Healthcare Data Curation Platform

To realize the value of analytics, machine learning, and artificial intelligence, healthcare organizations need a platform that can clean and curate data.

Electronic Medical Records (EMRs) have such low data quality that the raw data from today's system are not useful without spending significant resources by healthcare organizations to collect, clean, normalize, and group data. Today's EMRs are complicated systems and they require a lot of data to be entered correctly. Even minor errors can affect the quality of the data and thus care and outcomes. Data are often incomplete, inconsistent, and they lack standardization. EMR systems were mainly built to support the documentation needed for billing and for reducing liabilities. Therefore, these systems were not built to prepare data to make the captured data analytics or AI ready.

Poor data quality in EMRs can have several negative consequences. It makes timely and cost-effective decision-making very difficult and expensive. It can also make it difficult to conduct research and improve healthcare quality. The lack of data quality and standardization prevents health systems from achieving the value realization of AI and data science. In healthcare today, data quality remains one of the most intractable problems preventing the creation of solutions that would enhance quality of care, address health inequities, improve financial performance, or optimize operational efficiency.

Improving the quality of data in EMRs is an important challenge, but it is one that must be addressed. Until the healthcare ecosystem solves these problems, the expectations of artificial intelligence to create needed efficiencies, leveraging health equity initiatives to address disparities, or augmenting clinical care and operations via pragmatic decision-support tools, will not be realized.

We posit that data quality is the most pressing problem in healthcare today that drives many of the downstream complexities that administrators, clinicians, and researchers seek to solve today.

The Fullsteam Health data pipeline is a platform technology that curates, normalizes, groups, and monitors historical and real-world data from electronic medical records and other operational data sources. The Fullsteam Health data pipeline was built by a leading academic healthcare institution, meant for use by healthcare institutions to advance health and equity.

The EMR raw data is cleaned and converted into an actionable form to be used for administrative decision-making, clinical decision-making, and for any data science (AI/ML) model and solution development for operational and clinical deployment. This makes the data readily available for large language models (LLMs) and applications. Without clean data readily accessible, the same data elements are cleaned several times every time it is used, which increases cost and redundancies of efforts. Moreover, as data analysts have their own ways of cleaning data which leads to different interpretations of the same clinical concept. With our platform, data gets cleaned once at the enterprise level and is readily available for all downstream applications. This reduces the total cost of operations and accelerates time-to-value of solutions that require clean, curated and actional data.

The data curation platform is completely containerized, which means it can be deployed inside the health systems using either on-premises or public cloud infrastructures. Our platform stays behind the firewall of the healthcare institution and does not send data outside of the firewall. The platform also has tools to check the quality of the data and notify the relevant people of any issues.

The Fullsteam Health data pipeline uses a set of open-source software with proprietary code that extracts historical and real-time data from a plurality of data sources. The extracted data are cleaned, normalized, standardized. Meta data and data are continuously mapped into clinical and operational concepts. The platform incorporates a notification engine that monitors data quality. This general-purpose notification engine can be used to alert when data quality issues are observed and also for sending AI/ML inference outputs to appropriate care providers for clinical workflow uses.

This data curation platform has been in operation at Duke University Health System for over 5 years. 

Fullsteam Health helps healthcare organizations use selected data to overcome clinical and operational challenges that were hard to deal with in healthcare before. The platform assists both front-line clinicians, who provide patient care, and executives (CEO, COO, CMO) in making operational and strategic choices.

The Fullsteam Health data pipeline has been beneficial for patients, as it has helped to lower the death rate of sepsis patients by 22% through the Sepsis Watch solution. The Fullsteam Health data pipeline has also enabled many machine learning models in the past five years, including inpatient models that predict surgical complications, inpatient mortality and emergency department triaging, as well as outpatient use cases such as chronic kidney disease and incident HIV. The data is of high quality for research and has supported drug-related research projects, such as studying the occurrence of RSV in young children.

During the COVID-19 pandemic, the Fullsteam Health data pipeline was used to monitor health disparities for inpatient and outpatient care. The Fullsteam Health data pipeline is a production-ready system that supports operational efficiency and provides insight into current and historical disparities in care.

We work with healthcare organizations to find out the problems they have in their community and use design-thinking methods to come up with realistic and sensible solutions for those problems. We use selected data to drive the solutions so that administrators, clinicians, and researchers can address the problem faced by the community.

As the examples above show, we don't just offer solutions that fix particular operational and clinical problems. We go beyond that, by enabling decision-makers to tackle complex issues that make sure everyone in a community gets better service. Our firm conviction is that problem-solvers need clean, curated, and constantly checked and sorted data to put into action the best solutions for the challenges faced by the diverse players in the healthcare delivery ecosystem.

The Fullsteam Health team consists of people who have worked for decades at leading healthcare enterprises that focused exclusively on the creation, development, and implementation of solutions. Our team also includes some of the top experts in the health intelligence platforms coordination at health systems across the country. Our team members have the qualifications (MBA, MHA) and experience working with various healthcare institutions in both well-resourced and under-resourced settings that will be used to demonstrate value. Our team also has a comprehensive understanding of the IT components used in healthcare delivery. In short, our team is made up of data science and business and operational innovation professionals who understand the needs to transform healthcare and the needs to support operational change.

We bring this expertise, but we also ask for advice from market and industry experts such as clinicians, researchers, and patient communities to make sure that the solutions we are applying have the best potential to address the issues that worsen the problems observed in healthcare quality, healthcare operations, healthcare access, and healthcare disparities.

We have expertise in research and will use the platform to help the R&D activities from industry, NIH, and regulatory bodies to speed up product development and evidence generation efforts.

We are putting together an advisory board of experts from different areas of the healthcare sector. We will have leaders from healthcare systems, payers, and the pharma/biotech industry to help us create a product that suits their needs.

We plan to hire more staff to work on business development/sales and product development to help us create new solutions and features for the platform. By participating in the Cure Xchange Challenge, we can improve the product and also work on go-to-market strategies and sales strategies and get advice on getting non-dilutive funds (e.g. SBIR/STTR). 

Our data curation platform is a novel approach to ensuring actionable data for health systems and hospitals. Through a containerized platform, we extract data from a plurality of data sources (historical and real-time), cleanse and normalize the data, group the data into clinical and operational concepts, and monitor data and meta data continuously to ensure that data concepts and statistical models are maintained. Our platform also leverages a proprietary notification engine that allows for inferences derived from solutions to be delivered into the most appropriate workflow through multi-modal approaches (EMR writeback, Microsoft Teams message, Email, SMS, Pager, Spok message).

The problem of data quality is acute and necessary to be addressed if the healthcare ecosystem is going to leverage the value of analytics, machine learning, and artificial intelligence. Our platform is the only data curation platform built for purpose by healthcare institutions, for healthcare institutions. 

Since we have a data pipeline to continuously curate and clean data, it will allow us to identify chronic disease patients and the exacerbation of their disease early on so that we can keep them healthy through the identification of appropriate care pathways.

It is also possible, with our data curation platform, to set up national level surveillance with de-identified information for identification of pandemics, even before they become widescale.

In alignment with the goals of ensuring healthy lives and promoting well-being for all at all ages, we can identify patients in communities who need pre-exposure prophylaxis (PrEP) early on. The goal with each of these use cases that align with the UN Sustainable Development goals would be to identify where there are opportunities to reduce the total cost of care to put patients on correct therapeutics and care pathways and lower the overall total cost of care.

For artificial intelligence to create lasting benefit in healthcare requires curated data. We are the primary source for good, curated data. This becomes the platform to scale novel AI/ML to advance equity in healthcare. We can immediately deploy artificial intelligence and machine learning models for Sepsis identification and risk, proactive identification and intervention to address gaps in care for patients with peripheral artery disease, assistance in the triage of patients with traumatic brain injury, a maternal early warning system for obstetric patients at risk for clinical deterioration, as well as others. We will also seek to be productive partners with others in the cohort to assist them with access to the curated data we are able to provide as well as supporting the validation and deployment of the models in workflow use. 

We are acutely aware of the need to ensure ethical and responsible use of AI. Our platform was initially developed at Duke and works in conjunction with Health AI Partnership (healthaipartnership.org). Every model developed we work on, in partnership with Duke, is analyzed for bias.

Through the Health AI Partnership, we assess and monitor AI performance, audit the AI solutions through a framework used to enhance equity and reduce bias in any model that we put into routine use to determine whether models need to be updated or decommissioned. To do this effectively, we leverage The Health Equity Across the AI Lifecycle (HEAAL) framework. Through this, we assess accountability, fairness, fitness for purpose, reliability and validity, and transparency.

In the next year, we will be deploying our solution into two health systems in Louisiana for the purpose of identifying patients who would benefit from pre-exposure prophylaxis (PrEP) to ensure they receive the therapeutics and care needed.

We are also engaged with health systems across the country in business development efforts to show the value that our data curation platform can provide them in making their data actionable by leveraging clean, curated data for administrative, clinical, and research purposes.

Over the next five years, we will continue these business development efforts as well as to create value-creating solutions and features that will sit on top of the data curation platform that health systems can use to enhance decision-making, both operationally and clinically. 

We currently have 1 full-time employee with plans for growth planned for data engineers and product engineers.

We also leverage part-time data engineering contracted resources to assist in the extraction of real-time data as well as the deployment of the data engineering platform at health systems. We have part-time clinical, business, and technical advisors that curate our tech stack optimization plans, clinical and business use cases, as well as support go-to-market strategies. 

The Fullsteam Health data curation platform has been operational at Duke University Health System for 5 years. The team, whether employees, contractors, or advisors, were many of the people who developed the platform and have optimized the platform as data engineering concepts and technologies have evolved.

Our platform has been de-risked from a technical and operational standpoint, having been built, scaled, and deployed in one of the most prestigious academic medical centers. 

Currently, our team of co-founders, contracted resources, and advisors totals 7 individuals. Two of which are Caucasian, two are Middle Eastern, two are South Asian, and one is of Asian decent.

As our work over the years in deploying solutions focused on health equity and addressing bias in healthcare artificial intelligence, we deeply understand that addressing any challenges requires men, women, and people from different races and cultures to ensure that thinking is not limited, and old approaches are appropriately challenged. 

Our model to execute and deliver impact is twofold; deploying our data curation platform at health systems and hospitals to delivery high quality, curated data and creating and deploying value-creating solutions on top of the platform for these health systems and hospitals to achieve clinical, administrative, and research utility from the curated data they have access to via the platform.

To accomplish this, we have already been engaged with health systems across the country from a business development standpoint to illustrate what Duke has been able to achieve that no other health system has been able to at this point. We show the value of curated data through live demos of modules and visualizations that Duke University Health System uses to make action-oriented decisions.

Over the next year, we will bring in revenues to support NIH-sponsored activities in Louisiana and Pennsylvania to enhance the adoption of PrEP and triaging for traumatic brain injury.

Seeking ways to further our reach into health systems and hospitals, we have created a partnership with a global consultancy to be a distinctive solution for them to offer to their clients as they engage in data analytics and data architecture engagements. Many consultancies approach health systems and offer solutions that move data from one database to other databases that are easier to manipulate data. However, when you move unclean, messy data from one place to another, it is still unclean, messy data. We are the only company that can solve this problem. Working with this global consultancy will enhance our outreach and we will be able to leverage their global expertise in market framing, go-to-market strategies, and product engineering.

From an operational standpoint, we will make strategic hires by recruiting a Chief Technology Officer to assist with further optimization of the tech stack and also strategically hire product engineers to build the necessary features and solutions, leveraging design thinking and adoptable user interfaces, so clients can also purchase value-creating use cases to help make decision-making at an institutional level for clinical, operational and research initiatives. 

We will go-to-market with a SAAS business model for health systems and hospitals for the data curation platform. We will amplify this revenue model via customized AI models and solutions for health systems that they can purchase to optimize decision-making.

In the short term, we have begun raising financing through a $400k SAFE note. In addition to this, we are being brought in on government-sponsored research studies that will bring money into the company to support the deployment of the data curation platform at health systems across the country to enable these research opportunities.

We will also submit for other non-dilutive government-associated financing through programs such as SBIR/STTR as well as the ARPA-H Biomedical Data Fabric initiative.

Long term, our business model will be sustainable through the SAAS model providing our data curation platform as well as ancillary solutions that enhance health system value.

The current operation costs are in stealth mode with our current operational costs for deployment totaling $120,000 in direct labor costs, $15,000 in travel costs, and $40,000 in other direct costs. These current operational costs total $175,000.

We will also be exploring the need for co-working office space which is currently not allocated in our operating expenses. 

We are seeking $100,000. This funding will be used to further accelerate supporting the tech stack optimization through the hiring of a Chief Technology Officer. We will also use the money to increase the speed at which we can bring in contracted product engineers to assist with the development of solutions in our roadmap to show enhanced value to health systems and hospital in what can be done with the clean, curated data our platform provides. 

By participating in the Cure Xchange Challenge, we can improve the product and also work on go-to-market strategies, sales strategies, receive advice on getting non-dilutive funds (e.g. SBIR/STTR), and enhance our networking opportunities to engage with health systems, hospitals, and other channel partners. We are also excited about the opportunity to work with individuals familiar with product development as we accelerate our product and solution offerings.