning (ML) techniques, it optimizes the efficiency of searching for grant opportunities and lists funding data sorted by grantors, agencies, and others. G-Search’s algorithm is based on the principals of a pointwise approach to an ML ranking algorithm. This powerful tool can display results dynamically by assigning metrics to all search results. Metrics are a combination of weighted and established criteria with ranking based on user interaction. This changes the priority of grant search results. G-Search can filter results, display charts on funding amounts per organization, list open and close dates, and more, all while offering advanced search capabilities to reduce time and resources used in grant research.
e-Approve – Faster decisions are essential to meeting health agency missions, including to enhance health, lengthen life, and reduce the burdens of illness and disability. However, agency staff are often called upon to complete repetitive tasks, detracting from more complex work that needs to be done. Octo’s Robotic Process Automation (RPA)-fueled e-Approve robot automates verification of online forms, scanning documents and checking for accuracy. Serving as a digital assistant, e-Approve can examine structured and unstructured data in PDFs, text-based documents, spreadsheets, scanned images, and more. The robot holds administrative access to examine each document uploaded by the user. It then compares submitted data to data in an existing database. If the information provided in the database matches the document data, the robot can record a decision. (For example, it might approve an application or mark a user as current.) If not, the robot denies document submission. If documents are missing, the robot can send an email to an employee to request investigation and follow up. e-Approve can reduce processing time from two hours to under five minutes while eliminating errors.
Medical experts and the Federal Government need tools like these to manage data, streamline processes, and ensure success for end users, whether those end users represent the government or the public. AI is the way to get there, no matter who your agency serves, saving lives while reducing time consuming, repetitive tasks and associated costs.
For a demo of any of these technologies or to discuss how Octo can customize an AI solution for your organization, email CTO@octo.us.