By: Balaji Ramadoss, Coauthored with Peter Pronovost, MacArthur Fellow, Chief Transformation Office at University Hospitals, Cleveland and co-author of Safe Patients, Smart Hospitals.
Healthcare in the United States harms too often, costs too much, and learns slowly and improves to gradually. Ten percent of all patients are harmed during their interaction with healthcare, and medical error is the third leading cause of death. Only 60% of patients receive recommended therapies resulting in needless suffering, death, and costs.
Healthcare has the ignoble distinction for its massive spending on information technologies while realizing negative labor productivity. Twenty years ago, a 1,000-bed academic health system was staffed with 3,000 employees; today, it is staffed with 12,000 despite the same number of discharges. No other industry has this kind of negative productivity.
Physicians and other care providers spend 40% to 50% of their time on a computer documenting in the EMR. Nurses waste over 20% of their on the following activities:
· Hunting for supplies to manage the "last 10 feet" of the supply chain,
· Double-checking medications when an electronic double-check would be more accurate and efficient
These trends contrast starkly with other high-risk industries that have seen radical improvements in safety and productivity over time. Aviation, rail, nuclear energy, the military, and firefighting improved safety and productivity. Healthcare has not. These statistics are shocking. However, despite our awareness, the outcomes have not substantially improved over the last 20 years. This trillion-dollar waste reduces people's salaries, companies' profits, and money states and municipalities can spend on other social goods and burdens the Federal Government with Dept. The natural question is why. Removing the Trillion-dollar waste from the system can balance budgets, save social security, and allow us to invest in a healthy future. It is for this reason we call it #TheTrillionDollarProblem. In this essay, we explore why healthcare lags behind other safe industries.
One-third of healthcare cost is waste
Healthcare focuses on local improvements at the expense of system-wide improvements
Safe industries have mechanisms and procedures for the sector and system-wide improvements. Aviation, nuclear energy, rail, and military each employ a system-centric focus. Aviation has the Commercial Aviation Safety Teams to conduct sector-wide root cause analysis and implement system-wide solutions to major risks. The National Transportation Safety Board conducts rigorous investigations of accidents. The Nuclear industry created the World Association of Nuclear Operators (WANO), a member organization of nuclear companies that uses standardized tools to conduct peer reviews of nuclear facilities. Although WANO lacks regulatory authority and cannot sanction, it has successfully created a culture of learning rather than judging. Its peer reviews are rigorous and transparent. Risks are shared with the organization, and best practices are distributed to everyone. Consider when a defect was found in the 737. The fleet was grounded, and the problem was fixed.
This sector contrasts starkly with healthcare. Many more people are harmed by the poor usability of medical technology, including faulty design. The solution is always "re-educate" the clinicians. Errors are generally investigated locally by people who lack formal training in risk sciences — most are nurses or lawyers who may identify the clinical components but not the systemic components. Emergency Department patient flow issues cannot be impacted by just focusing on ED alone. Despite this, health systems look to segment process, people, and technology solutions. Segmenting leads to duplication of investments in data, communication, and associated processes.
The investigations are usually fast and superficial, with a limited engagement of device makers to innovate the harm out of the system. As such, innovation to reduce harm sector-wide is slow.
Old-School Organizational Structure
Antiquated operating models also stall innovation. The decades-old "management-consulting" mindset has compartmentalized healthcare into siloes — distinct and separate service lines and business units — attempting to break down the "complexity" into manageable chunks. This compartmentalization was driven by, and a cause for, the reimbursement models of the '80s and the '90s. The push for electronic medical records in the 2000s digitally "hardwired" these silos and "codified" them into place.
As a result, data needed to run operations and improve quality were produced, stored, and locked in in these silos, usually in disparate information systems. As a result, healthcare IT shops spent
the better part of the 2010s integrating data sources. We solved the symptom, not the root cause. Health systems attempted to innovate within, not between, these silos. As a result, health systems lack access to their data to innovate and improve. Compartmentalization is stifling innovation. We believe innovation resides in abundance in 'system-ness.'
"System"-Avoidance
Venture capital and other investors support small-scale innovations, focusing on problems that can be solved in 3 to 5 years with less than $20 million. They avoid systemic healthcare problems, like a moon mission, that take a long time and more money. For example, it is easy to fund an app that helps manage diabetes. While a critical function, the app does not address how to prevent diabetes or use data across a patient's journey to living well with diabetes and coordinate care for the other renal and cardiac diseases these patients often have. Government Funding invests heavily in basic and clinical research yet preciously little in healthcare systems research. The size, magnitude, and complexity of healthcare operational challenges discourage systemic innovation and transformation. That only leaves more apps and fills the cracks and fissures of each silo.
Hiding from complexity
"The lowest hanging fruit is often already rotten or the quickest to rot," Alan Ravitz from The Johns Hopkins University Applied Physics Laboratory. Our mad rush to build and license software and technologies to create "unicorns" has pushed us to go after the easy problems. While there is no small problem in healthcare, the market has focused on the lowest hanging fruit. Healthcare rarely uses disciplined engineering approaches to solve complex problems. And most complex projects lack disciplined requirements documents. As a result, in solving most complex problems, healthcare is like the blind man grabbing the elephant; they feel the part and are blind to the whole. Predictably, solutions intended to improve rarely work and often deteriorate performance. Today's risks in healthcare were yesterday's solution.
The Prediction Addiction
As healthcare organizations continue to implement new and advanced predictive analytics, identifying potential at-risk patients, such as sepsis patients or patients with high potential for
readmission, becomes standard procedure, and operationalizing the data from these tools becomes the focus.
Prediction is a substantial advancement in healthcare operations. However, predicting patient status is not impactful or worthwhile without an associated operational protocol that intervenes based on that prediction. Knowing ten patients may be at risk for sepsis is a giant leap forward. Now, armed with that knowledge, how do we ensure the patients remain safe and healthy? Someone must do something, in the appropriate order, in a prescribed time frame. Is this being done?
All too often, mining data and generating 'output' is touted as the end goal for addressing complex issues. Once a prediction is generated, someone, or something, still needs to intervene. If the workflow for intervention is the same as before, you will not likely realize the model's full value. Prediction without action squanders resources and productivity.
The archaic architecture
Today, information technology and data platforms perpetuate the barriers with architectures that mimic the silos based on the function-based division of labor. Contemporary data platforms are designed around these function-based divisions, which have not evolved much since the 1990s. For example, storage and normalization specialists are still the bottlenecks for today's as-a-service platforms — another specialized silo.
Bloated investments in data architectures such as multiple warehouses, on-premises data marts, cloud-based data lakes, and disparate business intelligence tools tend to make data mired in their operational silos. This trend is observed across all industries. In many cases, business lines within industry verticals tend to have different solutions perpetuating the need for specialists to care for and feed data and technology platforms.
Lack of Aligned Incentives and learning labs
For technologies to be impactful, they need to be designed iteratively with significant user input and deployed with close collaboration with users to provide feedback and iteration. In other high-risk industries, this occurs routinely, and real work experience is simulated. For example, a flight simulator can mimic a flight.
Healthcare lacks incentives for collaborations between technology companies and health systems. Instead, for most technologies, vendors present their wares with limited user input into the design and usability; preciously, few have been evaluated in a clinical workflow. Clinicians recognize opportunities to improve usability and assess the impact of the technology on patient outcomes. Yet these efforts take time and some resources.
Tech companies rarely support this type of effort, either because they are a startup and lack capital, or they are more mature and worry about protecting their IP. Clinicians then are usually reluctant to share their ideas to improve. The only compensation is a higher fee for the technology when the vendor incorporates their ideas.
As a result, society, patients, vendors, and provider systems suffer. They bring clunky and clumsy technologies that often do not improve outcomes, or we do not know if they do. We would accelerate innovation if we created learning labs with mutually beneficial incentives for vendors and clinicians to design and implement high-impact technologies iteratively
The Future
To effectively tackle #TheTrillionDollarProblem we demand a new operating model. With this deeper understanding of healthcare that goes beyond the traditional nuances of regulatory and reimbursement limitations, the newer operating model should enable value engineering, systemic redesign, architecture modernization, and incentive and organizational alignment.
We anticipate venture capitalists and investment funds to pay attention to the need for a systemic redesign of healthcare rather than going after the low-hanging fruits.
A modern healthcare system will effectively manage complexity, not hide from it