Butterworth Health Systems: Redesigning Healthcare Delivery
This is post 17 of 23 in a series on Systems Thinking: Managing Chaos and Complexity by Jamshid Gharajedaghi (ISBN: 978-0-7506-7973-2).
Chapter 9 is a case study. Butterworth Health Systems wanted to redesign itself from the ground up. Not a minor tweak. A full rethink of how healthcare gets delivered. Gharajedaghi and his team at INTERACT worked with Butterworth through six iterations to build a shared vision. The original design document was 200 pages. This chapter condenses it to about 40.
What makes this chapter stand out is that it applies systems thinking to something everyone cares about: healthcare. And the problems it tackles are problems we still deal with today.
The Mess: How Healthcare Got Here
The current healthcare system started as sickness care. You got sick, you went to the doctor, you got treated. Simple enough. But the payment model created a monster.
Three things combined to break it. Fee for service means you pay based on how much care you get. Cost-plus means the provider adds a markup on top of costs. Third-party payer means the person getting care isn’t the one paying. Insurance companies and government programs pick up the tab.
Put those together and you get a system with zero self-control. Patients keep asking because it feels free. Providers keep giving because they get paid more. Insurance companies pass costs along with their own margin. Demand feeds on itself.
Technology made it worse. Every breakthrough extends life expectancy, which creates more demand for the next breakthrough. The cost of delaying death keeps going up exponentially.
The system hit a wall. HMOs showed up to “manage care.” But most didn’t actually manage care. They negotiated volume discounts. That’s not care management. That’s wholesale shopping.
When HMOs did try to manage care, they used bureaucratic rules. A bureaucrat telling a doctor what procedures are allowed? Neither patients nor doctors accepted that. You can’t manage the most emotional part of human life with a mechanical checklist.
So the system pivoted to capitation. Give each doctor a fixed amount per patient and let them figure it out. Sounds reasonable until you realize it breaks the math of insurance. Insurance works because risk gets spread across a large population. Split patients into small groups per doctor and one expensive case can blow up an entire budget.
Gharajedaghi makes a key point here. Even the perception of conflict of interest, whether real or not, is enough to destroy trust. And trust between patient, doctor, and specialist is the single most important thing in any healthcare model.
Design Specifications
So what should the new system actually do? The design team came up with a clear list.
It should balance supply and demand so patients get quality care without fueling wasteful demand. It should dissolve structural conflicts between patients, providers, payers, and administrators. It needs to be flexible, capable of continuous learning, and represent the best current thinking without locking anyone into irreversible social experiments.
And it needs to handle the existing mess without letting old problems infect new solutions. That last point is really smart. Social institutions, once created, take on a life of their own. Causes go away but their effects stick around.
The Architecture: Market Access
The systems architecture has three dimensions: market (how users access care), output (where care happens), and input (knowledge and shared services).
On the market side, users fall into familiar groups. Fee-for-service patients through traditional insurance. HMO members with capitated plans. Self-insured employers. Medicare for seniors. Medicaid for those without resources.
Then there’s a group the existing system largely ignores: noncovered customers. People with no insurance at all. They end up in emergency rooms that get little or no reimbursement. Butterworth proposed giving these people a Health Coverage Card funded through the Butterworth Foundation. Make the community contribution visible and measurable instead of hidden in unpaid ER bills.
I like this idea. Instead of pretending the problem doesn’t exist, you build it into the system design. You create a real mechanism to serve uninsured patients properly.
The Care System: Four Types of Care
This is the core of the redesign. Instead of treating all healthcare as one big blob, the design team split it into four distinct modules.
Some context first. The existing system is almost entirely focused on sickness care. Prevention and wellness are afterthoughts. Nobody built real protocols for preventive care because there was no funding for it.
The design team also recognized that changing how you pay for healthcare isn’t reform by itself. Payment matters, but a real solution has to address operations, technology, and human behavior too.
Common Features Across All Care Types
Every care module shares certain traits. Each one handles capitated patients, fee-for-service patients, or any combination. Each one operates as a standalone business unit with its own finances, product manager, and stakeholder board.
The system encourages people to sign up for all four types of care. But a school might only want preventive care for students. A corporation might only need interventional. Pricing makes the full package cheaper, which is a smart incentive.
Every module has its own learning system. It states assumptions, monitors performance, and adjusts based on what it learns. Every module carries its share of indigent patients.
Rewards work on three levels. Throughput measures volume. Latency measures outcomes and population health over time. Synergy measures cooperation with other modules.
Preventive Care
Preventive care focuses on keeping healthy people healthy and catching problems early. It develops screening protocols, targets groups for testing, rewards early detection, and gives patients open access to clinical expertise. Patients choose their own prevention strategy and primary provider.
Interventional Care
This is what most people think of as “healthcare.” Restoring health through active treatment. It has three levels: primary care for non-life-threatening issues, secondary care for more intensive intervention, and tertiary care for the most complex cases.
The payment model here is a hybrid. Primary care can be capitated. But secondary and tertiary care use a trust fund with fee-for-service elements. This is deliberate. Complex treatments involve too much uncertainty and too many specialists to cap cleanly. The trust fund spreads risk across the system instead of dumping it on individual doctors.
Viability Care
Viability care handles functional limitations from accidents, illness, birth defects, or aging. It has two levels: rehabilitative care (working to restore function) and supportive care (maintaining current function and preventing further decline). The distinction comes down to duration and realistic chances of recovery.
Terminal Care
This is the one nobody wants to talk about. Terminal care addresses patients who are irretrievably dying. Healthcare providers whose duty is to save lives find themselves in an impossible position when facing death. Some try to delay the inevitable at any cost. Others give up entirely.
The Butterworth team made this explicit. They designed a module for supporting patients and families through death. Education and counseling for caregivers. Cost effectiveness built in. Making end-of-life questions part of the system instead of pretending they don’t exist.
I think this is one of the bravest parts of the design. Most healthcare redesigns avoid this topic completely.
Output Dimension: Where Care Actually Happens
Care doesn’t only happen in hospitals. It happens in clinics, offices, labs, nursing homes, fitness centers, and homes. The output dimension organizes all these locations into a network.
The design team considered two structures. A functional structure groups similar facilities together (all hospitals under one leader, all clinics under another). Easy to implement but keeps services fragmented. A modular structure groups all facility types by community. Harder to implement but integrated and community-focused.
Two big ideas stand out. First, emergency rooms should stop being the default entry point for everyone. Decentralized primary care centers handle non-emergency cases, freeing up ERs for real emergencies and saving huge amounts of money. Second, the Health Care Card for uninsured patients makes this restructuring fair. Nobody is denied access. They’re just routed to the right place.
This case study shows systems thinking applied to something as complicated and high-stakes as healthcare. The next post covers knowledge systems, shared services, and governance.
Previous: The Oneida Nation - Learning and Business
Next: Butterworth Health Systems - Knowledge, Services, and Leadership