Rom Eizenberg
How hospitals can improve patient journeys with AI
August 23, 2024
By Rom Eizenberg
Today, we can walk into a FedEx shop and drop a parcel. That parcel will go through numerous trucks and regional, national, and international distribution centers. The parcel will move through multiple airports and several airplanes, and somehow miraculously it appears at the desired destination. Even though drivers might be sick and trucks might break, the parcel will show up – potentially on the other side of the planet – in about 24 hours.
Compared with the often slow-moving and inefficient journey of the average patient – cumbersome forms, insurance nuances, triage, staffing delays, and lengthy wait times – this leads to some uncomfortable questions for health systems. Why can't we do this in healthcare? Why do healthcare leaders feel that they are continuously coping with crisis after crisis? Why do they feel that they are incapable of predicting the future and aligning resources?
These questions concern the factors behind many of the inefficiencies that contribute to the incredible cost of delivering care across the U.S. As an industry, health systems are struggling not with structural challenges, but with modernizing the process of managing health systems, aligning resources around care delivery, and catching up so a hospital can be as predictable, manageable, and optimal as delivering a parcel across the globe in 24 hours.
What prevents this from happening? Two words: data silos. It is almost impossible today for a health system leader to look at the inpatient journey end-to-end, from admission to discharge. Each part of the process of delivering care is contained within a different legacy system, with siloed data, often held by different vendors, and, quite frankly, no one has the big picture. As a result, while today’s hospitals generate an enormous amount of data by second from their care operations, much of the data remains untapped, representing a significant missed opportunity to create value.
Using machine learning (ML) and artificial intelligence (AI), we can dive deep into the data of patient journeys to uncover previously inaccessible insights and opportunities, and ultimately improve hospital operations and patient experiences.
Begin by exploring “known unknowns” and then “unknown unknowns.” In other words, start by finding answers to the questions you know; then evolve into new answers that you aren’t even aware of today.
Examples of “known unknowns” include:
● The discrepancy between planned patient discharge time and actual discharge time
● What factors affect the length of stay (LOS) for patients, such as staff-to-patient ratio, equipment readiness, bed turnover, etc., and to what extent
● Whether the right asset is available in the right place at the right time for efficient load balancing of assets
● The bottlenecks in your hospital that impact the patient flow, e.g., the average wait time for shared services
● Predicting the need or time to transfer patients from one department to another and anticipating the necessary equipment and staff to transport patients efficiently
The next step is to discover the “unknowns” that can be known, such as:
● What other patterns or bottlenecks can be identified when analyzing inpatient data – from admission to discharge – and more so getting predictions and recommendations for optimization
● What are the questions that we never thought to ask but should
● What additional data should we gather to further identify patterns and opportunities
Once you begin this process, some opportunities present themselves almost immediately. For example, by predicting when the patient will be discharged, hospitals can automate room sanitation and preparation workflows to reduce turnaround time. By integrating EHR, you may discover that patients with certain ailments are consistently above the Geometric Mean Length of Stay (GMLOS). Now drill down the data by time or compare it against other facilities – you will surface correlation factors that explain why the LOS is higher. The potential is limitless once you start conversing with your data and transforming it into impactful and value-creating decisions.
The waste, inefficiency, and frustration now part of the common patient journey should be a wake-up call for health systems. It’s time to move beyond data creation and collection to “data mobilization.” By tapping into your data with AI and ML, health systems can discover invaluable insights into the patient journey and uncover new opportunities that save costs and improve care.
About the author: Rom Eizenberg is CRO and head of product innovation at Kontakt.io, a leader in healthcare RTLS and digital transformation solutions.
Today, we can walk into a FedEx shop and drop a parcel. That parcel will go through numerous trucks and regional, national, and international distribution centers. The parcel will move through multiple airports and several airplanes, and somehow miraculously it appears at the desired destination. Even though drivers might be sick and trucks might break, the parcel will show up – potentially on the other side of the planet – in about 24 hours.
Compared with the often slow-moving and inefficient journey of the average patient – cumbersome forms, insurance nuances, triage, staffing delays, and lengthy wait times – this leads to some uncomfortable questions for health systems. Why can't we do this in healthcare? Why do healthcare leaders feel that they are continuously coping with crisis after crisis? Why do they feel that they are incapable of predicting the future and aligning resources?
These questions concern the factors behind many of the inefficiencies that contribute to the incredible cost of delivering care across the U.S. As an industry, health systems are struggling not with structural challenges, but with modernizing the process of managing health systems, aligning resources around care delivery, and catching up so a hospital can be as predictable, manageable, and optimal as delivering a parcel across the globe in 24 hours.
What prevents this from happening? Two words: data silos. It is almost impossible today for a health system leader to look at the inpatient journey end-to-end, from admission to discharge. Each part of the process of delivering care is contained within a different legacy system, with siloed data, often held by different vendors, and, quite frankly, no one has the big picture. As a result, while today’s hospitals generate an enormous amount of data by second from their care operations, much of the data remains untapped, representing a significant missed opportunity to create value.
Using machine learning (ML) and artificial intelligence (AI), we can dive deep into the data of patient journeys to uncover previously inaccessible insights and opportunities, and ultimately improve hospital operations and patient experiences.
Begin by exploring “known unknowns” and then “unknown unknowns.” In other words, start by finding answers to the questions you know; then evolve into new answers that you aren’t even aware of today.
Examples of “known unknowns” include:
● The discrepancy between planned patient discharge time and actual discharge time
● What factors affect the length of stay (LOS) for patients, such as staff-to-patient ratio, equipment readiness, bed turnover, etc., and to what extent
● Whether the right asset is available in the right place at the right time for efficient load balancing of assets
● The bottlenecks in your hospital that impact the patient flow, e.g., the average wait time for shared services
● Predicting the need or time to transfer patients from one department to another and anticipating the necessary equipment and staff to transport patients efficiently
The next step is to discover the “unknowns” that can be known, such as:
● What other patterns or bottlenecks can be identified when analyzing inpatient data – from admission to discharge – and more so getting predictions and recommendations for optimization
● What are the questions that we never thought to ask but should
● What additional data should we gather to further identify patterns and opportunities
Once you begin this process, some opportunities present themselves almost immediately. For example, by predicting when the patient will be discharged, hospitals can automate room sanitation and preparation workflows to reduce turnaround time. By integrating EHR, you may discover that patients with certain ailments are consistently above the Geometric Mean Length of Stay (GMLOS). Now drill down the data by time or compare it against other facilities – you will surface correlation factors that explain why the LOS is higher. The potential is limitless once you start conversing with your data and transforming it into impactful and value-creating decisions.
The waste, inefficiency, and frustration now part of the common patient journey should be a wake-up call for health systems. It’s time to move beyond data creation and collection to “data mobilization.” By tapping into your data with AI and ML, health systems can discover invaluable insights into the patient journey and uncover new opportunities that save costs and improve care.
About the author: Rom Eizenberg is CRO and head of product innovation at Kontakt.io, a leader in healthcare RTLS and digital transformation solutions.