Augmented Intelligence – The Operational ROI Engine

Brian Hammond, Chief Technology Officer, Tampa General Hospital

One positive outcome from the pandemic has been the acceleration of advanced analytics within healthcare organizations. At Tampa General Hospital (TGH), predictive modeling for hospitalization, ICU capacity and ventilator usage has proven its worth to senior leadership in the way of managing surgical volumes, staffing and biomedical equipment heightened by the response efforts.

Daily reports, interactive dashboards and weekly projections have been the staple information provided by analytic teams to assist operations. However, throughout this journey there is the realization that decision making continues to rely on the experience of team members. Due to the complexity of the environment, consequences of these decisions, at times, are largely unknown. The key to advancing efficiency lies in wrangling the vast amount of available data into applications that can analyze different scenarios and provide insights to help leaders make more informed decisions or select the “best” pathways forward.

The main challenges to implementing operational analytic applications are the readiness of the organization, sufficient technology platforms, and enough developers with the appropriate skills. As a result, there are a plethora of companies developing targeted solutions; however, many are immature and there is a limit to the number of these solutions that most organizations can afford. The large analytic companies continue to mature their very robust analytic platforms that offer healthcare-related libraries and algorithms and are ready to accelerate analytic applications. So how do you proceed to create a robust infrastructure with support teams capable of maximizing the potential of these applications and platforms? These are the steps we’re taking at TGH to build the organizational and technical foundation to implement and support the strategy:

• Develop your overall vision alongside members of your organizational analytics steering committee to gain their insight and support.

• Create a data literacy and governance program with leaders outside of Information Technology to help set expectations and adoption of these programs.

• Foster deep working relationships with operations to target projects and gain their commitment to invest their time for development and testing. Creating an atmosphere of trust where honest feedback can be given is important to developing useful applications.

• Determine the analytic software companies that will be your partners. These partnerships need to include experts with their software to assist with development and training your team.

• Engage your internal team in the review of analytic platforms to provide a wealth of information for management to help select the appropriate software partners. All platforms have their strengths and weaknesses, thus giving your team the time to effectively vet these is critical.

• Embed your data engineers with software partner experts so they learn the underlying applications. Eventually you’ll want to pivot work more towards your internal teams to reduce costs.

• Invest in building and developing your integration team. You’ll need internal teams that can implement application interfaces, FHIR and HL7 messaging to be successful. Real-time data is critical to the success of the program.

Combining user selections with real-time data and AI can provide the augmented intelligence that operators need to improve decision making and organizational efficiency. While improvements in technology inherently result in advancements in analytics tools, the real key to success is the creation of an organizational team that is committed to implementing and supporting these advanced tools. An innovative culture among leaders in the organization is the most important aspect to a successful analytics program.